Copyright TechTarget - All rights reserved ComputerWeekly’s best articles of the day https://cyber.law.harvard.edu/rss/rss.html Techtarget Feed Generator en Wed, 11 Mar 2026 06:44:50 GMT https://www.computerweekly.com editor@computerweekly.com <p>With the pace quickening in the quest to enable intelligent robots designed to work safely and effectively alongside humans across industrial, service and other environments, <a href="https://neura-robotics.com/">Neura Robotics</a> has announced a long-term strategic collaboration with Qualcomm Technologies to advance next-generation robotics and physical artificial intelligence (AI) platforms.</p> <p>By aligning high-performance, power-efficient edge AI with full-stack robotic platforms, the companies say they are endeavouring to help move robotics from research into production-ready deployment at scale.</p> <p>The strategic alliance is also said to reflect Neura’s belief that the race to build cognitive and humanoid robots will not be won alone, but through strong partnerships. By joining forces with leading domain expert technology players, Neura said it was looking to accelerate the real-world deployment of the next generation of <a href="https://www.computerweekly.com/news/366639389/DOCOMO-Keio-University-claim-5G-robot-teleoperation-first">intelligent robots</a>.</p> <p>Together, the companies plan to foster a global developer ecosystem and marketplace for physical AI and robotics applications, encouraging third-party innovation and supporting a build-once, deploy-across-multiple-form-factors approach. The collaboration will also seek to emphasise functional safety, real-time responsiveness and human-centric design as foundational principles, using a data-driven approach to continuously improve reliability, determinism and AI performance.</p> <p>To simplify how physical AI moves from development into production, the partnership includes plans for a standardised runtime and deployment interface to support how AI workloads are deployed, validated and updated across robotic platforms, supporting faster iteration while maintaining reliability and determinism.</p> <p>Said to be consistent with this approach, the collaboration will see the companies focus on what are described as “Brain + Nervous System” reference architectures that combine high-level cognition (perception, reasoning and planning) with “ultra-low” latency and real-time control for robotics systems.</p> <p>It will combine Qualcomm Technologies’ background in AI compute, connectivity and robotics platforms with Neura’s deep robotics system expertise and embodied AI software, with the shared goal of accelerating scalable, real-world robotic intelligence.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about robotics</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366636818/CES-2026-Qualcomm-expands-IEIoT-portfolio">CES 2026: Qualcomm expands IE‑IoT portfolio</a>: Edge AI technology made available for developers, enterprises and OEMs, integrating chipsets, software distribution and tools to scales across verticals.</li> <li><a href="https://www.computerweekly.com/news/366639389/DOCOMO-Keio-University-claim-5G-robot-teleoperation-first">Docomo, Keio University claim 5G robot teleoperation first</a>: Japan’s leading mobile operator and global research institute for haptics announce what they call world's first stable, high-fidelity robot via commercial 5G using low-latency slicing.</li> <li><a href="https://www.computerweekly.com/feature/AI-XR-digital-twins-set-to-transform-robotics">AI, XR, digital twins set to transform robotics</a>: The availability of advanced sensors, artificial intelligence, digital twins, XR and robotics has changed technology-driven markets. We look at how the intersection of these technologies will create commercial opportunities.</li> <li><a href="https://www.computerweekly.com/feature/AI-and-digital-twins-to-serve-increasingly-complex-robot-management">AI and digital twins to serve increasingly complex robot management</a>: Proliferating fleets of robots, diffusion of humanoids, and the emerging robot class of androids will increase complexities of robot management in the future. AI and digital twins will become increasingly important.</li> </ul> </div> </div> <p>Specifically, it will see use of Qualcomm robotics processors, including the <a href="https://www.qualcomm.com/news/releases/2026/01/qualcomm-introduces-a-full-suite-of-robotics-technologies-power">Dragonwing IQ10 Series</a> – bringing together heterogeneous edge computing, edge AI, mixed-criticality systems, software, machine learning operations and an AI data flywheel – physical AI acceleration and software stack and connectivity platforms. These will be paired with Neura’s hardware platforms and embodied AI software stack to enable scalable systems that are claimed to be designed specifically for real-world deployment.</p> <p>It will also align the Qualcomm end-to-end robotics architecture with Neura’s platform strategy to help accelerate robust deployment across multiple robotic form factors.</p> <p>Neura’s Neuraverse platform will serve as a core environment for simulation, training, orchestration and lifecycle management of physical AI workloads for Neura Robots running on Dragonwing Robotics processors. Neura’s robotic systems – including robotic arms, mobile robots, service and household robots, and humanoid platforms – may also serve as reference platforms for development, testing and real-world validation.</p> <p>Neura said it was aiming to accelerate a future where cognitive robots operate safely alongside humans. “This collaboration marks a major step toward making physical AI real: open, scalable and trusted,” said David Reger, CEO and founder of Neura Robotics. “By bringing together our cognitive robotics platforms and the Neuraverse ecosystem with Qualcomm Technologies’ leadership in edge AI and connectivity, we’re aiming to accelerate a future where cognitive robots operate safely alongside humans across industries and throughout everyday life.”</p> <p>Nakul Duggal, executive vice-president and group general manager of automotive, industrial and embedded IoT and robotics at Qualcomm Technologies, said: “Robotics represents one of the most demanding <a href="https://www.computerweekly.com/news/366637293/Motive-accelerates-Edge-AI-safety-for-automotive-operations">edge AI use cases</a>, where decisions must happen instantly, reliably and locally, without relying solely on the cloud for safety-critical responses.</p> <p>“Continued ecosystem development with companies like Neura Robotics helps accelerate scalable, on-device intelligence. Neura’s approach to cognitive robotics reflects a growing shift toward bringing perception and reasoning directly onto the device.”</p> Robotics firm inks strategic collaboration with chip giant to advance next-generation robotics and physical AI, and work jointly on reference architectures for full-stack robotics systems https://cdn.ttgtmedia.com/rms/computerweekly/Neura-Robotics-hero.jpg https://www.computerweekly.com/news/366639921/NEURA-Robotics-accelerates-next-generation-physical-AI Tue, 10 Mar 2026 11:00:00 GMT <p>Two years after it proposed the transition from the mobile internet era to the mobile artificial intelligence (AI) era, leading to the rapid adoption of agents in B2B applications and 30 million agents applied over the past 12 months, Huawei has introduced the AgentVerse, predicting a 10,000-fold increase in agent-handled work in networks by 2030.</p> <p>The proposal of a new paradigm for communications came on the back of the comms tech giant’s Agentic Core Summit at <a href="https://www.mwcbarcelona.com/">MWC 2026</a>, which centred on the strategic theme of building an agentic network with device-network-service synergy.</p> <p>At the summit, Huawei revealed that it had worked with global mobile trade association the <a href="https://www.gsma.com/">GSMA</a> and a range of operators and industry organisations across the Middle East, Asia Pacific, Europe, Latin America and other regions to explore AI-driven advancements for the core network. Together, they unanimously agreed that the 5G core network has entered into “a new phase” called the Agentic Core.</p> <p><a href="https://www.huawei.com/en/news/2026/3/mwc-agenticcore-summit">Huawei’s Agentic Core system</a> integrates AI into mobile internet, voice, operations and maintenance (O&amp;M) and telco cloud infrastructure to allow networks to evolve and main service offerings to be reshaped. Huawei sees AI as extending a core network with three “transformative” abilities: real-time experience awareness; global experience evaluation and resource coordination; and intelligent interaction and execution.</p> <p>This architecture is designed to give rise to a “network brain” that drives a closed-loop experience monetisation model where experiences are definable and assessable, service offerings are marketable, quality is guaranteed and exclusive user identities are perceptible.</p> <p>The intelligent O&amp;M part of the solution is built to &nbsp;transform network operations into an automated and intelligent ecosystem, driving the core network toward Autonomous Network (AN) L4 Phase 2. Phase 1 focuses on the intelligent assistant, NOEMate, which delivers automated closed-loop management for both faults and changes. Building on this, Phase 2 introduces hierarchical autonomy and builds an unmanned factory, achieving full single-domain autonomy within the core network.</p> <p><a href="https://www.computerweekly.com/news/366623324/Digital-Catapult-joins-EU-6G-innovation-consortia">Looking toward the 6G era</a>, Huawei Agentic Core also supports ubiquitous AI agent access, building an agent-based communication network that spans across devices and ecosystems. The Cloud Core Network is designed for an evolving communication infrastructure that will act as an interchange for AI agent network.</p> <p>And these, said Huawei Eric Zhao, vice-president and CMO of Huawei's wireless solution, would operate in the AgentVerse: “Mobile AI is sparking a paradigm shift across the communications industry. With a trillion-scale surge in AgentVerse connections on the horizon, mobile networks need an urgent upgrade.</p> <p>“To unlock the full potential of <a href="https://www.computerweekly.com/news/366623050/T-Mobile-rolls-out-5G-Advanced-across-US">5G-Advanced,</a> the industry should accelerate end-to-end upgrades and innovation, building multidimensional network capabilities that can meet the demands ahead.”</p> <p>At MWC, Huawei argued that agents were reshaping mobile network demands – for example, by evolving into engines of industrial automation and broad societal change. It offered the example of productivity agents making fully automated manufacturing possible through autonomous learning and the precise coordination of thousands of robots. It calculated that by 2030, the global market is expected to reach trillions of intelligent connections worldwide.</p> <p>Zhao added: “AI’s development has gone wide and far beyond our imagination, and it is now becoming clear that the application of AI will be [through] agents. We believe that in the future, every industry, terminal, organisation and individual will be served by agents – and this is why we propose the AgentVerse. Just in last year alone, there was 30 million agents applied in different industries, significantly improving the productivities of verticals; the adoption pace of agents is incredibly fast.</p> <p>“It is estimated that by 2030, the amount of work handled by agents will grow by 10,000 times. Agents adoption means the introduction of changes in communication methods and communication objects. That means, in the future, agents will introduce new interactions, agents will interact with people, agents will interact with agents. This is why we think that the time has changed and the wireless industry needs to be prepared to welcome new services.”</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about AI in networking</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366639409/NTT-DATA-Ericsson-team-to-scale-private-5G-physical-AI-for-enterprises">NTT Data, Ericsson team to scale private 5G, physical AI for enterprises</a>: Global comms tech provider forges global partnership with business and technology services firm to establish 5G as the foundational operating layer for enterprise AI, looking to scale AI-driven operations securely and consistently.</li> <li><a href="https://www.computerweekly.com/news/366639431/Firecell-CloudRANAI-collaborate-to-cut-cost-and-complexity-of-private-5G">Firecell, CloudRAN.AI collaborate to cut cost and complexity of private 5G</a>: Hot on the heels of the Accelleran merger, purpose-built private 5G connectivity company unveils radio portfolio integration to give system integrators faster, more affordable deployment route across industrial applications.</li> <li><a href="https://www.computerweekly.com/news/366636440/2025-a-transformative-year-for-private-connectivity">2025 transformative year’ for private connectivity</a>: Research finds standalone private 5G networks well-positioned to become predominant wireless connectivity medium for Industry 4.0 applications in manufacturing and process industries.</li> <li><a href="https://www.computerweekly.com/news/366638797/Private-LTE-5G-networks-reached-6500-deployments-in-2025">Private LTE/5G networks reached 6,500 deployments in 2025</a>: Analysis of private 5G market finds steadily growing market that is increasingly driven by organic growth.</li> </ul> </div> </div> Comms tech giant paints picture of AI-led communications infrastructure as industry transitions towards a new world of agents https://cdn.ttgtmedia.com/visuals/ComputerWeekly/HeroImages/5G-6G-mobile-network-next-gen-Yingyaipumi-adobe.jpg https://www.computerweekly.com/news/366639878/Huawei-agent-oriented-mobile-networks-to-define-Agent-Verse Tue, 10 Mar 2026 05:51:00 GMT <p>Having<b> </b>seen its products become the cornerstone of business in the artificial intelligence (AI) era, Nvidia is turning its attention on 6G infrastructures, announcing a collaboration with leading telcos to build the world’s “next generation” of wireless networks on AI-native, open, secure and trustworthy platforms.</p> <p>The initiative, announced at <a href="https://www.mwcbarcelona.com/">Mobile World Congress 2026</a>, is intended to represent a shared commitment to ensure 6G infrastructure is open, intelligent, resilient, accelerates innovation and safeguards global trust. Nvidia’s initial partners include Booz Allen, <a href="https://www.computerweekly.com/news/366585832/BT-Group-pushes-back-digital-switchover-deadline">BT Group</a>, Cisco, Deutsche Telekom, Ericsson, Mitre, Nokia, <a target="_blank" href="https://ocudu.org/news/linux-foundation-announces-ocudu-ecosystem-foundation-to-accelerate-open-source-ai-ran-innovation" rel="noopener">OCUDU Ecosystem Foundation</a>, ODC, SK Telecom, SoftBank&nbsp;and T-Mobile.</p> <p>Explaining its participation in the project, Nvidia noted that beyond traditional connectivity, 6G wireless networks will become the fabric for <a target="_blank" href="https://www.nvidia.com/en-us/glossary/generative-physical-ai/" rel="noopener">physical AI</a>, enabling billions of autonomous machines, vehicles, sensors and robots and significantly increasing demands for security and trust.</p> <p>Yet it noted that <a href="https://www.computerweekly.com/news/366639259/Global-5G-standalone-dynamic-shifts-from-coverage-to-capability">legacy wireless architectures</a> were not designed to meet these requirements, creating challenges as networks increase in complexity. To address this, Nvidia said that it was bringing the industry together to advance AI-native, <a href="https://www.computerweekly.com/news/366639378/SK-Telecom-outlines-mid-to-long-term-6G-network-evolution">software-defined wireless platforms</a> built on open and trusted principles.</p> <p>Nvidia added that by embedding AI across the radio access network (RAN),<a href="https://www.computerweekly.com/news/366639110/Telefonica-activates-commercial-Edge-services-in-Spain"> edge and core</a>, 6G networks must enable secure integrated sensing and communications, intelligence and decision-making while supporting interoperability, supply chain resilience and faster innovation.</p> <p>The tech giant predicted that 6G will be AI-native and software-defined, enabling wireless networks to advance at the pace of innovation. Nvidia is a founding member of the&nbsp;<a target="_blank" href="https://ai-ran.org/press-releases/mwc-2026-momentum" rel="noopener">AI-RAN Alliance</a>, which now has more than 130 participating companies driving networking innovation.</p> <p>6G networks built on AI-RAN architecture will continuously evolve through software, enabling real-time intelligence and rapid advancement, according to Nvidia. This transformation opens the door for a diverse ecosystem of participants – from global operators and technology providers to startups, researchers and developers – all contributing through open and programmable platforms.</p> <p>As part of its vision for 6G, Nvidia is participating in global private and public initiatives to advance 6G innovation, contributing open source software, accessible platforms and joint research and development. Among the global projects it is involved in is the FutureG Office-led&nbsp;<a title="OCUDU Initiative" target="_blank" href="https://ocudu.org/news/linux-foundation-announces-ocudu-ecosystem-foundation-to-accelerate-open-source-ai-ran-innovation/" rel="noopener">Ocudu Initiative</a> in the US, aligning with government and industry partners to accelerate open, software-defined and AI-native 6G architectures.</p> <p>In Korea, Nvidia is collaborating with an&nbsp;<a target="_blank" href="https://news.yonsei.ac.kr/en/news/detail?bbSeq=35359" rel="noopener">industry consortium</a> to help shape intelligent, secure, programmable 6G networks from the ground up. Meanwhile in the UK, Nvidia is collaborating with the&nbsp;<a target="_blank" href="https://www.gov.uk/government/publications/memorandum-of-understanding-between-the-uk-and-nvidia-on-ai-and-advanced-connectivity-technologies/memorandum-of-understanding-between-uk-and-nvidia-on-ai-and-advanced-connectivity-technologies" rel="noopener">Department for Science, Innovation and Technology</a>&nbsp;to advance applied research, ecosystem development and trusted AI-native network design.</p> <p>Together, Nvidia stressed that such collaborations represent a unified commitment – supported by like‑minded governments, operators and technology partners – to shape secure, intelligent and trusted global connectivity for the next generation of wireless technology.</p> <p>“AI is redefining computing and driving the largest infrastructure buildout in human history – and telecommunications is next,” said Jensen Huang, founder and CEO of Nvidia. “Together with a global coalition of industry leaders, Nvidia is building AI-RAN to transform the world’s telecom networks into AI infrastructure everywhere.”</p> <p>Allison Kirkby, chief executive of BT Group, added: “Connectivity is the backbone of economic growth, and with this collaboration, we’re helping lay the foundations for a future ecosystem that is intelligent, sustainable and secure. By building on open and trustworthy AI-native platforms, we can simplify future technologies like 6G, ensuring they build on the strengths of today’s 5G networks while still unlocking powerful new capabilities at scale.”</p> <p>Tim Höttges, CEO of Deutsche Telekom, stated: “With an open, intelligent and trusted 6G infrastructure, we are laying the foundation for the era of physical AI and unlocking new value for our customers, for industry and for society.”</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about 6G</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366639378/SK-Telecom-outlines-mid-to-long-term-6G-network-evolution">SK Telecom outlines mid- to long-term 6G network evolution</a>: A paper on 6G development by South Korean operator highlights direction of infrastructure in the AI era, with vision rooted in fundamental network values such as security, stability and quality.</li> <li><a href="https://www.computerweekly.com/news/366631268/Verizon-launches-6G-innovation-forum">Verizon&nbsp;launches 6G innovation forum</a>: Leading US communications company brings together comms technology producers to take early step in shaping the next generation of AI-enabled mobile.</li> <li><a href="https://www.computerweekly.com/news/366637167/Samsung-claims-virtualised-RAN-first-accelerating-AI-native-6G-ready-nets">Samsung claims virtualised RAN first, accelerating AI-native, 6G-ready nets</a>: CE giant Samsung completes commercial call using its own vRAN solution powered by Intel Xeon 6700P-B processor series with up to 72 cores, on a Tier 1 US mobile operator’s live network.</li> <li><a href="https://www.computerweekly.com/news/366633732/NVIDIA-Nokia-pioneer-AI-platform-for-6G-comms">Nvidia, Nokia ‘pioneer’ AI platform for 6G comms</a>: Global comms tech provider and AI leader establish a strategic partnership to enable accelerated development and deployment of next generation AI native mobile networks and AI networking infrastructure.</li> </ul> </div> </div> AI behemoth and global operators and infrastructure providers team to build next-generation mobile infrastructure to advance AI-native 6G innovation based on open and trusted software-defined wireless platforms https://cdn.ttgtmedia.com/visuals/ComputerWeekly/Hero%20Images/6g-mobile-network-broadband-iaremenko-adobe.jpg https://www.computerweekly.com/news/366639876/NVIDIA-teams-with-global-telecom-leaders-for-6G-development Mon, 09 Mar 2026 12:37:00 GMT <p>As artificial intelligence (AI) reshapes the connectivity market to become more aligned with tomorrow’s agentic-oriented business world, comms tech provider Qualcomm has set out its vision, emphasising the evolution from cloud-based AI to <a href="https://www.computerweekly.com/news/366623846/Qualcomm-boosts-UAE-presence-with-AI-IoT-Edge-development">edge AI</a>, enhancing performance, user experience and privacy.</p> <p>Kicking off the company’s presence at MWC 2026, Enrico Salvatori, senior vice-president and president of Qualcomm Europe, said the company would move throughout 2026 by advancing AI models with reduced parameters – particularly in smartphones, wearables and PCs – and launching a new generation of the Snapdragon 8 elite platform featuring a 37% faster neural processing unit.</p> <p>He added that Qualcomm’s hybrid AI architectures would see use in future 6G infrastructures and datacentres, with the aim of integrating cloud and edge AI “seamlessly”. Central to the company’s strategy would be to highlight the importance of execution alongside “visionary” goals, maintaining and extending partnerships with companies that have launched commercial devices using Qualcomm technologies, such as Samsung, Honor and Lenovo.</p> <p>The point, said Salvatori, was executing on plans introduced in 2025. “The opportunity is to have more devices that are smart, adopting AI and [highlighting] the evolution of the original cloud AI with complementary support happening at the edge,” he said. “We are able to grow the performance of platforms. Two things are converging: from one side, AI models are working with better performance [with a] reduced number of parameters, [and] at the edge, you can see more and more inference happening … that is part of the overall architecture of [new] AI, that we call hybrid AI.”</p> <p>One key area in which hybrid was being introduced rapidly was with smartphones, where user experiences in particular have evolved by introducing agentic AI. “We are moving from the original application-centric use of the smartphone to agentic AI,” said Salvatori. “The agent is interfacing with all the apps on the device so you don’t have to enter what you want to do … automatically, the device knows what you want to do. This is not talking about the future. This is what is happening today.” And there was a generality to this – hybrid would extend across the product lines and industry services where, he said, “AI is the new UI [user interface]”.</p> <p>Specific technology announcements included expanded efforts to move from 5G Advanced to <a href="https://www.computerweekly.com/news/366636975/CES-2026-rubber-hits-the-road-for-Qualcomm-automotive">AI-native 6G</a> – &nbsp;revealing that in this area, pre-commercial work is expected in 2028 to coincide with the Los Angeles Olympics, with commercialisation targeted from 2029. It also unveiled AI-driven RAN automation and commercial RAN AI features; a new Wi‑Fi 8 portfolio spanning client and networking infrastructure, a <a href="https://www.qualcomm.com/wearables/products/snapdragon-wear-elite-platform">Snapdragon Wear Elite</a> platform for on-device “personal AI” wearables; an X105 5G Advanced Modem‑RF positioned as the first Release 19-ready modem-RF system; and an on-premise industrial AI and private 5G autonomous factory demo with Siemens.</p> <p>Looking at 6G, the conference saw Qualcomm state 6G was being designed as an AI-native system that builds on three key pillars: connectivity, wide-area sensing, and high-performance compute. These next-generation networks are said to feature new and advanced capabilities, including intelligent radios with integrated wide-area sensing capabilities, virtualised and cloud RAN with high-performance and energy-efficient compute, AI-based network autonomy, and edge and centralised datacentres for entirely new AI workloads.</p> <section class="section main-article-chapter" data-menu-title="Generational opportunity"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Generational opportunity</h2> <p>Qualcomm saw 6G as a generational opportunity for the transformation and growth of the telecom sector, enabled by the combination of wireless, efficient computing and AI. 6G systems would support higher levels of efficiency and performance for telecommunication applications, new agentic consumer and enterprise devices, and new classes of AI-enabled services. These were said to range from context-relevant data, low-altitude aerial and terrestrial traffic management, data insights, and analytics at scale.</p> <p>The company also outlined the partner ecosystem that it was working with. These include firms such as Airtel, Amazon, Asus, BT Group, Cisco, Dell, e&amp;, <a href="https://www.computerweekly.com/news/366627074/Ericsson-recommits-to-UK-6G-research-programme">Ericsson</a>, FPT Corporation, Fujitsu/1finity, Google, HP, HPE, Humain, KDDI, KT, Lenovo, LG Electronics, LG Uplus, Meta, Microsoft, Motorola, NEC Corporation, Nokia, NTT Docomo, Reliance Jio, Samsung Electronics, Sharp, Siemens, <a href="https://www.computerweekly.com/news/366639378/SK-Telecom-outlines-mid-to-long-term-6G-network-evolution">SK Telecom</a>, Snap Inc, Stellantis, Swisscom, Tejas Networks, Telstra, TIM Group, T-Mobile, Viettel Group, VNG and YTL.</p> <p>Ericsson has been a key partner for Qualcomm on developing technology for every mobile generation, and at the Qualcomm MWC event, Marie Hogan, the comms tech provider’s head of 6G portfolio strategy, operations and technology portfolio, business area networks took a step back from mobile generations and looked at the demands on modern networks – in particular, what she said was the “dramatic rise” of AI over the past few years.</p> <p>“We’re moving from having just centralised AI models and datacentres to seeing distributed, autonomous agents, in devices, in cars, in wearables, in humanoid robots,” said Hogan. “And this means a dramatic impact to our mobile networks. In fact, with the rise of AI agents and intelligent AI-native devices, that places a lot of requirements on the mobile infrastructure that we build. It’s not enough just to add more compute. We need more advanced networks. We need to add more uplink, a lot more uplink. Since 2G, we optimised for downlink, especially for latency and performance. Now, we need to look at other aspects.</p> <p>“We [now] need to connect, smart devices; smart wearables,” she said. “<a href="https://www.computerweekly.com/news/366636975/CES-2026-rubber-hits-the-road-for-Qualcomm-automotive">We need to connect vehicles</a>. We need to connect hundreds of millions of homes and enterprises and billions of devices and sensors. We are starting to add intelligence to selected functions in the networks, for example, for autonomy in certain cases, for optimisation of performance in certain use cases. And this is a stepping stone towards the fully capable AI networks that we see when it comes to 6G. With native 6G, as we’re starting to call it, AI will be built as a fundamental principle of the network.”</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about AI in networking</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366639409/NTT-DATA-Ericsson-team-to-scale-private-5G-physical-AI-for-enterprises">NTT Data, Ericsson team to scale private 5G, physical AI for enterprises:</a> Global comms tech provider forges global partnership with business and technology services firm to establish 5G as the foundational operating layer for enterprise AI, looking to scale AI-driven operations securely and consistently.</li> <li><a href="https://www.computerweekly.com/news/366639431/Firecell-CloudRANAI-collaborate-to-cut-cost-and-complexity-of-private-5G">Firecell, CloudRAN.AI collaborate to cut cost and complexity of private 5G</a>: Hot on the heels of the Accelleran merger, purpose-built private 5G connectivity company unveils radio portfolio integration to give system integrators faster, more affordable deployment route across industrial applications.</li> <li><a href="https://www.computerweekly.com/news/366636440/2025-a-transformative-year-for-private-connectivity">2025 transformative year’ for private connectivity</a>: Research finds standalone private 5G networks well-positioned to become predominant wireless connectivity medium for Industry 4.0 applications in manufacturing and process industries.</li> <li><a href="https://www.computerweekly.com/news/366638797/Private-LTE-5G-networks-reached-6500-deployments-in-2025">Private LTE/5G networks reached 6,500 deployments in 2025</a>: Analysis of private 5G market finds steadily growing market that is increasingly driven by organic demand from end users, with WAN and enterprise segments of near equal worth.</li> </ul> </div> </div> <p>In a practical sense, this will see Ericsson construct, for the first time, AI embedded in every layer of the network, from the physical portion, to autonomous control, up to service orchestration. That is creating a network architecture that can bring together compute, connectivity and sensing. Hogan said the task of Ericsson, working with Qualcomm, would be to build an architecture that is both intelligent and “enabling the fabric of the world around it”.</p> <p>The other key new-era comms technology launch centred around Wi-Fi 8. Qualcomm unveiled at MWC an AI-native Wi‑Fi 8 Portfolio that it said would unify client and network connectivity for AI-era performance. The <a target="_blank" href="https://www.qualcomm.com/wi-fi/products/fastconnect/fastconnect8800" rel="noopener">FastConnect 8800 Mobile Connectivity System</a> offers Wi-Fi 8, Bluetooth, Ultra Wideband 802.15.4ab and Thread in a single chip package.</p> <p>It is claimed to be the world’s first Wi-Fi 8 service with a 4x4 Wi-Fi radio configuration, enabling speeds beyond 10 Gbps, up to twice the performance of the company’s previous Wi-Fi 7 generation and up to three times the gigabit wireless range. In addition, through Bluetooth High Data Throughput, top Bluetooth transmission speeds have increased from 2Mbps to 7.5Mbps.</p> <p>The range is intended to see use in smartphones, tablets, laptops, robotics and other connected equipment. All offerings in the Qualcomm Wi-Fi 8 portfolio are sampling immediately, with commercial products expected in late 2026.</p> <p>Outlining his company’s plans for the wireless standard to Computer Weekly, Ganesh Swaminathan, Qualcomm vice-president and general manager of wireless infrastructure and networking, predicted that likely enterprise use cases would be built on multiple access points (APs) and roaming, as well as what he described as “critical features” such as dynamic bandwidth expansion.</p> <p>“In the enterprise, if you have multiple APs in a floor, and oftentimes [for example] during an event, clients are specifically located or connect to a particular AP while other APs are not loaded heavily,” he said. “So, what happens is every AP is trying to blast for the same bandwidth, and they’re trying to take the same amount of time. What Wi-Fi 8 brings to the table is dynamic bandwidth expansion, which means that wherever the client workload is, that AP can maximise the bandwidth while the other APs can be coordinated to manage the lesser bandwidth. You basically put bandwidth where it’s needed and you can manage it across the spectrum.”</p> </section> Leading comms tech platform provider unveils agentic radio access network management service and AI enhancements for commercial RAN platforms to accelerate value for telcos on the path to 6G https://cdn.ttgtmedia.com/visuals/ComputerWeekly/HeroImages/Qualcomm-Snapdragon-Spaces-Dual-Render-Fusion-PR-hero.jpg https://www.computerweekly.com/news/366639874/Qualcomm-plots-out-6G-Wi-Fi-8-future-with-AI-as-the-new-user-interface Mon, 09 Mar 2026 12:30:00 GMT <p>Nick Pearson, CIO at <a href="https://www.computerweekly.com/microscope/opinion/The-UKs-AI-skills-fork-in-the-road">Ricoh Europe</a>, describes his job as energising. The company’s shift in business model presents challenges and significant opportunities for him to draw on previous experience at other blue-chip firms.</p> <p>Pearson joined the printer and tech services supplier in December 2023. He was previously head of IT platforms at Vodafone and held senior tech roles at RS Group and PepsiCo, where he spent almost a decade and was latterly UK IT director. In his role as CIO at Ricoh, Pearson is a member of the European executive board engaged in business transformation.</p> <p>“We’re going through a seismic change that pivots Ricoh from a device and manufacturing company – an asset-based firm – into a services organisation,” he says.</p> <p>“That shift on all levels – the way people think, the way we build systems, the way we sell, the way we operate, and the way we deliver – is energising. And that change is occurring against the backdrop of everything happening in the wider technology space.”</p> <section class="section main-article-chapter" data-menu-title="Assuming responsibility"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Assuming responsibility</h2> <p>Tokyo-listed Ricoh employs close to 80,000 people globally across three core divisions: Japan and Asia-Pacific, EMEA and North America. Pearson reports to the European CEO, who reports to the global CEO.</p> <p>Pearson says the company runs a federated business across its three regions, loosely coupled back to its Tokyo headquarters. The European division employs about 17,000 people and boasts annual revenues of more than £3bn.</p> <p>“For me, it’s a nice opportunity in a chunky organisation,” he says.</p> <p>One element of the loosely coupled relationship with the Japan headquarters is that the company doesn’t have a global IT organisation. Pearson is part of a board in Japan that governs technology and creates a strategy across North America, Europe and Japan. This IT leadership approach was one of the major attractions of the role.</p> <blockquote> <div class="imagecaption alignLeft"> <img src="https://cdn.ttgtmedia.com/rms/computerweekly/Nick-Pearson-Ricoh-PR-140x180px.jpg" alt="Photo of Nick Pearson, CIO at Ricoh Europe"> </div> <p><span style="font-size: 14pt;"><strong><span style="color: #34495e;">“There wasn’t a CIO [at Ricoh] before me. Running technology was a role someone adopted as part of their C-suite responsibilities. Now, a big part of my role is talking about the benefits of technology”</span></strong></span></p> <p><em><span style="color: #34495e;">Nick Pearson, Ricoh Europe</span></em></p> </blockquote> <p>“This model allows us to operate a fairly autonomous but linked IT estate, without over-globalising platforms and services. <a href="https://www.computerweekly.com/news/366570412/CIO-interview-Nigel-Richardson-European-CIO-PepsiCo">At PepsiCo, we pushed globalisation to the extent where, as a CIO or IT director</a>, your role shifted at a regional level into more of a business relationship – you were the director of how you provide those globalised services,” he says.</p> <p>“Ricoh is loosely coupled. And that setup excited me as a CIO, because it gives you the innovation space and the opportunity to deliver solutions that are ultimately closer to the customers here in Europe, rather than a global standard becoming the accepted index.”</p> <p>Pearson says another appealing factor was that he knew Ricoh and respected its work. He had partnered with the firm at PepsiCo, where the technology specialist ran his helpdesk. Pearson recognised from initial discussions with senior executives at Ricoh that reporting directly to the European CEO would allow him to showcase the crucial role of digitisation.</p> <p>“The conversations were about putting technology at the very heart of the business,” he says. “Before I joined, we never really had an IT leadership voice on the board. There wasn’t a CIO before me. Running technology was a role someone adopted as part of their C-suite responsibilities. Now, a big part of my role is talking about the benefits of technology.”</p> </section> <section class="section main-article-chapter" data-menu-title="Shifting priorities"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Shifting priorities</h2> <p>Pearson’s role has two main components. First, he assumes a traditional CIO position, overseeing artificial intelligence (AI), data, digital transformation, cyber security and business applications, while managing an internal team and a growing network of partners.</p> <p>“I provide technology services to 17,000 people in Europe, about 200 applications, and all the services that run enterprise IT,” he says. “That’s my CIO stomping ground where I’m feeding, watering, supporting and operationally running IT.”</p> <p>The other half of Pearson’s role is external facing. As a technology company, Ricoh sells products and services to CIOs. His conversations with his digital peers focus on his company’s developments across a range of areas, from private cloud to software that supports customers to technologies that form part of the company’s <a href="https://www.computerweekly.com/resources/Internet-of-Things-IoT">internet of things</a> (IoT) platform.</p> <p>“I am not the product owner of those end-solutions,” he says. “That responsibility sits in the commercial business. But the platforms and the technologies that support our solutions are a thread of mine that we’re growing. We’re moving away from just being a back-office function and looking at those front-office solutions that we’re providing.”</p> <p>In short, suggests Pearson, some of the products and services his team develops in the back office might become front-office solutions for the firm’s customers. He describes this approach as a customer-zero strategy, where Ricoh Europe uses its own technologies before releasing them to the public.</p> <p>“I’m trying to bring that approach through,” he says. “So, for instance, what are we doing for <a href="https://www.computerweekly.com/microscope/news/365531795/Ricoh-exposes-gap-in-workplace-transformation-experiences">workplace experience</a>? What are we building out in terms of capabilities, and where do they land? Private clouds are a good example. Our customers need a private cloud solution, and so do I. It’s about building something that’s a scalable solution and a win/win for us.”</p> </section> <section class="section main-article-chapter" data-menu-title="Enabling transformation"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Enabling transformation</h2> <p>Pearson says the broad aspiration is to use technology to help power the company’s business transformation from products to services. This shift is already underway. He estimates that about 55% of the company’s revenue now comes from services.</p> <p>Some growth has come from the development of new services in key areas, such as cyber security, cloud, document management and meeting room spaces. Another key growth factor has been acquisitions, including the 2023 purchase of digital infrastructure and managed services company PFH Technology Group.</p> <p>“My job collectively is to look at some decent growth, some declining growth, and explore opportunities to pivot,” says Pearson, reflecting on the transformation.&nbsp;“I need to consider how we run a model for everything from acquisitions through to manufacturing and print, our distribution business, while also growing our IT services organisation.”</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more interviews with tech company CIOs</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366635509/Interview-Art-Hu-global-CIO-Lenovo">Art Hu, global CIO, Lenovo</a>: The IT chief at the PC, servers and storage supplier is using his experience of rolling out tech internally to boost the growing services ambitions of the Chinese tech giant.</li> <li><a href="https://www.computerweekly.com/news/366620510/Interview-Cynthia-Stoddard-CIO-Adobe">Cynthia Stoddard, CIO, Adobe</a>: After nearly 10 years in post, Adobe’s CIO is still driving digital transformation and looking to deliver lasting change through technology.</li> <li><a href="https://www.computerweekly.com/news/366623881/Interview-Rom-Kosla-CIO-Hewlett-Packard-Enterprise">Rom Kosla, CIO, Hewlett Packard Enterprise</a>: The IT chief of one of the world’s biggest tech companies still has to get the technology right – and he’s building on cloud, application consolidation and data to bring new capabilities in AI to his users.</li> </ul> </div> </div> <p>Pearson says his IT strategy to enable this effort has “two brains” – a fit-for-growth element for internal IT and a second part that focuses on growth. When it comes to internal IT, Pearson is striving for operational excellence. He is working to speed up day-to-day technology delivery, and that’s meant taking a detailed look at the underlying architecture.</p> <p>“Ricoh was on a journey, using quite monolithic stacks around Oracle and big back-office systems,” he says.</p> <p>“We need to get to a more agile model, so we’ve got choices and flexibility. So, for example, we’re driving a dual strategy in an area like ERP [enterprise resource planning]. We’ve got Oracle, but we don’t want to slow the business down. We want to offer choice.”</p> <p>When it comes to the second part of his strategy, supporting growth, Pearson says the organisation needs to ensure that cyber security awareness becomes a key part of its DNA, particularly on the manufacturing and distribution side of the business. Meanwhile, continuing to grow the services side of the business means giving people secure access to data-enabled platforms, such as the internet of things (IoT), private cloud and automation.</p> <p>“Our approach to emerging technology is that we’re using AI where it matters,” he says.&nbsp;“We are a very much taking an ‘N-minus-one’ view in the sense that we say, ‘Let’s leverage, let’s partner, let’s scale, let’s use what’s working for our customers and for us internally,’ and we think that’s the best way we’re going to get value from this technology.”</p> </section> <section class="section main-article-chapter" data-menu-title="Driving growth"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Driving growth</h2> <p>Pearson has developed Ricoh Europe’s AI strategy over the past 12 months. He’s set up a tri-party AI council alongside the head of service operations and the commercial manager in Spain, which is one of the company’s most progressive businesses.</p> <p>This council explores opportunities to buy, build and reuse emerging technologies operationally and commercially. Internally, the company is investigating how to use Microsoft Copilot technologies to boost operational productivity. Externally, the organisation considers how data can be used to improve customer processes, such as using AI to boost room-booking capabilities and exploit office space more effectively.</p> <blockquote class="main-article-pullquote"> <div class="main-article-pullquote-inner"> <figure> We’ll be doing AI where it matters, people will understand where it matters, and they’ll be using it where it matters, not because of the fear of missing out </figure> <figcaption> <strong>Nick Pearson, Ricoh Europe</strong> </figcaption> <i class="icon" data-icon="z"></i> </div> </blockquote> <p>Pearson is eager to make some key developments through 2026. His team is rolling out ServiceNow field service management technology to about 2,000 engineers across Europe.</p> <p>“That means we can get better engineer automation, better fixed rates for when people visit clients, and we’ve got a lot of work we’re doing to automate our helpdesk facilities as customers call contact centres to get better problem resolution,” he says.</p> <p>Another priority relates to an overarching initiative called One Ricoh, where his team ensures acquired businesses and their staff can benefit from using a standardised application estate.</p> <p>“We’re looking at our back-office systems, especially around ERP, where we’re bringing companies across Europe into a consolidated solution set. That’s driving efficiency,” he says.</p> <p>Pearson also wants to drive business growth by exploiting information. Ricoh Europe uses the Snowflake data platform and has invested in Microsoft technologies, such as Power BI and Fabric.</p> <p>“We want to power a step change in business capabilities,” he says. “We’re setting up a data factory, which hopefully helps our AI journey as well, with more employees shifting into becoming data product owners on that side.”</p> <p>Looking forward, Pearson suggests the platform he’s building will deliver a strong, digitally enabled business two years from now.</p> <p>“Number one, we will have capabilities in data and applications that mean, whether you’re a small or a big business in Ricoh, selling a single line or every line of our portfolio, you’ll have standard tooling to enable you to do that,” he says.</p> <p>Second, he expects his customer-zero strategy to produce big benefits. “We’ll be out there, proud and innovative in the way we run our company, and showing our workplace experience outside-in.”</p> <p>Finally, Pearson says the organisation will have established a clear data and AI strategy that is recognised by internal staff and external customers.</p> <p>“We’ll be doing AI where it matters, people will understand where it matters, and they’ll be using it where it matters, not because of the fear of missing out,” he says.</p> </section> Working for a company undergoing a major pivot in its business model means variety and opportunity for the supplier’s tech chief https://cdn.ttgtmedia.com/visuals/ComputerWeekly/Hero%20Images/printing-ink-cyan-magenta-yellow-black-rock_the_stock-adobe.jpg https://www.computerweekly.com/news/366639330/Interview-Nick-Pearson-CIO-Ricoh-Europe Mon, 09 Mar 2026 10:15:00 GMT <p>Nscale, an artificial intelligence (AI) factory builder and key plank in the UK government’s AI growth strategy, has announced $2bn of C-series funding.</p> <p>The announcement takes funding for the datacentre builder to just over $4.9bn in just over a year. The funding round is based on a valuation for Nscale of $14.6bn.&nbsp;</p> <p>Much of that is based on Nscale’s datacentre development pipeline of more than <a href="https://www.computerweekly.com/resources/Data-centre-capacity-planning">1.3GW of capacity</a>. It has also reported it has contracted supply for 200,000 <a href="https://www.computerweekly.com/news/366639658/Huge-grid-and-heat-challenges-ahead-as-Nvidia-set-for-1MW-rack">Nvidia GB300 graphics processing units</a> (GPUs) and agreements to supply capacity for Microsoft via many of its planned sites.</p> <p>Last year, Nscale announced its intention to invest $2.5bn in UK datacentres up to 2028. These include fixed sites, but also <a href="https://www.computerweekly.com/news/366639248/First-HPE-datacentre-modules-set-to-land-at-Derbyshire-AI-site">modular datacentres</a>.</p> <p>Key among these is its Loughton development in Essex, where 50MW scaling to 90MW of capacity – claimed as the UK’s largest AI datacentre – using 23,000 GB300 GPUs is slated to go live in Q4 2026.</p> <p>Loughton is a combination of modular build for rapid deployment and fixed capacity, which will take longer to become available.&nbsp;</p> <p>Nscale acquired modular datacentre provider Kontena in mid-2024. A modular datacentre allows for prefabricated units that can be delivered in months, compared with fixed site builds that take 24 to 36 months.&nbsp;</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Nscale is also building datacentres at:</h3> <ul class="default-list"> <li>Glomfjord, Norway, where it has 30MW of capacity and a planned 60MW powered by hydroelectricity. It is operational with modular units;</li> <li>Narvik, Norway, with plans for 230MW expanding to add another 290MW in to so-called Stargate Norway project with OpenAI and Aker SA;&nbsp;</li> <li>Barstow in Texas in a leased facility with services for Microsoft, where it will deploy 104,000 of the GPUs in a 234MW development starting in Q3 of 2027 that is planned to scale to 1.2GW.</li> <li>Sines in Portugal, where initial power draw on 12,600 12,600 Nvidia GB300 GPUs will be around 20MW and will use seawater cooling;&nbsp;</li> <li>Keflavik, Iceland, where deployment of around 4,600 Nvidia GPUs and 15MW of power draw is planned starting during 2026. <a href="https://www.computerweekly.com/news/366621028/The-datacentre-energy-deficit-How-worried-should-operators-be-about-power-outages">Geothermal and hydro power</a> are its energy sources.&nbsp;&nbsp;</li> </ul> </div> </div> <p>Nscale also has available datacentre capacity at leased space in datacentres in two other sites in Norway, and one each in Slough, Iceland and North Carolina.</p> <p>Nscale is registered in the UK with Companies House, and has a registered address in Cheshire with a business administration service.</p> <p>Its founder, Josh Payne, was a founder of Australia-based Arkon Energy, which provided infrastructure for crypto-currency mining.</p> <p>Leaders in the current round of funding are Norwegian industrial investment company Aker ASA, plus US-based private equity investor 8090 Industries. Further investors in this round include Dell, Lenovo, Nokia and Nvidia.&nbsp;&nbsp;</p> <p>The company has described itself as the UK’s “only full-stack sovereign AI infrastructure provider”, and been name-checked as “British firm Nscale” in two government press releases – one relating to work with OpenAI to create sovereign AI compute capacity, and another about its involvement in the Loughton datacentre.</p> <p>Also announced alongside the new funding is the appointment of former Liberal Democrat leader and deputy prime minister Nick Clegg as board member. Clegg is also a partner at investment company Hiro Capital, and former president of global affairs at Meta.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about Nscale</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/feature/Nscale-explained-Everything-you-need-to-know">Nscale explained: Everything you need to know</a>. AI infrastructure provider Nscale has risen to prominence in UK tech circles over the course of the past year, having aligned itself with the government’s AI strategy. But what is Nscale, and who is behind it?</li> <li><a href="https://www.computerweekly.com/news/366631959/AI-infrastucture-provider-Nscale-secures-11bn-in-series-B-funding">AI infrastructure provider Nscale secures $1.1bn in series B funding</a>. A week after it was revealed that Nscale would be working with Microsoft to build the UK's largest supercomputer, the company announces it has secured $1.1bn in series B funding.</li> </ul> </div> </div> Nscale has a pipeline of 1.3GW of capacity across the UK, Norway and the US, with contracted supply of 200,000 Nvidia GPUs, and is name-checked as a British supplier of AI factories https://cdn.ttgtmedia.com/rms/German/AdobeStock_460716446_SergeyBitos-hero.jpeg https://www.computerweekly.com/news/366639801/AI-factory-builder-Nscale-announces-another-2bn-of-funding Mon, 09 Mar 2026 07:15:00 GMT <p>The key question regarding <a href="https://www.computerweekly.com/news/366581516/Edge-AI-explained-Everything-you-need-to-know">edge artificial intelligence (AI)</a> is no longer about its vast business potential, but about where it can be most efficient and deliver faster, measurable results. Early uses across the manufacturing, retail and infrastructure sectors have to date focused on issues such as predictive machine maintenance, tailored, localised analytics in retail stores, and grid monitoring.&nbsp;</p> <p>However, cost constraints, latency and data residency continue to require careful consideration by organisations looking to scale edge AI strategies.&nbsp;</p> <p>“Early deployments should focus on narrow beachhead use cases where ethical, legal and security risks are limited – or clearly outweighed by the benefits,” observes Michaël Bikard, professor of strategy at the <a href="https://www.insead.edu/executive-education/masters-and-execed">Insead business school</a>. “That’s how new technologies have historically entered safety-sensitive domains.”</p> <section class="section main-article-chapter" data-menu-title="Edge AI is being used practically right now"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Edge AI is being used practically right now</h2> <p>Many global businesses have <a href="https://www.computerweekly.com/news/366630160/Some-87-of-enterprises-see-private-wireless-edge-ROI-in-a-year">adopted edge AI in some capacity</a>. However, most deployments remain relatively small and highly specialised, prioritising speed, reliability and energy efficiency over huge, datacentre-like models. They also depend significantly on human oversight and intervention.&nbsp;</p> <p>Current models focus on minimising edge AI’s limitations, rather than ultra-sophisticated models. Most deployments are still hybrid, with humans handling most of the training and performance evaluation, while the model handles local inference.&nbsp;</p> <p>Edge AI systems are optimised to recommend the best course of action, rather than make fully independent decisions. In highly regulated or safety-critical businesses, humans still have the final say.&nbsp;</p> <p>Successful deployments highlight that edge AI is more about ensuring that reliable decisions can be taken closer to where the data is generated, rather than more intelligent technology itself.&nbsp;</p> </section> <section class="section main-article-chapter" data-menu-title="What’s working: Predictive machine maintenance in manufacturing"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What’s working: Predictive machine maintenance in manufacturing</h2> <p><a href="https://www.computerweekly.com/microscope/news/366613959/Schneider-Electric-shares-French-lessons-with-the-UK">Schneider Electric</a> believes it has significantly advanced the <a href="https://www.computerweekly.com/news/366638922/Direct-to-device-connectivity-set-to-underpin-next-generation-of-industrial-IoT">industrial internet of things (IIoT)</a> by using edge AI for real-time predictive maintenance on the factory floor, through local controllers, servers and devices. This is designed to improve operational efficiency while strengthening data security and decreasing latency as well.&nbsp;</p> <p>The company uses edge AI systems to analyse factors such as real-time temperature, vibration and performance to predict machine issues before they occur, which helps decrease production stoppages.&nbsp;</p> <p>It also employs edge AI for automated inspection and image-based barcode reading, which improves product quality. The Cognex AI-based technology can detect objects and shapes, allowing conveyor cameras to automatically reject flawed products.&nbsp;</p> <blockquote class="main-article-pullquote"> <div class="main-article-pullquote-inner"> <figure> Predictive maintenance succeeds when AI is embedded into operational workflows and decision processes, not deployed as a standalone analytics layer </figure> <figcaption> <strong>Himanshu Rai, director at IIM Indore</strong> </figcaption> <i class="icon" data-icon="z"></i> </div> </blockquote> <p>Schneider Electric also focuses on enhanced autonomous machine control through its EcoStruxure Automation Expert virtualised controller system. This connects shop floor IoT devices to edge controllers. The company also uses edge AI to grow yield by analysing variables in real time and reducing waste.&nbsp;</p> <p>Automotive giant <a href="https://www.computerweekly.com/news/366622135/Renault-charges-Ampere-to-drive-future-with-software-defined-vehicles">Renault</a> has also deployed edge AI tools for predictive manufacturing maintenance. This is mainly achieved by supervising welding robots to ensure that welding defects and failure anticipations are flagged in real time, to minimise downtime.&nbsp;</p> <p>Renault’s Industrial Metaverse uses edge AI heavily to analyse real-time data from 12,000 connected machines, which strengthens production lines. This is said to have <a href="https://media.renaultgroup.com/with-re-industry-renault-group-is-launching-an-ambitious-plan-to-transform-its-industrial-base/">helped Renault Group save €270m</a> in 2023.&nbsp;Similarly, Renault’s autonomous control systems conduct visual inspections through edge AI, further freeing up operator time.&nbsp;</p> <p>“Predictive maintenance has emerged as one of the most commercially successful AI use cases; however, technology alone is insufficient. Stalled or underperforming deployments may cite poor data integration, fragmented ownership, or constraints from legacy systems as root causes,” says Himanshu Rai, director at IIM Indore. “Predictive maintenance succeeds when AI is embedded into operational workflows and decision processes, not deployed as a standalone analytics layer.”</p> </section> <section class="section main-article-chapter" data-menu-title="Real-time inventory tracking and decreasing food waste in retail"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Real-time inventory tracking and decreasing food waste in retail</h2> <p><a href="https://www2.hm.com/en_ph/index.html">Fast fashion retailer H&amp;M</a> has partnered with <a href="https://avassa.io/avassa-for-edge-ai/">Avassa</a> in using edge AI to modernise in-store facilities, streamline operational efficiency and improve customer experience. Another focus is making sure applications keep working even when connectivity is down.&nbsp;</p> <p>One of the biggest uses of edge AI is through RFID-enabled tracking, a highly accurate system allowing inventory to be tracked with real-time data. This helps staff find in-store items immediately, significantly cutting down on customer wait times.&nbsp;</p> <p>Other in-store edge AI deployments include smart mirrors in fitting rooms, which connect to local networks and deliver product recommendations. They let buyers see which items are in stock in real time and ask for other sizes if required, without having to leave the fitting room, which considerably enhances the customer experience.&nbsp;</p> <p>Customers can look for items using photos through the TensorFlow Lite edge AI system on the H&amp;M app, too, further speeding up performance.&nbsp;</p> <p>H&amp;M is partnering with Honeywell to use edge AI to optimise lighting, heating and air-conditioning across 90 European stores as well. By gathering data from smart meters and sensors, the system improves real-time energy usage, decreasing costs and carbon footprint at the same time.&nbsp;</p> <p>Similarly, grocery giant <a href="https://www.tesco.com/">Tesco</a> has leaned heavily into edge AI with a recent three-year partnership with Mistral AI to optimise its supply chain and reduce food wastage. One of the models employs dynamic expiry pricing. The system evaluates expiry dates and how fresh produce is, and automatically reduces prices for items expiring soon.&nbsp;</p> <p>This has helped bring Tesco a step closer to its goal of reducing food waste, with <a href="https://www.tescoplc.com/how-tesco-is-working-to-reduce-food-waste/#:~:text=At%20Tesco%2C%20we%20are%20making,change%20within%20the%20food%20industry.">wastage levels</a> across UK operations down by 45% in 2025, compared with 2016/2017 levels. Another major deployment is real-time logistics and shipments tracking across more than 3,000 locations through solar-powered sensors. Tesco also saves 100,000 miles per week by using AI to search for the most efficient delivery routes.&nbsp;</p> <p>Edge AI is used for product demand prediction as well, improving fresh produce shelf life, which decreases the risk of overstocking. This reduces the need for manual checking and improves inventory management across the board.&nbsp;</p> <p>Self-checkout processes have been upgraded with edge AI too, with stores now including smart systems with cameras that use AI and computer vision to monitor real-time packaging behaviour and flag incorrectly scanned items.&nbsp;</p> </section> <section class="section main-article-chapter" data-menu-title="Grid monitoring and maintenance in energy and infrastructure"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Grid monitoring and maintenance in energy and infrastructure&nbsp;</h2> <p>Siemens Energy is successfully revolutionising legacy grid infrastructure to active, intelligent networks through edge AI, enabling them to automatically handle rising demand and fluctuating renewable energy levels.&nbsp;</p> <p>The process includes AI systems such as substations, transformers and sensors, which allows predictive grid maintenance and real-time decision-making. Online sensor devices, such as the Sensformer advanced unit, keep tabs on high-voltage equipment and transformers.&nbsp;</p> <p>Edge AI flags irregularities in temperature, vibration and torque through local data analysis. Operators can then maintain machines as per their current condition, rather than routine checks, avoiding expensive surprise downtimes.&nbsp;</p> <p>Some sensors are virtual physics-informed neural networks (PINNs), used to virtually predict hotspots on things like transformer bushings without physical sensors.&nbsp;</p> <blockquote class="main-article-pullquote"> <div class="main-article-pullquote-inner"> <figure> New technologies gain traction not by being universally superior, but by outperforming the status quo in narrow contexts. In infrastructure, that often means environments requiring continuous, real-time monitoring at a scale or speed that humans or centralised systems cannot sustain </figure> <figcaption> <strong>Michaël Bikard, Insead</strong> </figcaption> <i class="icon" data-icon="z"></i> </div> </blockquote> <p>Another edge AI deployment, dynamic line rating (DLR), analyses line data factors like wind speed and temperature in real time and remits current transmission line capacities. Unlike potentially conservative static numbers, this unlocks 10% to 15% of <a href="https://press.siemens.com/global/en/pressrelease/siemens-strengthens-gridscale-x-offering-enabling-tsos-accelerate-energy-transition#:~:text=%E2%80%9CGridscale%20X%20Dynamic%20Line%20Rating%E2%80%9D%20maximizes%20the%20utilization%20of%20the,reach%20billions%20of%20euros%20annually.">additional capacity</a> more than&nbsp;90% of the time.&nbsp;</p> <p>Siemens also implemented intelligent substations for a hybrid approach, which processes data locally and only shares relevant information to the cloud, improving bandwidth.&nbsp;</p> <p>“New technologies gain traction not by being universally superior, but by outperforming the status quo in narrow contexts. In infrastructure, that often means environments requiring continuous, real-time monitoring at a scale or speed that humans or centralised systems cannot sustain,” Bikard observes.</p> <p>Similarly, Ørsted uses edge AI systems for wind farm optimisation, by analysing data from thousands of turbine sensors, which optimises predictive maintenance too. It also monitors and analyses localised weather patterns like cloud cover and sun intensity, using the technology to better boost battery storage utilisation and solar energy production.&nbsp;&nbsp;</p> </section> <section class="section main-article-chapter" data-menu-title="Edge AI failures"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Edge AI failures</h2> <p>Despite several successful edge AI deployments in the past few years, there are some models which have failed – often very publicly. McDonald’s AI-driven voice ordering trial, deployed across about 100 drive-throughs, was one such case. The fast-food chain launched a three-year partnership with IBM for this project in 2021, which ended in 2024 following several bad reviews.&nbsp;</p> <p>Viral, embarrassing social media videos posted by customers highlighted the system misunderstanding orders, sometimes resulting in hundreds of dollars’ worth of food being included. Mistakes such as adding bacon to ice cream were also common.&nbsp;</p> <p>Other problems included issues with background noise, different human accents and dialects, and unusual local requests.&nbsp;</p> </section> <section class="section main-article-chapter" data-menu-title="What drove success – and where models broke down"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What drove success – and where models broke down</h2> <p>Successful edge AI deployments across Schneider Electric, Tesco and Siemens Energy, among others, had one common trait: they all focused on extremely narrow processes, within broader organisational structures.&nbsp;Launched in very controlled environments, they only scaled incrementally, after rigorous testing and iterations.&nbsp;</p> <p>“Each stage generates learning, not just about performance, but about failure modes, governance and acceptable risk. Those lessons make it possible to move from tightly controlled settings to more complex environments,” Bikard points out.</p> <p>These models also have a very clear ownership and accountability structure, with specific people being responsible for outcomes or issues. These include operators, supervisors, production line managers, shop managers or similar.</p> <blockquote class="main-article-pullquote"> <div class="main-article-pullquote-inner"> <figure> Data quality and domain expertise are more critical than algorithmic sophistication </figure> <figcaption> <strong>Florian Stahl, Mannheim Business School</strong> </figcaption> <i class="icon" data-icon="z"></i> </div> </blockquote> <p>Constant human supervision meant that any issues or downtime with the models could be immediately addressed with minimal repercussions. A hybrid approach between cloud and edge AI was consistently prioritised as well.&nbsp;</p> <p>Successful deployments did not involve any absolutely critical processes either. Even in cases of predictive maintenance, both on factory floors and grids, their purpose was mainly to speed up and optimise the process, rather than take over completely.&nbsp;</p> <p>On the other hand, one of the biggest pitfalls of the McDonald’s model was taking human oversight almost completely out of the loop and giving the system more autonomy than it was designed to handle as a pilot project. This led to serious mistakes, such as adding hundreds of dollars of extra food to orders going nearly unchecked, with customers having little recourse.&nbsp;</p> <p>Another mistake was launching the initial trial across around 100 locations, instead of a few, well-monitored locations, and introducing far too much data at once through various human accents.&nbsp;</p> <p>The model in question was also ill-suited to handling open-ended inputs, the kind which should be expected in a fast food restaurant, which sees a high volume of personalised requests.&nbsp;</p> <p>Finally, McDonald’s being a well-recognised global brand also meant the company had a very small margin of error to launch new features before being potentially criticised by clients, thus requiring far more testing before launch.&nbsp;</p> <p>“One key lesson is that data quality and domain expertise are more critical than algorithmic sophistication,” observes Florian Stahl, chair of quantitative marketing and consumer analytics at <a href="https://www.mannheim-business-school.com/en/">Mannheim Business School</a>. Many early failures can be traced to poorly labelled data, sensor drift, or insufficient understanding of underlying physical processes.</p> </section> <section class="section main-article-chapter" data-menu-title="What’s next?"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What’s next?</h2> <p>As successful edge AI use cases increase, businesses are likely to move away from isolated experiments to more widespread deployments, through cameras, sensors, robots and other machines.&nbsp;</p> <p>This may decrease cloud reliance while speeding up decision-making at the edge. However, the fundamental principle driving successful deployments will remain the same.&nbsp;</p> <p>The most successful edge AI models will still be those that address highly specific tasks and scale incrementally, while having clear oversight, ownership and accountability structures, even if the number of endpoints grows.&nbsp;</p> <p>“Framing adoption as a human-versus-AI contest misses where the real opportunities lie. What matters instead is identifying situations where existing solutions are clearly insufficient,” Bikard concludes.&nbsp;</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about edge AI</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366581516/Edge-AI-explained-Everything-you-need-to-know">Edge AI explained – everything you need to know</a>: In this essential guide, Computer Weekly investigates the current rapid proliferation of artificial intelligence deployed at the edge of networks.</li> <li><a href="https://www.computerweekly.com/news/366637421/Sovereign-and-edge-AI-drive-return-to-on-premise-Kubernetes">Sovereign and edge AI drive return to on-premise Kubernetes</a>: While public cloud services remain popular, the need to control sensitive data and maximise GPU performance is pushing enterprises to deploy Kubernetes in their own datacentres.</li> <li><a href="https://www.computerweekly.com/news/366637293/Motive-accelerates-Edge-AI-safety-for-automotive-operations">Motive&nbsp;accelerates&nbsp;Edge AI&nbsp;safety for automotive operations</a>: Commercial vehicle AI dash cam said to be able to deliver three times more AI processing power, stereo vision and hands-free two-way communication in an all-in-one device.</li> <li><a href="https://www.computerweekly.com/news/366626073/European-telcos-team-to-gain-edge-AI-factories">European telcos team to gain edge, AI factories</a>: Orange, Fastweb, Swisscom, Telefónica and Telenor announce partnership with AI technology leader to build infrastructure to support generative and agentic AI services for regional enterprises.</li> </ul> </div> </div> </section> In the past few years, edge AI has moved beyond experimental pilots and into real deployments across organisations, yet most uses stay narrow in scope rather than focusing on company-wide initiatives https://cdn.ttgtmedia.com/visuals/ComputerWeekly/Hero%20Images/Intel-edge-computer-hero.jpg https://www.computerweekly.com/feature/Edge-AI-Whats-working-and-what-isnt Fri, 06 Mar 2026 11:43:00 GMT <p>Suppliers of <a href="https://www.computerweekly.com/news/366637759/Saudi-Arabia-ordered-to-pay-3m-to-UK-dissident-targeted-with-Pegasus-spyware" target="_blank" rel="noopener">commercial spyware</a> have edged ahead of nation-state threat actors when it comes to the exploitation of zero-day vulnerabilities at scale, according to data released by the Google Threat Intelligence Group (GTIG).</p> <p>In a report titled <i>Look what you made us patch: 2025 zero-days in review</i>, the GTIG team said that of 42 unique zero-days it tracked in 2025, it was able to firmly attribute first exploitation of 15 to <a href="https://www.techtarget.com/searchsecurity/news/366568215/Citizen-Lab-details-ongoing-battle-against-spyware-vendors" target="_blank" rel="noopener">commercial surveillance vendors</a> (CSVs), compared with 12 that were first exploited by nation-states – seven by China, and nine by financially motivated cyber criminals.</p> <p>The data additionally highlight three zero-days that were “likely” exploited by China, and one possibly at the intersection of cyber crime and nation-state activity.</p> <p>The GTIG team, comprising researchers Casey Charrier, James Sadowski, Zander Work, Clement Lecigne, Benoît Stevens and Fred Plan, wrote that despite CSVs increasingly focusing on operational security to obscure their unethical activity, the growth in their activity <a href="https://www.techtarget.com/searchsecurity/news/366569061/Google-Spyware-vendors-are-driving-zero-day-exploitation" target="_blank" rel="noopener">reflected a trend dating back several years</a>.</p> <p>“Historically, traditional state-sponsored cyber espionage groups have been the most prolific attributed users of zero-day vulnerabilities,” they said. “[But] over the last few years, the increase of zero-day exploitation attributed to CSVs and their customers has demonstrated the growing ability of these vendors to provide zero-day access to a wider range of threat actors than ever before.</p> <p>“GTIG has reported extensively on the capabilities CSVs provide their clients, as well as how many CSV customers use zero-day exploits in attacks which erode civil liberties and human rights,” they added.</p> <p>“<a href="https://www.icij.org/investigations/cyprus-confidential/greek-court-convicts-intellexa-founder-tal-dilian-three-others-in-wiretapping-scandal/" target="_blank" rel="noopener">In late 2025, we reported on how Intellexa</a>, a prolific procurer and user of zero-days, adapted its operations and tool suite and continues to deliver extremely capable spyware to high paying customers.”</p> <section class="section main-article-chapter" data-menu-title="China-nexus threat actors"> <h2 class="section-title"><i class="icon" data-icon="1"></i>China-nexus threat actors</h2> <p>Beyond CSVs, China-nexus threat actors were the most prolific exploiters of new zero-days, predominantly focusing on edge and networking devices that are hard to monitor, as they seek to gain long-term footholds in their targets’ operations.</p> <p>GTIG said it was clear that China-nexus espionage actors have become increasingly adept at developing and sharing exploits among themselves, demonstrating their government is prepared to shower them with plentiful technical, and presumably financial, resources – compared with the other “Big Four” states of Iran, North Korea and Russia.</p> <p>Russian cyber criminals, on the other hand, continue to make a killing and remain able to similarly invest in technical expertise, as evidenced last year by <a href="https://www.computerweekly.com/news/366632397/Oracle-patches-E-Business-suite-targeted-by-Cl0p-ransomware" target="_blank" rel="noopener">Cl0p’s extortion campaign targeting flaws in Oracle E-Business Suite</a>, and the exploitation of <a href="https://cloud.google.com/blog/topics/threat-intelligence/exploiting-critical-winrar-vulnerability" target="_blank" rel="noopener">a flaw in the WinRAR file archiver</a> by a group with possible links to <a href="https://www.computerweekly.com/news/366612515/Unmasked-The-Evil-Corp-cyber-gangster-who-worked-for-LockBit" target="_blank" rel="noopener">the long-standing and ever-present Evil Corp crew</a>.</p> </section> <section class="section main-article-chapter" data-menu-title="Overall zero-day volumes remain on par"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Overall zero-day volumes remain on par</h2> <p>All this said, more widely, GTIG observed a total of 90 zero-days under active exploitation during 2025, lower than 2023’s record high of 100, but generally in the 60 to 100 range that has become established since the Covid-19 pandemic.</p> <p>Of these 90 flaws, the raw number and proportion – 43% and 48%, respectively – of these targeted enterprise technology, with zero-days increasingly affecting security and network edge devices, favoured by both cyber criminals and nation-states alike.</p> <p>CSVs, on the other hand, tended to prefer mobile and browser exploits, the overall volume of which is ebbing and flowing – well up on 2024, but about on par with 2023 – likely thanks to more focused actions from the likes of Google on Android and Apple on iOS, which have forced such threat actors to expand or adjust their techniques, leading to the peaks and troughs.</p> <p>Broken out by supplier, GTIG found that the clear majority of zero-days understandably target Microsoft, which accounted for 25 in total. This was followed by Google, with 11; Apple, with eight; Cisco and Fortinet, tied on four; and Ivanti and VMware, with three. Six more suppliers had two zero-days each, and the remaining 20 were split across 20 suppliers.</p> <p>Looking ahead into 2026, GTIG said that as supply-side actors continue their work to make zero-day exploitation tougher for the bad guys – particularly in the mobile space – adversaries will unfortunately continue to hone their skills as well, foreshadowing more expansive techniques and a growing diversity of targets.</p> <p>The team said that enterprise exploitation in particular will widen thanks to the sheer breadth of applications and devices now in use, with only a single-point-of-failure needed for threat actors to engineer a breach.</p> </section> <section class="section main-article-chapter" data-menu-title="The AI factor"> <h2 class="section-title"><i class="icon" data-icon="1"></i>The AI factor</h2> <p>The team also expects artificial intelligence (AI) to accelerate the race between attackers and defenders, with AI increasingly used to automate and scale attacks by accelerating recon activity and, critically, exploit discovery and development.</p> <p>This will put more pressure on defenders to detect and respond to zero-days, but at the same time, they will of course be able to take advantage of AI tools – like agents – in their own work.</p> <p>GTIG also indicated an emerging paradigm for zero-day exploitation in 2026, heralded by <a href="https://www.cybersecuritydive.com/news/china-actor-us-entities-brickstorm-malware/807166/" target="_blank" rel="noopener">the Brickstorm malware campaign</a>, in which data theft “has the potential to enable long-term zero-day development”.</p> <p>Rather than merely stealing sensitive client data, Brickstorm’s actors – known as Warp Panda – used it to target their intellectual property, such as source code and development documents, something they could use to work angles on new zero-days in their victims’ software.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about vulnerabilities and zero-days</h3> <ul class="default-list"> <li>Google and Qualcomm have tag-teamed a serious vulnerability in the chipsets used in Android mobile devices, <a href="https://www.computerweekly.com/news/366639578/Zero-day-in-Android-phone-chips-under-active-attack" target="_blank" rel="noopener">which has been exploited in the wild as a zero-day</a>.</li> <li>Microsoft releases patches for six zero-day flaws in its latest monthly update, many of them related to <a href="https://www.computerweekly.com/news/366638958/February-Patch-Tuesday-Microsoft-drops-six-zero-days" target="_blank" rel="noopener">security feature bypass issues</a>.</li> <li>The number of vulnerabilities to be disclosed in 2026 is almost certain to exceed last year's total, and may be heading towards 100,000, <a href="https://www.computerweekly.com/news/366638949/CVE-volumes-may-plausibly-reach-100000-this-year" target="_blank" rel="noopener">according to analysis</a>.</li> </ul> </div> </div> </section> Exploitation of zero-days by commercial surveillance and spyware developers outpaced exploitation by nation-state actors last year, according to a report https://cdn.ttgtmedia.com/visuals/ComputerWeekly/HeroImages/data-virus-cyber-attack-freshidea-adobe.jpg https://www.computerweekly.com/news/366639774/Spyware-suppliers-exploit-more-zero-days-than-nation-states Thu, 05 Mar 2026 10:00:00 GMT <p>The impending growth of <a href="https://www.computerweekly.com/resources/Data-centre-energy-efficiency-and-green-IT">datacentre rack power draw</a> to 1MW within two years will bring a step change in energy use and potential to waste massive amounts of heat.</p> <p>This means datacentre operators must become <a href="https://www.computerweekly.com/news/366639772/National-Grid-Nebius-and-Emerald-hail-datacentre-power-throttling">responsible partners to the electricity grid</a>, and with just one 1MW rack set to produce as much heat as 200 5kW ovens, the datacentre industry, government and local authorities must remove obstacles to heat re-use.</p> <p>Those are the views of Schneider Electric vice-president, secure power and datacentre division, UK and Ireland, Matthew Baynes, who spoke this week at the Datacentre World event in London about the energy issues around so-called artificial intelligence (AI) factories.</p> <p>“We’re seeing a major shift into very, very extreme high-density applications with Nvidia GPUs [graphics processing units]. This poses extreme challenges for us as an industry. But the number one challenge is energy production, access to energy and being responsible with this critical asset of electricity,” said Baynes.</p> <p>“Recently, we were designing up to 10, 15, 20, 40kW per rack in traditional datacentres housing cloud applications, but they are <a href="https://www.computerweekly.com/news/366623023/How-AI-workloads-are-reshaping-datacentre-design">moving to becoming AI factories</a>, so that predictable nature we had is harder for us to understand. The pace at which Nvidia is moving with the GPU technology makes it very, very difficult,” he added.</p> <p>Here, the context is Nvidia’s roadmap for GPU products. While right now the densest datacentre racks running Nvidia Blackwell GPUs draw well under 200kW, that’s set to multiply by five or six times from around 2028.</p> <p>Firstly, <a href="https://www.computerweekly.com/news/366636948/Nvidia-unveils-Vera-Rubin-architecture-to-power-AI-agents">Nvidia’s Rubin</a> and Rubin-plus GPUs are set to take rack power draw from around 240kW this year to 600kW-plus in 2027. Then will come Nvidia’s Feynman GPU hardware, which will attain 1MW in a rack. Their power draw will be up to 2kW, each, with 576 per rack.</p> <p>From a technical point of view, this will mandate direct-to-chip liquid cooling, where air cooling has sufficed until now. It will also require a move to 800V DC power distribution to the rack, which is where Schneider’s product set intersects with the issue.&nbsp;</p> <p>At the high level, this all equates to massive and rapidly increasing demand for power, with new datacentre capacity projected to reach 240GW globally by 2030.&nbsp;</p> <blockquote class="main-article-pullquote"> <div class="main-article-pullquote-inner"> <figure> Datacentres need to be a stable asset on the grid, not one that is causing disruption </figure> <figcaption> <strong>Matthew Baynes, Schneider Electric</strong> </figcaption> <i class="icon" data-icon="z"></i> </div> </blockquote> <p>For the datacentre industry, Baynes said, that translates to the need to mitigate energy supply constraints, which include flexible use of the grid and making best use of the heat produced in datacentre operations.</p> <p>“I was at a conference yesterday with ministers, and the grid situation is challenging, with something like 8GW of applications on the London grid for datacentres alone,” said Baynes.&nbsp;</p> <p>“Datacentres need to be a stable asset on the grid, not one that is causing disruption, not one that is having dynamic heat loads on and off. You’ve got to offer technology and be an active asset to the grid so that datacentres can run at their most efficient and be the least disruptive they can be to the national grid,” he added.</p> <p>“We also need to innovate on heat re-use,” said Baynes. “We need to harness some of that hybrid waste heat. We need to work with communities and see how we can leverage that and actually be part of society as well as an industry.</p> <p>“The challenges are government and regulation, [the lack of] district heat networks, the planning permission for that, and then having off-takers for it. It’s not as simple as grabbing the heat and putting it into a swimming pool. There’s a lot more in between those two equations that need to happen, a lot more technology that needs to be implemented.”</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about datacentre energy use&nbsp;</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366634715/Datacentre-energy-demands-set-to-soar-by-2030-as-AI-growth-accelerates-predicts-Gartner">Datacentre energy demands set to soar by 2030 as AI growth accelerates, predicts Gartner</a>: IT market watcher Gartner has shared its projections about how the energy consumption habits of datacentres are set to change as AI takes off.&nbsp;</li> <li><a href="https://www.computerweekly.com/news/366639772/National-Grid-Nebius-and-Emerald-hail-datacentre-power-throttling">National Grid, Nebius and Emerald hail datacentre power throttling</a>: In a UK-first trial, Emerald AI acts as intelligence in datacentre energy management to throttle demand at peak loads, including being able to respond rapidly to energy system stress.&nbsp;</li> </ul> </div> </div> With Nvidia Feynman in 2028, 1MW datacentre racks will produce as much heat as 200 5kW ovens. Industry and government must respond, says Schneider Electric UK datacentre executive https://cdn.ttgtmedia.com/visuals/ComputerWeekly/Hero%20Images/solar-flare-storm-Earth-Sun-SN-adobe.jpg https://www.computerweekly.com/news/366639658/Huge-grid-and-heat-challenges-ahead-as-Nvidia-set-for-1MW-rack Thu, 05 Mar 2026 08:15:00 GMT <p>Over the past few years, <a href="https://www.techtarget.com/searchenterpriseai/definition/edge-AI">edge artificial intelligence (AI)</a> has quickly transformed from a niche technology to a vital and strategic necessity. This is mainly because it helps resolve or minimise some of the key bottlenecks of traditional cloud-based AI. These include data volume, latency, privacy and cost, among others, while allowing companies to make instant decisions to keep up with modern and increasingly automated operations.&nbsp;</p> <p>As a result, the deployment of edge AI is no longer only a technical architecture choice, but one that is actively reshaping risk, cost, compliance and responsibility for enterprises. Businesses are increasingly choosing to store sensitive information mainly on local networks, instead of relying on cloud providers, which has further driven the growth of edge AI.</p> <p>Rather than asking whether or not to adopt edge AI, the crucial question for most companies is how to do so without creating new security, cost and governance issues. As a relatively new technology still, several companies risk implementing edge AI simply to jump on the AI bandwagon, without being fully aware of which situations can most benefit from it.&nbsp;</p> <p>“Edge AI attracts a lot of enthusiasm because it enables real-time, autonomous decisions. However, the real danger is a false sense of technological maturity,” notes Michaël Bikard, professor of strategy at the <a href="https://www.insead.edu/executive-education/masters-and-execed">Insead business school</a>. “Edge AI can work well locally while producing fragile outcomes at the system level. Historically, that’s when failures occur. Not because the technology fails, but because it is trusted too early, before institutions, organisations and governance are ready.”</p> <p>As such, understanding the consequences of edge AI deployment is paramount to deciding long-term strategy.&nbsp;</p> <section class="section main-article-chapter" data-menu-title="Why businesses are moving from cloud-first to hybrid"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Why businesses are moving from cloud-first to hybrid</h2> <p>Businesses are increasingly choosing a more <a href="https://www.computerweekly.com/news/366623846/Qualcomm-boosts-UAE-presence-with-AI-IoT-Edge-development">hybrid AI approach</a> over a cloud-first strategy, driven mainly by larger and more complex AI workloads. Many firms have also been disappointed by the savings achieved by adopting a full public cloud strategy, instead being faced with sharply surging operational costs.&nbsp;</p> <p>These costs, exacerbated by data-heavy applications, mainly arose from moving large datasets to and from the cloud and between providers. Surprise fees and unpredictable bills have further strained IT budgets and complicated budgeting and forecasts.&nbsp;&nbsp;</p> <blockquote class="main-article-pullquote"> <div class="main-article-pullquote-inner"> <figure> Edge AI attracts a lot of enthusiasm because it enables real-time, autonomous decisions. However, the real danger is a false sense of technological maturity </figure> <figcaption> <strong>Michaël Bikard, Insead</strong> </figcaption> <i class="icon" data-icon="z"></i> </div> </blockquote> <p>On the other hand, with edge AI, companies can run stable and predictable workloads on-premise much cheaper than in the cloud.&nbsp;</p> <p>Latency is another overarching concern. Edge AI can often be better than the cloud to minimise latency for applications which need real-time, high-speed processing. These include operational control systems and local analytics, among others.&nbsp;</p> <p>In highly regulated industries such as finance and healthcare, some data may only be stored within certain jurisdictions, which has further driven the shift to edge AI or&nbsp;<a href="https://www.computerweekly.com/news/366630160/Some-87-of-enterprises-see-private-wireless-edge-ROI-in-a-year">on-premise solutions</a>.</p> <p>Major, single cloud providers can also come with supplier lock-ins, while multicloud environments are increasingly complicated to manage, also leading to hybrid approaches.</p> <p>A hybrid strategy lets companies use public cloud to train and update applications which need to scale fast, while keeping high-volume, sensitive or stable data on-premise.&nbsp;This allows organisations to balance agility, cost efficiency and operational resilience, especially in a global context where real-time intelligence is increasingly valuable.&nbsp;</p> </section> <section class="section main-article-chapter" data-menu-title="Edge AI business drivers: What’s real and what’s noise"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Edge AI business drivers: What’s real and what’s noise&nbsp;</h2> <p>At present, most businesses using edge AI have adopted the technology due to practical operational needs. Successful deployments have focused on solving specific, cloud-only limitations, rather than trying to overhaul entire company tech infrastructures.</p> <p>The need for real-time decision-making has primarily driven edge AI adoption, especially in sectors like infrastructure, logistics, manufacturing and transport. This is especially as latency can have far-reaching operational and financial consequences, which the technology can help significantly in cutting down.&nbsp;</p> <p>Applying edge AI to these sectors helps companies process data closer to where it is generated, which enables them to react faster during times of lost central connectivity.</p> <p>The technology also helps organisations dealing with sensitive data stay legally and financially compliant in jurisdictions with especially strict data storage laws.&nbsp;</p> <p>For companies working on critical operations, edge AI can greatly improve operational resilience by making sure that data and intelligence are distributed throughout a number of locations. This helps reduce dependence on centralised systems, which in turn decreases the <a href="https://www.computerweekly.com/news/366630419/Rising-network-outages-are-proving-costly-to-businesses">impact of outages</a>.</p> <p>However, some business drivers are vastly overestimated when it comes to influencing the need to implement edge AI. The biggest of these is short-term cost savings. Edge AI can certainly cut down on transfer and cloud data consumption costs in the long-run.</p> <p>However, it initially needs significant capital expenditure, mainly in the form of hardware device upgrades. There are also ongoing maintenance, monitoring and software update costs following implementation. In some cases, integration with legacy systems may be slower than expected and businesses may have to hire specialised labour as well. Edge AI systems also use considerable amounts of power, leading to higher energy bills.</p> <p>These factors can all cause costs to be higher in the first few months, requiring businesses to have a long-term view when it comes to seeing strategic benefits from edge AI.</p> <p>Another notion that is often overestimated is edge AI being able to deliver anything like “super-intelligence”, by running huge, complicated models like <a href="https://www.computerweekly.com/opinion/A-call-to-action-to-rebuild-the-UKs-semiconductor-industry">datacentre graphics processing units</a>. However, given current computing and power restrictions in most cases, this scenario is highly unlikely at the moment.</p> <p>Similarly, expectations of businesses being able to switch entirely to edge AI, instead of a hybrid approach, are also unrealistic, mainly because of practical deployment, integration and maintenance limitations across various locations.&nbsp;</p> </section> <section class="section main-article-chapter" data-menu-title="How edge AI is changing security, governance and ownership"> <h2 class="section-title"><i class="icon" data-icon="1"></i>How edge AI is changing security, governance and ownership</h2> <p>As edge AI becomes more embedded in hybrid business tech strategies, risk management, enterprise security and governance are also changing, moving away from centralised IT control. These areas are now being shaped by local operational teams taking increasingly autonomous decisions, factoring in the real-time conditions of critical physical infrastructure.</p> <p>Rising edge AI usage could heighten security concerns as well, as it widens organisational attack surfaces through multiple distributed devices and infrastructure. These then need to be protected, monitored and updated equally, following a set of standard guidelines, despite each of them presenting their own unique limitations.&nbsp;</p> <blockquote class="main-article-pullquote"> <div class="main-article-pullquote-inner"> <figure> AI systems can perform exceptionally well under conditions similar to their training data, yet fail abruptly under rare, extreme, or novel scenarios – precisely the situations that matter most in critical infrastructure </figure> <figcaption> <strong>Florian Stahl, Mannheim Business School</strong> </figcaption> <i class="icon" data-icon="z"></i> </div> </blockquote> <p>“AI systems can perform exceptionally well under conditions similar to their training data, yet fail abruptly under rare, extreme, or novel scenarios – precisely the situations that matter most in critical infrastructure,” remarks Florian Stahl, chair of quantitative marketing and consumer analytics at <a href="https://www.mannheim-business-school.com/en/">Mannheim Business School</a>.</p> <p>Patch management can pose more issues with edge AI as well, with thousands of endpoints and vulnerabilities causing potential delays and discrepancies in maintenance.&nbsp;</p> <p>With edge AI being all about local deployments, more questions around version control, oversight and audit issues can arise. This means that companies may need to maintain more in-depth and regular records about data inputs, decision-making processes and operational factors.&nbsp;Highly regulated industries may especially demand evidence trails and seek greater accountability, which can impact company reputations and licences.&nbsp;</p> <p>“Real-time AI systems, particularly those based on machine learning, often operate as ‘black boxes’, making it difficult to explain or audit decisions when failures occur. This lack of transparency is problematic in infrastructures where accountability and post-incident analysis are essential,” Stahl adds.&nbsp;</p> <p>As autonomous decisions taken locally can have very real financial, safety and compliance consequences, businesses may be compelled to take accountability far more seriously if they choose to use edge AI.&nbsp;</p> <p>Senior leadership may also need to adapt centralised organisational and governance models to a more distributed intelligence strategy, all while keeping costs low.&nbsp;</p> <p>These factors have led to edge AI becoming a structural change just as much as a technical one, impacting how and where decisions are taken, how risk is evaluated and overall accountability.</p> </section> <section class="section main-article-chapter" data-menu-title="What leaders should consider before implementing edge AI"> <h2 class="section-title"><i class="icon" data-icon="1"></i>What leaders should consider before implementing edge AI&nbsp;</h2> <p>Given the considerable initial investment required by most edge AI models, leaders should prioritise long-term strategic impact, rather than the hype of the latest technology. This means that while evaluating company-readiness, apart from timing, the potential scope of the intended edge AI model is paramount.</p> <p>The biggest factor to consider is which processes or systems are most likely to benefit from using edge AI first and which can wait for a few more months. Ideally, businesses should prioritise any processes where latency, operational risk and data locality are most critical. By doing this, organisations can spread out costs while testing new deployments in a relatively lower-risk manner.&nbsp;</p> <p>“Importantly, organisations should evaluate AI deployments not only through efficiency metrics, but also through risk-adjusted performance indicators, recognising that marginal efficiency gains are rarely justified if they introduce disproportionate systemic or ethical risks,” Stahl advises.</p> <p>The next question is: to scale or not to scale? In several cases, a pilot edge AI deployment is either enough for the short-term, does not deliver the expected results, or highlights many hidden costs and operational issues.&nbsp;</p> <p>In these cases, decision-makers need to evaluate whether it is worth taking the risk to scale, which will need more investment, specialised skills and manpower.</p> <p>However, knowing when not to use edge AI, and when it could cause more harm than good, is equally important for businesses. This is primarily in cases where data volumes are still low, latency is not crucial, or the company does not have the means to appropriately handle several distributed endpoints.</p> <p>“Edge AI should not be deployed in sectors where use cases are broad, stakes are high, and the consequences of errors are poorly understood,” Insead’s Bikard states. “That combination usually signals a timing problem rather than a technological one. In open, highly interconnected environments, even small mistakes can cascade before organisations have time to respond.”</p> <p>In such cases, exercising strategic restraint is far more instrumental to long-term value.&nbsp;</p> </section> <section class="section main-article-chapter" data-menu-title="From tech choice to organisational shift"> <h2 class="section-title"><i class="icon" data-icon="1"></i>From tech choice to organisational shift</h2> <p>Ultimately, implementing edge AI models should be primarily focused on delivering long-term, strategic value, rather than a trend-based decision. This is especially true if latency and real-time data analysis pose real risks. Businesses need to consider that edge AI use is likely to reshape everything from cost structures and decision-making to autonomy and risk, and prepare accordingly.</p> <p>“There are real potential gains from using AI for predictive maintenance, but those gains rarely come from the technology alone. For AI to pay off, the surrounding organisation – its incentives, culture, structures and skills – must also adapt. Predictions only create value if people are empowered to act on them,” Bikard concludes.&nbsp;</p> <p>Enterprises that treat edge AI as an entire operational shift, rather than an independent feature to be tacked onto legacy systems, will inevitably be able to take advantage of it better in the long run.&nbsp;&nbsp;</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about edge AI</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366637421/Sovereign-and-edge-AI-drive-return-to-on-premise-Kubernetes">Sovereign and edge AI drive return to on-premise Kubernetes</a>: While public cloud services remain popular, the need to control sensitive data and maximise GPU performance is pushing enterprises to deploy Kubernetes in their own datacentres.</li> <li><a href="https://www.computerweekly.com/news/366637293/Motive-accelerates-Edge-AI-safety-for-automotive-operations">Motive&nbsp;accelerates&nbsp;Edge AI&nbsp;safety for automotive operations</a>: Commercial vehicle AI dash cam said to be able to deliver three times more AI processing power, stereo vision and hands-free two-way communication in an all-in-one device.</li> <li><a href="https://www.computerweekly.com/news/366626152/Ericsson-Supermicro-advance-enterprise-connectivity-for-edge-AI-systems">Ericsson, Supermicro advance enterprise connectivity for edge AI systems</a>: Global comms tech provider and application-optimised IT services firm collaborate to simplify and accelerate edge AI deployments by bundling 5G connectivity with pre-validated AI compute.</li> <li><a href="https://www.computerweekly.com/news/366626073/European-telcos-team-to-gain-edge-AI-factories">European telcos team to gain edge, AI factories</a>: Orange, Fastweb, Swisscom, Telefónica and Telenor announce partnership with AI technology leader to build infrastructure to support generative and agentic AI services for regional enterprises.</li> </ul> </div> </div> </section> Right now, rather than asking whether or not to adopt edge artificial intelligence, the crucial question for most companies is how to do so without creating new security, cost and governance issues https://cdn.ttgtmedia.com/visuals/German/Hero-Elshad-Karimov-Einstieg-AdobeStock_879947262.jpg https://www.computerweekly.com/feature/Edge-AI-Business-cost-risk-and-control Thu, 05 Mar 2026 07:40:00 GMT <p>A zero-day <a href="https://www.techtarget.com/searchsecurity/definition/zero-day-vulnerability" target="_blank" rel="noopener">vulnerability</a> in the Qualcomm chipsets used by many <a href="https://www.techtarget.com/searchmobilecomputing/definition/Android-OS" target="_blank" rel="noopener">Android mobile devices</a> is being actively exploited in the wild, according to Google, and system users should apply the relevant updates as soon as possible.</p> <p>Tracked as <a href="https://www.cve.org/CVERecord?id=CVE-2026-21385" target="_blank" rel="noopener">CVE-2026-21385</a>, the flaw is a memory corruption vulnerability that arises from an integer overflow or graphics wraparound condition. Left unaddressed, it enables a threat actor to bypass security controls and take over the targeted system.</p> <p><a href="https://docs.qualcomm.com/securitybulletin/march-2026-bulletin.html#_cve-2026-21385" target="_blank" rel="noopener">It affects well over 200 chipsets in widespread use</a>, according to Qualcomm, which said it was first reported in December 2025 by the Google Android Security Team, and it notified its own customers on 2 February 2026, with fixes rolling out as long ago as January.</p> <p>In its <a href="https://source.android.com/docs/security/bulletin/2026/2026-03-01" target="_blank" rel="noopener">March Security Bulletin</a>, which additionally addresses over 100 other flaws in Android and related components thereof, Google said “there are indications that CVE-2026-21385 may be under limited, targeted exploitation”.</p> <p>Google’s choice of wording suggests that CVE-2026-21385 is being used by a state-linked surveillance operation as, historically, this has been the case with a great many zero-days that ultimately endanger smartphone devices.</p> <p>However, Google has made no firm statement on this point, and nor has it provided any information on the minutiae of the attacks, or their victims, to date.</p> <p>In its bulletin, Google additionally flagged <a href="https://www.cve.org/CVERecord?id=CVE-2026-0047" target="_blank" rel="noopener">CVE-2026-0047</a>, a critical elevation of privilege (EoP) flaw, and <a href="https://www.cve.org/CVERecord?id=CVE-2026-0006" target="_blank" rel="noopener">CVE-2026-0006</a>, a remote code execution (RCE) flaw, as warranting close attention from defenders.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about mobile security</h3> <ul class="default-list"> <li>To manage security threats across devices and networks, IT administrators should implement tools and best practices <a href="https://www.techtarget.com/searchmobilecomputing/answer/What-mobile-network-security-tools-should-organizations-use" target="_blank" rel="noopener">for strong mobile network security</a>.</li> <li>To stay on top of new threats, IT pros can test their skills with mobile security training. Explore the top programmes <a href="https://www.techtarget.com/searchmobilecomputing/feature/Top-mobile-security-courses-and-certifications-for-IT" target="_blank" rel="noopener">to learn about mobile attacks, penetration testing and more</a>.</li> <li>Do concerns of malware, social engineering and unpatched software on employee mobile devices have you up at night? One good place to start is <a href="https://www.techtarget.com/searchmobilecomputing/tip/Building-mobile-security-awareness-training-for-end-users" target="_blank" rel="noopener">mobile security awareness training</a>.</li> </ul> </div> </div> <p>Adam Boynton, senior enterprise strategy manager at <a href="https://www.jamf.com/">Jamf</a>, a specialist in Android and iOS security, said the Qualcomm zero-day would be of particular concern to security teams because – although it has been patched by Google – it is OEMs and mobile carriers who really control when the patch trickles down and reaches the actual devices in people’s pockets.</p> <p>“In enterprise environments, that gap can stretch from days to months – and during that window, the vulnerability is public and the device is exposed,” he explained.</p> <p>“Mobile is no longer a secondary attack surface, and organisations that treat it as such, by delaying updates, will be the ones that end up in incident reports.”</p> <p>As of Tuesday 3 March, CVE-2026-21385 has also now been added to the Cybersecurity and Infrastructure Security Agency’s <a href="https://www.cisa.gov/known-exploited-vulnerabilities-catalog" target="_blank" rel="noopener">Known Exploited Vulnerabilities</a> catalogue. This obliges all agencies of the Federal Civilian Executive Branch in the US to apply the Android patches by 24 March, and is a further signal of the potential scope, and damage, of the vulnerability to the wider enterprise community.</p> <section class="section main-article-chapter" data-menu-title="Apple not neglected"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Apple not neglected</h2> <p>Meanwhile, on 3 March, Google’s in-house Threat Intelligence Group (GTIG) released details of a powerful exploit kit <a href="https://cloud.google.com/blog/topics/threat-intelligence/coruna-powerful-ios-exploit-kit" target="_blank" rel="noopener">targeting Apple iPhone models</a> running versions 13.0 through 17.2.01 of iOS.</p> <p>The so-called Coruna kit is said to contain a set of five comprehensive iOS exploit chains comprising 23 total exploits – the most advanced of which use exploitation techniques and mitigation bypasses that are not yet public.</p> <p>GTIG said it had tracked its use by a customer of an unnamed commercial spyware supplier, in a series of watering hole attacks targeting Ukrainian users, linked to Russian intelligence, and in a broad-scale campaign conducted by a financially motivated cyber criminal operator hailing from China – tracked as UNC6353.</p> <p>“How this proliferation occurred is unclear, but it suggests an active market for second-hand zero-day exploits,” the GTIG team said in their write-up.</p> <p>“Beyond these identified exploits, multiple threat actors have now acquired advanced exploitation techniques that can be reused and modified with newly identified vulnerabilities.”</p> <p>GTIG noted that Coruna is ineffective against devices running the latest version of iOS and encouraged all users to update their devices – or <a href="https://www.computerweekly.com/news/252522455/Amid-NSO-lawsuit-Apple-expands-spyware-protections" target="_blank" rel="noopener">enable Lockdown Mode</a> if this is not yet possible.</p> </section> Google and Qualcomm have tag-teamed a serious vulnerability in the chipsets used in Android mobile devices, which has been exploited in the wild as a zero-day https://cdn.ttgtmedia.com/visuals/ComputerWeekly/HeroImages/security-tablet-smartphone-device-andranik123-adobe.jpg https://www.computerweekly.com/news/366639578/Zero-day-in-Android-phone-chips-under-active-attack Wed, 04 Mar 2026 11:45:00 GMT <p>Expect datacentre spending to increase tenfold. That was among the claims Nvidia CEO Jensen Huang made during the company’s latest quarterly earnings call. He forecast that capital expenditure (CapEx) on datacentres would increase from the $300-400bn mark today to $3-4tn by 2030.</p> <p>Huang’s remarks at the end of February came a few days after <a href="https://www.computerweekly.com/news/366639308/Microsoft-CEO-opens-up-London-AI-tour-with-Copilot-push">Microsoft’s AI Tour London</a> event, when CEO Satya Nadella effectively called for enterprise software developers to use the capabilities now built into Microsoft 365 to create agentic AI workflows for streamlining business processes.</p> <p>Nadella discussed the need to have an efficient token factory, where phrases or tokens can be streamed into AI engines that interpret natural language for querying large language models (LLMs). The Microsoft vision of enterprise AI is built on the <a href="https://www.computerweekly.com/blog/Cliff-Sarans-Enterprise-blog/The-Q-in-AI">M365 foundation</a>, which acts as a knowledge store on which a new category of knowledge-based software can be built.</p> <p>During his keynote presentation, Nadella spoke about the intelligence that exists in the various IT systems used across the business. He said that businesses should be able to harness the intelligence that already exists enterprise-wide, starting with what he described as the “data underneath Microsoft 365”, which, according to Nadella, represents the people in the business, their relationship to coworkers, and work artefacts such as projects, calendars and communications data. “This is massive information,” he said, which can be used to bootstrap agentic AI projects.</p> <p>“Our goal is to have all of the innovation and the systems available in the token factory,” said Nadella. “That way you can build software which has the ability to use all of the capability [we provide] to train models and deliver models for inference.”</p> <p>In effect, he sees the Windows software developer ecosystem evolving to where it is now a Microsoft 365 ecosystem, where enterprise data is stored in AI-enabled office productivity tools such as Word, Excel, PowerPoint, Teams and Outlook, and these can be used as the foundation for a new generation of applications that can draw on these AI knowledge sources.</p> <p>It is this idea that all software will need to be knowledge-aware, which Huang spoke about during the company’s earnings call. “<a href="https://www.theserverside.com/tutorial/An-introduction-to-LLM-tokenization">Token generation</a> is at the centre of almost everything that relates to software in the future and relates to computing,” he said. “If you look at the way we use computing in the past, however, the amount of computation demand for software in the past is a tiny fraction of what is necessary in the future.”</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more AI hardware stories</h3> <ul class="default-list"> <li>Nvidia unveils <a href="https://www.computerweekly.com/news/366636948/Nvidia-unveils-Vera-Rubin-architecture-to-power-AI-agents">Vera Rubin architecture</a> to power AI agents: The AI chip giant has taken the wraps off its latest compute platform designed for test-time scaling and reasoning models, alongside a slew of open source models for robotics and autonomous driving.</li> <li><a href="https://www.techtarget.com/searchcloudcomputing/news/366637986/Microsoft-Maia-200-AI-chip-could-boost-cloud-GPU-supply">Microsoft Maia 200 AI chip</a> could boost cloud GPU supply: Industry watchers predict ancillary effects for enterprise cloud buyers from Microsoft's AI accelerator launch this week, from GPU availability to Nvidia disruption.</li> </ul> </div> </div> <p>According to Huang, the amount of computation necessary to run AI is 1,000 times higher than the computing power needed to run non-AI software. “The computing demand is just a lot higher,” he said. “And so, if we continue to believe there’s value in it, then the world will invest to produce that token.”</p> <p>When asked whether Nvidia is confident that its customers will continue to have the ability to spend more on AI infrastructure, which could impact Nvidia’s ability to grow, Huang spoke about the opportunity in enterprises to make use of agentic AI and its widespread usefulness across organisations.</p> <p>“We have now seen the inflection of agentic AI, and the usefulness of agents across the world and enterprises everywhere,” he said. “You’re seeing incredible compute demand because of it. In this new world of AI, compute is revenues. Without compute, there’s no way to generate tokens. Without tokens, there’s no way to grow revenues.”</p> <p>At least, that is how he positioned AI for the investment bank analysts on the earnings call. The company posted fourth quarter revenue of $68bn, up 73% year-over-year. Datacentre revenue increased by 75% to $62bn, which Nvidia said was being driven by demand for its Blackwell architecture and AI inference deployments. It also reported networking revenue of $1bn, up 3.5x year-over-year, fuelled by adoption of NVLink, Spectrum X and other Nvidia ethernet technologies.</p> <p>Last year, during his keynote presentation at the GTC conference in the US, Huang claimed that the lowest cost per token was being achieved using the most expensive GPU – which at the time was the Grace Blackwell NVLink 72.</p> <p>Nvidia describes the GB200 Grace Blackwell as a “superchip”, which connects two high-performance Nvidia Blackwell Tensor Core GPUs and the Nvidia Grace CPU with the NVLink-Chip-to-Chip (C2C) interface, capable of delivering 900 GBytes/s of bidirectional bandwidth.</p> <p>Significantly, the architecture means that applications have coherent access to a unified memory space. According to Nvidia, this simplifies programming and supports the larger memory needs of trillion-parameter LLMs, transformer models for multimodal tasks, models for large-scale simulations, and generative models for 3D data.</p> <section class="section main-article-chapter" data-menu-title="‘Huang’s Law’"> <h2 class="section-title"><i class="icon" data-icon="1"></i>‘Huang’s Law’</h2> <p>Some industry observers have coined the term “Huang’s Law” to describe his perspective of how each new generation of GPU delivers a 10x increase in performance, compared with Moore’s Law’s doubling of performance every 18 months.</p> <p>Nadella and Huang both spoke about how newer hardware is more energy-efficient at running AI workloads. During the Microsoft AI tour, Nadella noted that today’s system supports an entirely different memory hierarchy, which he said means “there’s now no latency with AI inference”.</p> <p>The messaging from both the Microsoft and Nvidia chiefs is that the best efficiency is achieved by taking advantage of the capabilities available in these new systems. “There’s an unbelievable renaissance happening with these systems and workloads, whether they’re training workloads or inference workloads, they are unlike anything we’ve seen in the past,” said Nadella.</p> <p>The tech sector is dead set on getting enterprises to adopt more and more AI. It is being built into knowledge-aware enterprise software likely to draw on the capabilities available in the newest generation of AI acceleration hardware.</p> <p>Clearly, the business models of Microsoft and Nvidia are tied to increased demand for AI. But it is also apparent that the cost of deploying advanced AI systems is not going to get any cheaper. If anything, capital expenditure on datacentres will continue to increase at a phenomenal rate, fuelled by demand for these new AI systems and the AI acceleration hardware they need.</p> </section> Looking at Nvidia’s latest financial results, it would seem that spending on compute is set to increase tenfold by 2030 https://cdn.ttgtmedia.com/visuals/German/article/GPU-datacentre-2-adobe.jpg https://www.computerweekly.com/news/366639773/Is-there-no-stopping-the-AI-spending-spree Wed, 04 Mar 2026 07:45:00 GMT <p>One of the most notable cloud technology trends in 2025 was the (seemingly) overnight emergence of the neocloud category of cloud providers, which specialise in the provision of niche, sovereign cloud and artificial intelligence (AI) infrastructure services.</p> <p>Neocloud providers, which include the likes of Nscale, CoreWeave and Carbon3.ai, are having a somewhat disruptive impact on the market by making huge commitments to build out hyperscale datacentres in support of the <a href="https://www.computerweekly.com/news/366628066/The-UK-governments-AI-Growth-Zones-strategy-Everything-you-need-to-know">UK government’s AI growth agenda</a>.</p> <p>These providers are also taking up <a href="https://www.computerweekly.com/news/366634475/CBRE-charts-rise-of-neocloud-providers-within-European-colocation-market">capacity in colocation datacentres</a> that some of the hyperscale cloud giants previously committed to renting space in, before pulling out, as they seek to rapidly build their footprint in the UK, particularly. &nbsp;&nbsp;</p> <p>As reported by Computer Weekly, real estate consultancy CBRE pinpointed <a href="https://www.computerweekly.com/news/366634475/CBRE-charts-rise-of-neocloud-providers-within-European-colocation-market">lower hyperscaler demand for colocation capacity</a> in the first nine months of 2025. In the aggregate, future AI-ready datacentre capacity was contracted for a total of 414MW, versus 133MW in the comparable 2024 period.</p> <p>A chunk of that will be to neocloud providers offering purpose-built AI services, such as bare metal or <a href="https://www.techtarget.com/searchcloudcomputing/news/366637986/Microsoft-Maia-200-AI-chip-could-boost-cloud-GPU-supply">graphics processing units (GPUs)</a> as a service (GPUaaS) or inference with pay-as-you-go pricing. But should enterprises be betting on neocloud? With AI infrastructure investments underpinning a Gartner forecast that annual enterprise IT revenues will see a 10.8% surge from 2025 to reach $6.2tn (£4.5tn) by the end of 2026, few want to be left behind.</p> <p>Mark Boost, CEO at cloud provider <a href="https://www.computerweekly.com/news/366632179/Cloud-provider-publishes-tech-sovereignty-plan-for-UK">Civo</a>, thinks some may have reasonable concerns about neoclouds, despite – or even because of – the vast investments in train. “The problem is there is too much hype right now. And with neocloud, you’re having companies that may be well capitalised but still have little experience in running cloud services.”</p> <p>They might tick multiple financial boxes and successfully procure datacentre space or GPUs, but that might be their limit. They might not be able to offer a mature, wide ecosystem of products and services. That may or may not be fine, depending on what IT buyers need. Some may be building themselves up in this space, by going down an open source route, for example, but it can represent a risk for customers to consider.&nbsp;</p> <p>“Your hyperscalers, your CoreWeaves and so on, do have a more mature ecosystem. But then, for sovereign infrastructure, beyond them, you’re really limited for choice,” says Boost. “Only a few have some form of software stack. Others are scrambling around to do it. Of course, if you do just want to buy a few GPUs and nothing else, they can hand you the keys and you’re on your own.”</p> <section class="section main-article-chapter" data-menu-title="Support needs for AI workloads"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Support needs for AI workloads</h2> <p>Many enterprises need far more than that in terms of support, however, especially with the rise of <a href="https://www.techtarget.com/searchitoperations/answer/Battle-of-the-buzzwords-AIOps-vs-MLOps-square-up">AIOps and MLOps</a>. Most organisations looking to benefit from AI and machine learning (ML) need a partner that can supply the required level and cadence of support. “There’s a consultancy and professional services element to consider,” says Boost. “And sovereignty is becoming a bigger and bigger thing. People have been burned. They crave control.”&nbsp;</p> <p>In summary, organisations need transparency around how data will be managed, stored and priced. They need to tread carefully when choosing cloud providers.</p> <blockquote class="main-article-pullquote"> <div class="main-article-pullquote-inner"> <figure> Neoclouds can raise the same sovereignty questions as traditional clouds. Do you really control your data? </figure> <figcaption> <strong>Enrico Signoretti, Cubbit</strong> </figcaption> <i class="icon" data-icon="z"></i> </div> </blockquote> <p>Enrico Signoretti, vice-president of product and partnerships at cloud storage firm Cubbit, adds that many neoclouds are just specialised clouds, operated or using a tech stack that’s largely based overseas. “[This means] they can raise the same sovereignty questions as traditional cloud,” he says. “Do you really control your data?”</p> <p>For sovereign AI, you need “home-grown champions”. European countries need to scale and fund their own new AI factories. The viable path is architectures that keep data sovereignty next to the GPU through encryption and the right data orchestration and governance. Otherwise, an enterprise’s data, which is its most important asset, remains exposed to risks linked to extraterritorial laws, he says.</p> <p>Thomas King, chief technology officer of internet exchange DE-CIX, says neocloud providers have competed so far by offering cheap GPUs for AI training. Rapid innovation in AI servers travels hand-in-hand with depreciation, which is estimated to be three to five times faster than for traditional hardware.</p> <p>“Usually, they are a lot cheaper because they focus on AI workloads only. They are not general-purpose cloud providers,” he says.</p> <p>The risk to the customer partly depends on the risk of provider lock-in that restricts long-term agility. That said, modern IT infrastructures usually have a lot of virtualisation in place. Moving from one provider to another is a lot easier than it was 10 years ago, says King.</p> <p>Additionally, moving to AI inference workloads instead of training is likely to prove more profitable. Training can be done cheaply, where land and power are affordable and datacentres are easy to build. But when you’re doing more, you need quality connectivity.</p> <p>“When it’s about using the AI models, neoclouds supported very closely can provide inference with very low latency,” he says.&nbsp;“In this case, you are usually also in an environment where you’re not only going with one AI provider anyway. You need to find the right mix to serve your customers best.”</p> <p>In addition, organisations do not usually go out of business overnight, with many neocloud firms publicly traded, which means regular market announcements. Warning signs, such as not keeping up with new GPU versions, mean you could start migrating elsewhere.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more neocloud stories</h3> <ul class="default-list"> <li><a href="https://www.techtarget.com/searchcloudcomputing/feature/AI-will-heavily-influence-cloud-related-decisions">AI will heavily influence</a> cloud-related decisions in 2026: AI is set to become a defining force in shaping cloud-related decisions and strategies this year. See what the experts are saying about multicloud, neoclouds, data sovereignty and serverless.</li> <li>Considerations for ensuring a <a href="https://www.computerweekly.com/feature/Considerations-for-ensuring-a-minimum-viable-digital-sovereign-cloud">minimum viable digital sovereign cloud</a>: Growing geopolitical uncertainty means CIOs need to consider the role of digital sovereignty in ensuring IT-powered services remain on tap.</li> </ul> </div> </div> <p>“If you do your IT infrastructure right, and build in the risk that your neocloud provider might go out of business, it shouldn’t be too hard to move your infrastructure,” he says.</p> <p>With the European Union’s <a href="https://www.eu-cloud-ai-act.com/">proposed Cloud and AI Development Act</a>, which is set to come into effect this year, neocloud providers may be able to offer control of data processing locations and ensure jurisdiction-aware interconnection and data pathways, he adds.</p> </section> <section class="section main-article-chapter" data-menu-title="Expansion tipped to continue"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Expansion tipped to continue</h2> <p>Estimates by Synergy Research suggest the doubling of the neocloud sector in the past year could be followed by further expansion at 69% per year through to 2030. &nbsp;</p> <p>“AI is a killer application for edge computing,” DE-CIX’s King says. “You have complex AI models. [Applications] need to be close to the user, because working on doing the calculations on the AI model already takes time. You can’t spend a lot of time on the transmission of the data back and forth.”</p> <p>Traditional hyperscale providers are also moving in a similar direction because a new market is developing, even if not as fast, with the return on investment (ROI) not being realised as quickly as many had hoped.&nbsp;</p> <p>“There are a lot of pros, including high margins in the inference space,” says King. “Not everyone will survive. But, in the end, everybody is looking into how we can make use of AI, and we are still in the beginning.”</p> <p>Suresh Vasudevan, CEO of AI platform provider Clockwork.io, notes that datacentre lifecycles run to 10 or 15 years, while GPU technology depreciates in four to six years. However, long-term contracts with foundation model builders or hyperscalers may reduce any risk.</p> <p>In many cases, neoclouds can offer lower GPU pricing, more predictable access to high-end capacity in a supply-constrained market, and sometimes bare metal environments where enterprises can bring and tune their own software stack for higher utilisation. When GPU supply is tight, guaranteed access to capacity and cost control can outweigh ecosystem convenience – although integration friction and enterprise readiness requirements cannot be underestimated.</p> <p>“Ultimately, the choice comes down to workload profile and economics,” adds Vasudevan.</p> </section> <section class="section main-article-chapter" data-menu-title="Consider independent benchmarks"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Consider independent benchmarks</h2> <p>Every neocloud will describe itself as enterprise-grade, so look for measurable operating data on the infrastructure reliability, utilisation, power, cooling and the like. Consider independent benchmarks like ClusterMAX from SemiAnalysis for useful comparative transparency, Vasudevan urges. “Enterprises should press for hard numbers,” he says. “What is your measured cluster-level availability? How often do interruptions occur at 1,000-GPU scale? What does your SLA truly guarantee?”&nbsp;</p> <blockquote class="main-article-pullquote"> <div class="main-article-pullquote-inner"> <figure> Enterprises should press for hard numbers. What is your measured cluster-level availability? How often do interruptions occur at 1,000-GPU scale? What does your SLA truly guarantee? </figure> <figcaption> <strong>Suresh Vasudevan, Clockwork.io</strong> </figcaption> <i class="icon" data-icon="z"></i> </div> </blockquote> <p><a href="https://www.computerweekly.com/blog/Computer-Weekly-Editors-Blog/Why-Big-Tech-wants-you-to-shift-the-definition-of-good-enough">Four or five nines availability</a> is expected in traditional central processing unit (CPU) environments, he points out. However, large GPU clusters can experience multiple disruptive interruptions per day. Failures are part of operating at scale, but must be consistently and efficiently managed.&nbsp;“The second differentiator is diagnostics. When jobs slow down or fail, does the provider offer deep, actionable telemetry to isolate the problem quickly? Without strong observability, GPU hours are lost and ROI erodes,” says Vasudevan.</p> <p>Hyperscale won’t be going away. For organisations with multi-year cloud commitments and significant data gravity, there are financial and practical incentives to continue building within that environment. “Hyperscalers bring breadth. They offer a deeply integrated ecosystem of microservices – identity, databases, security, networking and observability – that already sits alongside an enterprise’s existing data estate,” adds Vasudevan. &nbsp;</p> <p><a href="http://com/insights/figures/european-data-centres-figures-q3-2025">CBRE’s dataset also recorded</a> notable activity in the Nordics, where there are lower-cost renewable energy options. Power requirements may have more influence on the lease structures than square footage, and CBRE has also noted that neoclouds are attracting more interest where there are fewer hyperscale availability zones.</p> <p>Kevin Restivo, director and head of datacentre research for Europe at CBRE, says that generally, colocation providers may be offering space to neoclouds under different terms than those offered to hyperscalers. “The deals we see in the market between neoclouds and datacentre providers are typically shorter in length,” he says. “And contract terms change depending upon the amount of capacity contracted.”</p> <p>Meanwhile, rent prices of late are sometimes well above inflation. So it can be worth paying a premium and having shorter-term deals, the pay-off being greater flexibility and ability to migrate, as well as speed to market. “Neoclouds are trying to build out their infrastructure,” he says. “They need their kit in datacentres, and they need to do it quickly. Capacity is, as I like to say, an increasingly precious commodity in Europe and worldwide.”</p> <p>Through 2026, the supply bottleneck for compute-intensive workloads looks confirmed, relative to the perceived demand for access to GPUs, he adds.&nbsp;</p> <p>Of course, if that demand does not eventuate, there may be a need for fewer providers down the track. For now, neocloud will continue to play a key role in the datacentre landscape, by virtue of the capacity in train – that is, under construction for cloud purposes.&nbsp;“The real question is what enterprises make of AI services,” says Restivo. “Because there is great anticipation about investment in AI services on the part of European enterprises.”</p> <p>For most enterprises to move forward and begin employing AI at scale, the markets are going to need to see more early adopters succeed, demonstrating benefits and productivity.&nbsp;</p> </section> Anything that expands quickly can attract bubble accusations, and the market for cloud services in an AI age doesn’t look to be different https://cdn.ttgtmedia.com/rms/onlineimages/cloud_g943065362.jpg https://www.computerweekly.com/feature/Weighing-up-the-enterprise-risks-of-neocloud-providers Mon, 02 Mar 2026 04:00:00 GMT <p>Studies have shown that private 5G deployment is set for continued steady growth over the next five years, largely driven by new network deployment opportunities for system integrators (SIs), but despite the growth, SI-led deployments consistently run into complexity and total cost of ownership issues, obstacles that a partnership between <a href="https://www.cloud-ran.ai/">CloudRAN.AI</a> and <a href="https://firecell.io/powering-industrial-autonomy-firecell-and-accelleran-merge-to-launch-a-full-stack-private-5g-platform-that-eliminates-connectivity-failures/">Firecell</a> aims to address.</p> <p>Founded in 2021, Firecell serves system integrators and enterprises in ports, factories, defence, logistics and critical infrastructure, and develops purpose-built <a href="https://www.computerweekly.com/news/366632402/Airbus-climbs-in-industrial-digitisation-with-private-5G-deployment">private 5G connectivity for industrial environments</a>. Its services integrate hardware-agnostic, European sovereign core networks including programmable RAN, AI-driven network intelligence and management systems.</p> <p><a href="https://firecell.io/powering-industrial-autonomy-firecell-and-accelleran-merge-to-launch-a-full-stack-private-5g-platform-that-eliminates-connectivity-failures/">Following its recent merger with Accelleran</a>, the company now operates across France, Belgium, the UK, Germany and Poland, with deployments in Europe, the US and Asia.</p> <p>CloudRAN.AI provides private 5G and radio access services designed for enterprise and industrial deployments. Its portfolio encompasses scalable radio access network (RAN) architecture for outdoor and complex indoor coverage with artificial intelligence (AI)-native network planning (Megrez) and mission-critical communications (MCX).</p> <p>By combining Firecell’s delivery model with CloudRAN.AI’s modular radio architectures, the two companies aim to make private 5G faster to deploy and more cost-competitive against Wi-Fi on sites where the economics have historically made it hard to justify. Moreover, they wish to expand the radio hardware options available to SIs, focusing on reducing total cost of ownership and on-site deployment time.</p> <p>The combined architecture, available immediately across Europe and North America, is also designed to meet the performance requirements of demanding enterprise and industrial environments.</p> <p>“Private 5G is moving from large flagship deployments toward a much higher volume of mid-size industrial and enterprise sites,” said Firecell CEO Claude Seyrat. “At that scale, total cost of ownership and deployment speed determine whether a project gets approved. This partnership gives system integrators the hardware economics and deployment simplicity to compete on those projects, backed by the performance that the most demanding use cases require.”</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about private 5G</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366636440/2025-a-transformative-year-for-private-connectivity">2025 a ‘transformative year’ for private connectivity</a>: Research finds standalone private 5G networks well-positioned to become predominant wireless connectivity medium for Industry 4.0 applications in manufacturing and process industries.</li> <li><a href="https://www.computerweekly.com/news/366638797/Private-LTE-5G-networks-reached-6500-deployments-in-2025">Private LTE/5G networks reached 6,500 deployments in 2025</a>: Analysis of private 5G market finds steadily growing market that is increasingly driven by organic demand from end users, with WAN and enterprise segments of near equal worth.</li> <li><a href="https://www.computerweekly.com/news/366637131/Hutchison-Ports-completes-private-5G-network-at-UK-hub">Hutchison Ports completes private 5G network at UK hub</a>: Leading UK port authority taps into global mobile network to support autonomous trucks and future digital innovation.</li> <li><a href="https://www.computerweekly.com/news/366636384/Port-of-Tyne-charts-successful-course-with-private-5G">Port of&nbsp;Tyne charts successful course with private 5G</a>: One of the UK’s most dynamic and key deep-sea gateways reports strong operational gains after several years using private 5G across its 620-acre estate using sensor-driven processes.</li> </ul> </div> </div> <p>The integrated portfolio covers two common deployment scenarios. For outdoor coverage, Firecell will offer CloudRAN.AI’s 10W and 40W All-in-One radios, which combine baseband and radio in a single unit, reducing components and expediting installation.</p> <p>For complex indoor environments such as hospitals, multi-storey facilities and compartmentalised buildings, the companies have a Pico Radio Unit Active DAS architecture, in which multiple units operate as a unified coverage layer to reduce handover issues and maintain connectivity as devices move through a building.</p> <p>The collaboration also supports multi-region roll-outs with the frequency band diversity required across European and North American markets. For any project that is said to require “strong” outdoor or complex indoor coverage, Firecell assured that it could deliver the CloudRAN.AI radio portfolio through its partner network.</p> <p>“This partnership makes it easier for system integrators to deliver private 5G at the speed and cost enterprise sites require,” said Michel Trudelle, partnership director at CloudRAN.AI.</p> <p>“Through Firecell’s ecosystem, system integrators can access CloudRAN.AI All-in-One radios for efficient outdoor coverage and Active DAS for complex indoor environments – so they can standardise designs, reduce on-site complexity and deploy repeatably with performance that holds up in real conditions,” he said.</p> Hot on the heels of the Accelleran merger, purpose-built private 5G connectivity company unveils radio portfolio integration to give system integrators faster, more affordable deployment route across industrial, enterprise sites https://cdn.ttgtmedia.com/visuals/ComputerWeekly/HeroImages/mobile-network-5g-phone-mast-adobe.jpg https://www.computerweekly.com/news/366639431/Firecell-CloudRANAI-collaborate-to-cut-cost-and-complexity-of-private-5G Thu, 26 Feb 2026 09:28:00 GMT <p>Physical artificial intelligence (AI) is set to be one of the largest growth areas as industry moves into a new era of automation and efficiency, and looking to address this potentially hot market, NTT Docomo and Keio University <a href="https://haptics-c.keio.ac.jp/">Haptics Research Centre</a> have announced they have successfully conducted a demonstration of high-precision remote robot operation over a commercial 5G network.</p> <p>The Haptics Research Centre is aiming to make RealHaptics technology widely and universally available to consumers and companies worldwide. RealHaptics is an advanced&nbsp;<a href="https://www.google.com/search?q=force+feedback+system&amp;rlz=1C1GCEA_enGB1155GB1156&amp;oq=RealHaptics+technology&amp;gs_lcrp=EgZjaHJvbWUyBggAEEUYOTIJCAEQIRgKGKAB0gEIMzY5M2owajeoAgCwAgA&amp;sourceid=chrome&amp;ie=UTF-8&amp;mstk=AUtExfBuHob68Vq8ug2VoGOh9Ivbq40SqxoAxGDFFmFi_9UI24PS3D8xiHWhQ5gc9LXDPEYL-hjO32mMtwB57UWOks51bCR8V_Ya7diF7ULzgGh7gnPgEVbxRTAKS6u9nOqgZHU&amp;csui=3&amp;ved=2ahUKEwi7xoS5xfSSAxUBQkEAHS3HLGMQgK4QegQIARAD">force feedback system</a> that transmits, records and reproduces the tactile sensation of touch in real time, allowing remote operators to feel the resistance, softness or texture of objects handled by a robotic device.</p> <p>As a research institution for RealHaptics technology, Keio University holds a core set of related patents. The research centre operates a council for RealHaptics technology in which private companies also participate. The centre works closely with companies to carry out joint research to explore new applications for the technology.</p> <p>In essence, the demo has seen the combination of Configured Grant, a network slicing scheduling method in which the base station pre-allocates communication resources to specific devices – with^{<a href="https://www.docomo.ne.jp/english/info/media_center/pr/2026/0225_00.html#notice02"></a>} Keio’s Real Haptics technology, invented by Kohei Onishi, transmitting tactile and contact information bidirectionally and reproducing human force on robots.</p> <p>To convey precise force feedback when a remote robot interacts with objects controlled by an operator at a distant location, mobile data communication must maintain low latency and minimal jitter. High or fluctuating latency can disrupt synchronisation between the operator and remote robot. This compromises the precise force reproduction of Real Haptics, hindering delicate robot operation.</p> <p>Drilling deeper into the technological breakthrough, mobile devices normally communicate with a base station using the Dynamic Grant method. When a device has data to transmit, it first requests communication resources from the base station. The base station then allocates resources based on the status of other devices and notifies the requesting device.</p> <p>The time between the resource request and actual data transmission, the scheduling delay, always occurs under Dynamic Grant, and can fluctuate depending on network conditions. This poses a major challenge for stable remote operation of advanced robots, especially in congested networks.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about robotics</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/feature/AI-XR-digital-twins-set-to-transform-robotics">AI, XR, digital twins set to transform robotics</a>: The availability of advanced sensors, artificial intelligence, digital twins, XR and robotics has changed technology-driven markets. We look at how the intersection of these technologies will create commercial opportunities.</li> <li><a href="https://www.computerweekly.com/feature/AI-and-digital-twins-to-serve-increasingly-complex-robot-management">AI and digital twins to serve increasingly complex robot management</a>: Proliferating fleets of robots, diffusion of humanoids, and the emerging robot class of androids will increase complexities of robot management in the future. AI and digital twins will become increasingly important.</li> <li><a href="https://www.computerweekly.com/news/366629305/China-Unicom-5G-A-network-powers-worlds-first-humanoid-robot-games">China Unicom 5G-A network powers world’s first humanoid robot games</a>: Leading Chinese telco offers robot showcase next-gen 5G network that supports embodied AI realising seamless connectivity for both spectators and robots even during peak usage.</li> <li><a href="https://www.computerweekly.com/news/366636818/CES-2026-Qualcomm-expands-IEIoT-portfolio">CES 2026: Qualcomm expands IE‑IoT portfolio</a>: Edge AI technology made available for developers, enterprises and OEMs, integrating chipsets, software distribution and tools to scales across verticals.</li> </ul> </div> </div> <p>With Configured Grant, the base station pre-allocates exclusive communication resources to a device or line for a specific period. The device can transmit data without requesting resources, effectively eliminating scheduling delay. As a result, both latency and jitters are reduced, enabling more stable wireless remote operation.</p> <p>In the experiment, Configured Grant was applied to the radio link between the terminal and base station to minimise latency and jitter. Results confirmed that latency requirements for practical teleoperation of Real Haptics robots were met, while force reproducibility and operability were improved.</p> <p>Delicate force feedback and tactile sensations were transmitted stably, marking the world’s first demonstrationof Configured Grant enabling practical robot teleoperation over a commercial <a href="https://www.computerweekly.com/news/366619933/MWC-2025-NTT-Docomo-preps-for-beyond-5G-future">5G network such as Docomo’s</a>. By reducing the impact of latency in the wireless segment, the two organisations say that highly precise and delicate remote robot control can now be performed stably even under network congestion.</p> <p>Docomo and Keio University have previously collaborated on <a href="https://www.docomo.ne.jp/corporate/technology/rd/6g/007/?icid=CRP_EN_info_media_center_pr_2026_0225_00_to_CRP_CORP_technology_rd_6g_007">the development and testing of Real Haptics for robot teleoperation over mobile networks</a>. Going forward, the two organisations say they will continue developing and testing technologies to accelerate the practical use of advanced robot teleoperation.</p> Japan's leading mobile operator and global research institute for haptics announce what they call world's first stable, high-fidelity robot via commercial 5G using low-latency slicing. https://cdn.ttgtmedia.com/visuals/ComputerWeekly/HeroImages/robot-humanoid-butterfly-evolution-Mykola-adobe.jpg https://www.computerweekly.com/news/366639389/DOCOMO-Keio-University-claim-5G-robot-teleoperation-first Wed, 25 Feb 2026 09:45:00 GMT <p>The use of artificial intelligence (AI) in networking is increasing rapidly, with the latest stage in this technological evolution seeing comms tech firm Nokia announce a collaboration with Amazon Web Services (AWS) to bring the first agentic AI-powered 5G-Advanced network slicing solution in a live 5G network.</p> <p>The intent-based 5G slicing innovation combines <a href="https://www.computerweekly.com/news/366616411/Nokia-Du-deliver-successful-transport-network-slicing-trial">Nokia’s network slicing</a> with AWS AI platform technologies to empower comms providers in delivering premium services precisely where and when they are needed.</p> <p>Outlining the challenge that they believe is facing the industry, Nokia and AWS said that providers can face challenges optimising network performance during unpredictable events such as traffic surges, emergencies or <a href="https://www.computerweekly.com/blog/CW-Developer-Network/Palo-Alto-Networks-Global-incident-report-2026-analysis">mass gatherings</a>, which can result in suboptimal service quality and inefficient resource utilisation.</p> <p>Hosted by <a href="https://www.computerweekly.com/news/366639085/Ericsson-strikes-Microsoft-AWS-deals-for-enterprise-5G">AWS</a>, the autonomous network slicing intelligence dynamically adapts and manages even the most challenging traffic conditions across varied geographical areas.&nbsp;The agentic AI-powered approach is said to be the key to unlocking “significant” customer value across diverse applications and use cases by creating premium services that respond intelligently to dynamic conditions, ensuring optimal performance precisely where and when customers need it.&nbsp;</p> <p>The end-to-end advanced network slicing technology across RAN-transport-core especially utilises Nokia’s 5G AirScale base station, MantaRay SMO and agentic AI modules, which are seamlessly integrated with the Amazon Bedrock artificial intelligence platform.</p> <p>The agentic AI modules operate in multiple modes: chatbot, on-demand, scheduled and autonomous. All modules interact with Amazon Bedrock via APIs. Furthermore, applications and use cases powered by agentic AIs are enhanced with Nokia’s Edge Slicing solution, bringing cloud applications and workloads directly to mobile users and devices over high-capacity, secure and low-latency networks.</p> <p>Nokia noted that its AI slicing technology uses agentic AI to analyse real-world internet data, including locations, events, traffic, incidents and maps to deliver adaptive network slicing. Cited typical use cases for the agentic AI-powered 5G-Advanced network slicing include intent-based enterprise and industrial slicing, on-demand slicing with agentic AI, and agentic AI for mass events.</p> <p>As regards to the former, the intent-based enterprise and industrial slicing technology will see use in measuring live network KPIs such as bit rate and latency, and autonomously adjusts RAN policies to meet enterprise SLAs across campuses, business parks and city areas. This is intended to enhance premium slicing services for <a href="https://www.computerweekly.com/news/366629995/Private-mobile-network-deployments-return-to-growth">critical applications in manufacturing</a>, the internet of things (IoT), drones, smart cities, hospitals, energy transportation <a href="https://www.computerweekly.com/news/366637131/Hutchison-Ports-completes-private-5G-network-at-UK-hub">and ports</a>.</p> <p>On-demand slicing with agentic AI&nbsp;also boosts network performance for selected 5G base stations. When activated by external data, Nokia believes that this service will provide first responders and public safety authorities with better network connectivity during emergencies. On-demand network slicing with agentic-AI is said to preserve quality of service for premium 5G+/<a href="https://www.computerweekly.com/news/366628095/L4S-adoption-sees-T-Mobile-writing-next-chapter-of-5G">5G Advanced</a> and <a href="https://www.computerweekly.com/news/365531497/Nokia-advances-5G-network-slicing-FWA-across-Europe-and-Asia">FWA</a> customers using gaming, streaming, <a href="https://www.computerweekly.com/news/366619827/MWC-2025-Comms-tech-consortium-begins-XR-trials-on-5G-Standalone-network">XR</a> and AI applications in response to major traffic surges, weather conditions and environmental changes.</p> <p>Nokia added that agentic AI for mass events will deliver much broader capacity availability during high-demand moments such as <a href="https://www.computerweekly.com/news/366639043/Silverstone-takes-mobile-connectivity-to-full-throttle">concerts and sporting events</a>. AI analyses network data, infers patterns and sets slicing policies for scheduled events, optimising premium 5G slicing for VIP spectators, payment applications, fan engagement, video broadcasting and operational crews in arenas, parks and conference centres.</p> <p>Operators Du and Orange are the first to explore the innovation in their respective networks. In November 2024, Nokia and United Arab Emirates (UAE) <a href="https://www.computerweekly.com/news/366616411/Nokia-Du-deliver-successful-transport-network-slicing-trial">telecoms and digital services provider du</a> carried out a transport network slicing trial to gain benefits of dedicated infrastructure and reduce overall energy consumption.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about network slicing and 5G</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366634793/5G-Standalone-growth-spurs-differentiated-connectivity-services">5G Standalone growth spurs differentiated connectivity services</a>: Mobility Report shows 33 CSPs currently offer differentiated connectivity services based on network slicing, with a combined total of 65 offerings with around 1.4 billion people expected to be served by FWA.</li> <li><a href="https://www.computerweekly.com/news/366637291/Three-Sweden-launches-commercial-5G-SA-network">Three Sweden launches commercial 5G SA network</a>:&nbsp; Leading Swedish operator implements end-to-end 5G standalone network with focus on enterprises in urban areas and the fixed wireless access broadband market.</li> <li><a href="https://www.computerweekly.com/news/366618259/Singtels-5G-network-slicing-to-boost-Singapores-defence-and-security">Singtel’s 5G network slicing to boost Singapore’s defence and security</a>: Singapore’s defence and security technology agencies will leverage Singtel’s 5G network slicing technology nationwide to combat evolving security threats and enhance national security.</li> <li><a href="https://www.computerweekly.com/news/366617971/BT-unveils-5G-Standalone-network-slicing-deployment-first-in-Belfast">BT unveils 5G Standalone network slicing deployment first in Belfast</a>: UK’s leading telco announces first real-world deployment of 5G Standalone network slicing capabilities in a slicing trial that brought payments to Belfast Christmas Market in December 2024.</li> </ul> </div> </div> ‘Industry-first’ intent-based 5G-Advanced slicing with agentic AI to offer providers with premium network slicing services that respond to real-world situations and enable autonomous intelligence https://cdn.ttgtmedia.com/visuals/ComputerWeekly/Hero%20Images/network-abstract-1-adobe.jpeg https://www.computerweekly.com/news/366639478/Nokia-AWS-demo-agentic-AI-network-slicing-with-du-Orange Wed, 25 Feb 2026 07:52:00 GMT <p>Looking to transcend the physical constraint of every generation of connectivity to date – copper’s speed, fibre’s time to deploy and the scarcity of radio spectrum – Taara has revealed a “breakthrough approach” to commercial communications and connectivity infrastructure in the form of the “world’s first” wireless communication platform based on optical phased arrays.</p> <p>Designed for operators, enterprises and next-generation data infrastructure, Taara Beam is attributed with bringing <a href="https://www.computerweekly.com/news/366637204/UK-broadband-revolution-shows-no-sign-of-slowing-down">fibre-like speeds</a> to environments where traditional infrastructure is too slow, costly or impractical to build, marking a shift from fixed, physical networks to infrastructure that can evolve at the pace of demand.</p> <p>The new developments build on existing Taara work using beams of light to extend high-speed internet to places where traditional infrastructure is difficult to deploy. Its first system Taara Lightbridge is now deployed in more than 20 countries, with operators including <a href="https://www.computerweekly.com/news/366628320/Airtel-to-sell-its-in-house-tech-globally-inks-deal-with-Singtel">Airtel</a>, Digicel, T-Mobile, SoftBank and Liquid.</p> <p>Taara Beam is described as being designed for the next phase, shrinking Taara’s wireless optics technology into a form factor around the size of a shoe-box to “radically” increase network density and flexibility. It is intended to see use in enabling high-throughput, low-latency connectivity across urban environments, enterprise campuses, datacentre clusters and event venues without the delays and costs associated with building physical infrastructure.</p> <p>Taara sees Beam as being deployed on rooftops, poles or existing structures within hours, forming high-bandwidth mesh networks that support applications ranging from small-cell backhaul mounted on street furniture to fronthaul networks and AI-driven, real-time systems. By operating in the unlicensed optical spectrum, it avoids congestion and recurring spectrum costs while delivering performance at the speed modern networks require.</p> <p>Explaining the rationale for the launch and the fundamental technology foundations, <a href="https://www.taaraconnect.com/">Google’s Moonshot Factory</a> said that by moving the core functionality of high-speed wireless optical communication into an integrated circuit that controls light electronically, comms networks that can be deployed quickly, scaled more flexibly and improved over time, without the constraints of trenching fibre or securing scarce spectrum.</p> <p>The proprietary optical phased arrays were developed at X and Taara labs over the past several years. The first product built on the photonic platform will be Taara Beam, engineered to deliver up to 25 Gbps of high-speed, low latency connectivity over distances up to 10 kilometres in a compact, deployable form factor.</p> <p>Traditional free-space optical systems steer beams of light using mirrors, sensors and mechanical hardware. According to Taara, this is an approach that works but is physically constrained at scale. Taara Beam is said to represent a new architecture, shifting from mechanical control to increasingly solid-state control of light.</p> <p>At Taara Beam’s core is an integrated photonic module containing over a thousand miniature light emitters arranged in an optical phased array, a solid-state steering device. This phased array allows the platform to track, shape and steer light with greater precision, improving reliability and latency while significantly reducing size and mechanical complexity.</p> <p>“With light transmitted through the air, those constraints begin to disappear. Taara Beam is the first commercial product built on our photonics platform, and it’s just the beginning,” said Taara founder and CEO Mahesh Krishnaswamy. “We’re not just improving networks, we’re removing the limits that have defined them. We’re…building toward a future where connectivity feels less like infrastructure and more like the air we breathe – essential, abundant and almost invisible to the people who rely on it.”</p> <p>Devin Brinkley, senior vice-president of engineering at Taara, added: “Silicon photonics allows us to integrate the core functionalities of wireless optical communication into a single module. We’ve compressed most of the functionality of our previous systems into a photonic module the size of a finger. As the technology matures, it can scale across performance, cost and size – similar to the exponential pace at which semiconductor platforms evolve.”</p> <p>Taara Beam will make its official industry debut at the forthcoming <a href="https://www.mwcbarcelona.com/">Mobile World Congress (MWC) Barcelona 2026</a>. Operators, infrastructure providers and partners can now request early <a href="https://www.taaraconnect.com/">access to the technology</a>.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about optical communications</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/news/366639275/Cisco-Qunnect-claim-quantum-first-with-datacentre-connectivity">Cisco, Qunnect claim quantum first with datacentre connectivity</a>: IT and networking giant collaborates with scalable quantum networks firm for what is said to be a successful demonstration of quantum networking connecting a datacentre to two research facilities in New York City.</li> <li><a href="https://www.computerweekly.com/news/366635233/UK-partnership-extends-fibre-optic-tech-for-more-reliable-radio-comms">UK partnership extends fibre optic tech for more reliable radio comms</a>: Knowledge transfer partnership aims to make ‘radio-over-fibre’ technology available to new sectors, developing ways to amplify and extend radio signals through fibre optics for fewer delays and stronger signals.</li> <li><a href="https://www.computerweekly.com/feature/IOWN-Shining-light-on-the-future-of-communications">Shining light on the future of communications</a>: With rising demand for data and energy consumption due to the vast compute power required by applications such as AI and large language models, something needs to change in networks – this could be IOWN.</li> <li><a href="https://www.computerweekly.com/news/252529265/Optical-lights-up-nascent-space-communications-applications-market">Optical lights up nascent space communications applications market</a>: Research highlights potential of what it calls “exciting and rapidly developing” global market for optical communications terminals driven by NGSO broadband constellations.&nbsp;</li> </ul> </div> </div> Google Moonshot company unveils way to transmit ultra-high-speed internet through the air using light, shrinking core technology onto a single chip to deliver fibre-like speeds without cables or spectrum https://cdn.ttgtmedia.com/visuals/ComputerWeekly/HeroImages/energy-power-electricity-abstract-flashmovie-adobe.jpg https://www.computerweekly.com/news/366639454/Taara-unveils-photonics-platform-for-wireless-comms Tue, 24 Feb 2026 10:10:00 GMT <p>ChatGPT, launched in 2022, began making a significant impact on the market by late 2023, according to Synergy Research Group. The company’s chief analyst, John Dinsdale, points out that cloud market leaders have experienced accelerated revenue growth over time. Additionally, the emergence of numerous <a href="https://www.computerweekly.com/news/366634475/CBRE-charts-rise-of-neocloud-providers-within-European-colocation-market">neocloud companies</a> (<em>see box: </em><a href="http://preview.pg.techtarget.com:8080/ComputerWeekly/feature/Neoclouds-Meeting-demand-for-AI-acceleration?_dc=1770899156939&amp;vgnextrefresh=1"><em>What is a neocloud?</em></a>) has further strengthened the already positive momentum in the market.</p> <p>This sentiment is reflected in the <em><a href="https://www.weforum.org/publications/rethinking-ai-sovereignty/">Rethinking AI sovereignty</a></em> whitepaper, published to coincide with the <a href="https://www.computerweekly.com/news/366637503/Davos-2026-Smart-thinking-needed-for-sovereign-AI-investment">World Economic Forum</a>, which notes that surging demand for compute is spawning new AI infrastructure development models, such as neocloud providers, national cloud providers and industry-specific artificial intelligence (AI) clouds. While hyperscalers offer global reach and full-service cloud ecosystems, neoclouds provide specialised, high-performance compute infrastructure tailored to AI training and deployment.</p> <p>This surge in demand for AI acceleration has seen a surprising benefactor. According to <a href="https://reports.tiger-research.com/p/mining-eng">Tiger Research</a>, cryptocurrency mining firms, seeking to reduce their exposure to bitcoin’s volatile pricing, are redirecting their graphics processing unit (GPU) farms toward AI acceleration applications.</p> <p>One example is the Australian bitcoin mining company, Iris Energy. In 2021/2022, Neel Khokhani, a Dubai-based fund manager, acquired shares in the small Australian datacentre for $1 per share. By assisting the company in leveraging its substantial physical assets to transition into an AI infrastructure provider, the share price surged to $63 by 2026. This transformation led to a $60m increase in the company’s valuation, which is now operating under the name Iren.</p> <section class="section main-article-chapter" data-menu-title="More choice"> <h2 class="section-title"><i class="icon" data-icon="1"></i>More choice</h2> <p>Before the emergence of neoclouds a few years ago, if an organisation wanted to work with AI, it had no choice but to go to a hyperscaler like Amazon Web Services (AWS) or Google. While the hyperscalers offer AI infrastructure as part of their vast public cloud services portfolio, Roy Illsley, chief analyst at Omdia, says the hyperscalers tend to be expensive and, as he recalls, a few years ago, there was very little choice other than Google’s AI offerings.&nbsp;</p> <p>Analyst firm Gartner estimates that by 2030, neocloud providers will capture around 20% of the $267bn AI cloud market. Neoclouds are purpose-built cloud providers designed for GPU-intensive AI workloads. They are not a replacement for hyperscalers, but a structural correction to how AI infrastructure is built, bought and consumed. Their rise signals a deeper shift in the cloud market: AI workloads are forcing infrastructure to unbundle again.</p> <p>In a <a href="https://www.computerweekly.com/opinion/Gartner-Why-neoclouds-are-the-future-of-GPU-as-a-Service">recent Computer Weekly article</a>, Mike Dorosh, a senior director analyst at Gartner, said IT buyers face three interrelated constraints, which influence their AI infrastructure decisions. First, there is what Dorosh calls cost opacity, which he says is rising as GPU pricing becomes increasingly bundled and variable, often inflated by overprovisioning and long reservation commitments that assume steady-state usage. Then there are supply bottlenecks, which he says constrain access to advanced AI accelerators. This results in long lead times, regional shortages and limited visibility into future availability. For Dorosh, the third area of concern for IT buyers is performance trade-offs, where virtualisation layers and shared tenancy reduce predictability for latency-sensitive training and inference workloads.&nbsp;</p> <p>According to Dorosh, these pressures are no longer marginal. They create a market opening that neoclouds are designed to fill.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading"><a id="dig-sov"></a>The role of neoclouds in digital sovereignty</h3> <p>Along with artificial intelligence (AI) acceleration, neocloud providers are targeting demand for <a href="https://www.computerweekly.com/ezine/Computer-Weekly/Tech-nationalism-The-need-to-build-and-protect-UK-digital-sovereignty">digital sovereignty</a>. From an IT derisking strategy perspective, this means organisations might feel they are relying too heavily on the platform and services of a hyperscaler.</p> <p>In a recent video interview, Gartner senior director analyst Rene Buest told Computer Weekly the analyst firm is having more client conversations where IT leaders are seeking ways to diversify their cloud strategy. As such, Gartner is receiving more enquiries about sovereign clouds – local infrastructure alternatives to the hyperscalers.</p> <p>“Throughout 2025, I would say 90% of my Gartner enquiries with end-user customers were only about the topic of digital sovereignty. Their concerns have increased because they don’t know what they should do or how the world will look tomorrow. They just wanted to balance the risks,” he said.</p> <div class="youtube-iframe-container"> <iframe id="ytplayer-0" src="https://www.youtube.com/embed/S9k_os80Nm8?autoplay=0&amp;modestbranding=1&amp;rel=0&amp;widget_referrer=null&amp;enablejsapi=1&amp;origin=https://www.computerweekly.com" type="text/html" height="360" width="640" frameborder="0"></iframe> </div> <p>Buest said IT buyers are evaluating other types of cloud providers that can offer a higher level of sovereignty, or a level of sovereignty that their preferred hyperscaler cannot provide. And this aligns with the need to build out local and sovereign AI capabilities.</p> </div> </div> <p>One example of a neocloud provider is <a href="https://www.computerweekly.com/news/366637748/Scotland-gets-AI-growth-zone-boost-in-Lanarkshire">CoreWeave,</a> which the authors of the <em>Rethinking AI sovereignty</em> report say is undergoing a capacity expansion, having secured funding of $25bn since 2024. AI infrastructure buildout is also expanding through national cloud providers such as Humain (Saudi Arabia), G42 (United Arab Emirates), Outscale (France) and StackIT (Germany).</p> <p>Another neocloud company that has been making headlines is <a href="https://www.computerweekly.com/feature/Nscale-explained-Everything-you-need-to-know">Nscale</a>, which has committed to delivering approximately 12,600 Nvidia GB300 GPUs at the Start Campus datacentre in Sines, Portugal, in the first quarter of 2026. This multi-year agreement sees Nscale offering Nvidia AI infrastructure services to Microsoft while providing European customers with sovereign AI within the European Union.</p> <p>This deal builds on <a href="https://www.computerweekly.com/news/366631259/Microsoft-to-invest-30bn-in-expanding-its-operations-and-AI-infrastructure-footprint-in-UK">plans announced by Nscale and Microsoft</a> in September 2025 to deliver the UK’s largest Nvidia AI supercomputer at Nscale’s Loughton AI Campus. The 50MW facility, scalable to 90MW, is expected to house approximately 23,000 Nvidia GB300 GPUs from the first quarter of 2027 to power Microsoft Azure services.</p> <p>Gartner’s <a href="https://www.gartner.com/en/documents/7129230"><em>Neoclouds: The next offering arrow in the service provider quiver</em></a> report notes that the consumption-based economics and transparent pricing offered by neocloud providers address the overprovisioning and hidden costs often associated with the offerings from hyperscalers. In fact, Gartner reports that by offering transparent, usage-based billing, IT buyers can expect to see cost savings of 60-70% on GPU instances compared with hyperscalers.</p> <p>However, Dorosh says the more significant change is architectural rather than financial. Neoclouds encourage organisations to make explicit decisions about AI workload placement. Training, fine-tuning, inference, simulation and agent execution each have distinct performance, cost and locality requirements. Treating them as interchangeable cloud workloads is increasingly inefficient and often unnecessarily expensive.</p> <p>As a result, AI infrastructure strategies are becoming inherently hybrid and multicloud by design – not as a by-product of supplier sprawl, but as a deliberate response to workload reality. The cloud market is fragmenting along functional lines, and neoclouds occupy a clear and growing role within that landscape.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more about neoclouds</h3> <ul class="default-list"> <li><a href="https://www.computerweekly.com/feature/Sovereign-cloud-and-AI-services-tipped-for-take-off-in-2026">Sovereign cloud and AI services</a> tipped for take-off in 2026: Digital sovereignty is set to become a top investment priority in 2026, due to geopolitical and legislative changes.</li> <li><a href="https://www.computerweekly.com/news/366632731/Revenue-generated-by-neoclouds-expected-to-exceed-23bn-in-2025-predicts-Synergy">Synergy – revenue generated by neoclouds</a> expected to exceed $23bn in 2025: IT market watcher Synergy Research Group is predicting big things for the neocloud market through 2025 and beyond.</li> </ul> </div> </div> <p>“Neoclouds started as GPU as a service. If you needed GPUs, these companies bought or leased GPUs from Nvidia, and then they would slice them and sell them off to people in smaller groups and bundles,” says Omdia’s Illsley.</p> <p>However, over time, neocloud providers have added software stacks and developed other services to meet the demand of IT buyers who need GPU power and the software stack required for AI training or AI inferencing.</p> <p>Getting started on deploying AI workloads for inference or training is arguably not as simple as the one-click option offered on something like the AWS Marketplace, Illsley says the neocloud providers are maturing to a point where they have partnered with AI software providers and can therefore offer a full set of services to meet the requirements of IT buyers who need AI compute capacity. “They are saying that they have GPUs and now provide access through partnerships to the software to run AI workloads,” he says.</p> <p>As an example, CoreWeave and Nvidia recently expanded their relationship to accelerate CoreWeave’s build-out of more than 5GW of AI factory capacity by 2030. Along with the hardware commitment, according to a market insight report from Macquarie Group, the agreement shows that CoreWeave is also working with Nvidia to incorporate its AI-native software within Nvidia’s reference architectures for Nvidia’s enterprise clients and cloud partners.</p> <p>One neocloud benefit identified by Gartner is access for IT buyers to specialised hardware, since neoclouds tend to prioritise cutting-edge GPUs, often securing first-to-market access through strategic partnerships. They also cater to bare-metal performance and optimised networking, since neoclouds are able to eliminate the layers of server virtualisation needed in multi-tenanted hyperscaler installations. Instead, they are able to offer direct hardware access, which Gartner says reduces latency and makes it possible to deploy high-bandwidth connectivity such as NVLink and InfiniBand for optimal GPU-to-GPU communication.</p> </section> <section class="section main-article-chapter" data-menu-title="Choosing between a neocloud and a hyperscaler"> <h2 class="section-title"><i class="icon" data-icon="1"></i>Choosing between a neocloud and a hyperscaler</h2> <p>While they may have begun as GPU-as-a-service type offerings, the evolution of neoclouds means there is now less of a gap between their AI services and the full-blown AI platform offerings from the hyperscalers.</p> <p>Clearly, hyperscalers will eventually offer more attractive pricing to compete with neoclouds, but as Gartner senior director analyst Rene Buest points out, neocloud providers are trying to deliver more predictable pricing.</p> <p>“Hyperscalers are very transparent in terms of their pricing models, so pay as you go, but at the end of the month, you don’t really know what you will pay,” he says. In other words, when using hyperscaler IT infrastructure, the monthly cost of compute resources consumed cannot be determined in advance.&nbsp;</p> <p>IT leaders can benefit, at least in Buest’s view, from 70% cost savings by choosing a neocloud over a hyperscaler. “They also provide instant direct access to advanced GPUs, which tend to outpace the hyperscalers in speed and transparency,” he says.</p> <p>Buest says neoclouds are very niche, “providing purpose-built infrastructure for AI workloads”. This not only meets customer demand today, but also suggests that neoclouds will be viable in the foreseeable future.</p> <p>Khokhani’s successful investment in the former bitcoin miner Iris Energy, now known as Iren, suggests that the long-term AI capacity contracts secured by neocloud providers indicate a stable and robust business model.</p> <p>He says: “People still think of Iren through a bitcoin-mining lens, but that misses what the business has become. What attracted me was the transition to long-dated, contracted datacentre infrastructure. When you have multi-year take-or-pay style contracts with an investment-grade counterparty like Microsoft, the economic risk starts to resemble infrastructure credit rather than crypto volatility.”</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading"><a id="neo-def"></a>What is a neocloud?</h3> <p>A neocloud refers to a modern, next-generation cloud computing model that builds on traditional cloud infrastructure by incorporating advanced technologies, innovative architectures and enhanced capabilities.</p> <p>The term is not a formal industry standard, but is often used to describe cloud-based IT infrastructure that goes beyond the conventional public, private or hybrid cloud models. Neoclouds are designed to address the evolving needs of businesses, particularly in areas like scalability, flexibility and performance.</p> <ul class="default-list"> <li><strong>Advanced automation</strong>: Utilises AI and machine learning for process optimisation and resource management.</li> <li><strong>Edge computing</strong>: Processes data closer to the source for reduced latency and faster responses.</li> <li><strong>Multicloud and hybrid support</strong>: Integrates with multiple cloud providers and on-premise systems.</li> <li><strong>AI and data-driven</strong>: Optimised for AI workloads, big data analytics and machine learning applications.</li> <li><strong>Serverless computing</strong>: Enables developers to focus on applications without managing infrastructure.</li> <li><strong>Sustainability</strong>: Emphasises energy efficiency and green datacentres.</li> <li><strong>Enhanced security</strong>: Incorporates zero-trust architectures and real-time threat detection.</li> </ul> </div> </div> </section> We look at how neoclouds can deliver access to artificial intelligence acceleration faster and cheaper than public cloud providers https://cdn.ttgtmedia.com/visuals/ComputerWeekly/Hero%20Images/technology-digital-ai-robot-adobe.jpeg https://www.computerweekly.com/feature/Neoclouds-Meeting-demand-for-AI-acceleration Mon, 23 Feb 2026 04:00:00 GMT <div></div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{11}" paraid="453044816"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">For the past decade,&nbsp;</span><span data-ccp-parastyle="No Spacing">hyperscalers</span><span data-ccp-parastyle="No Spacing">&nbsp;have defined how</span><span data-ccp-parastyle="No Spacing">&nbsp;CIOs and IT leaders</span><span data-ccp-parastyle="No Spacing">&nbsp;think about</span><span data-ccp-parastyle="No Spacing">&nbsp;their&nbsp;</span><span data-ccp-parastyle="No Spacing">organi</span><span data-ccp-parastyle="No Spacing">s</span><span data-ccp-parastyle="No Spacing">ation’s</span><span data-ccp-parastyle="No Spacing">&nbsp;cloud infrastructure. Scale, abstraction and convenience became the default answers to&nbsp;</span><span data-ccp-parastyle="No Spacing">almost every</span><span data-ccp-parastyle="No Spacing">&nbsp;</span><span data-ccp-parastyle="No Spacing">compute</span><span data-ccp-parastyle="No Spacing">&nbsp;question.&nbsp;</span></span><span xml:lang="EN-GB" data-contrast="auto"><span data-ccp-parastyle="No Spacing">But </span></span><a rel="noreferrer noopener" target="_blank" href="https://www.gartner.com/en/ai"><span xml:lang="EN-GB" data-contrast="none"><span data-ccp-charstyle="Hyperlink">a</span><span data-ccp-charstyle="Hyperlink">rtificial intelligence</span></span></a><span xml:lang="EN-GB" data-contrast="auto"><span data-ccp-parastyle="No Spacing"> </span></span><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">(AI</span><span data-ccp-parastyle="No Spacing">)</span><span data-ccp-parastyle="No Spacing">&nbsp;is breaking the economics of cloud&nbsp;</span><span data-ccp-parastyle="No Spacing">computing</span><span data-ccp-parastyle="No Spacing">&nbsp;and&nbsp;</span><span data-ccp-parastyle="No Spacing">neoclouds</span><span data-ccp-parastyle="No Spacing">&nbsp;are&nbsp;</span><span data-ccp-parastyle="No Spacing">emerging</span><span data-ccp-parastyle="No Spacing">&nbsp;as the response.&nbsp;</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></p> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{13}" paraid="1656553237"><span xml:lang="EN-US" data-contrast="auto">Gartner estimates that by 2030,&nbsp;neocloud&nbsp;providers will capture around 20% of the $267bn AI cloud market.</span><span xml:lang="EN-US" data-contrast="auto">&nbsp;Neoclouds&nbsp;are purpose-built cloud providers designed for</span><span xml:lang="EN-GB" data-contrast="auto">&nbsp;</span><span xml:lang="EN-US" data-contrast="auto">graphics processing unit&nbsp;(GPU)-intensive AI workloads. They are not a replacement for&nbsp;hyperscalers, but a structural correction to how AI infrastructure is built,&nbsp;bought&nbsp;and consumed. Their rise signals a deeper shift in the cloud market: AI workloads are forcing infrastructure to unbundle again.</span><span data-ccp-props="{&quot;335559738&quot;:240,&quot;335559739&quot;:240}">&nbsp;</span></p> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{15}" paraid="1924758627"><span xml:lang="EN-US" data-contrast="auto">This is not a return to on-premises thinking, nor a rejection of the cloud operating model. It is the next phase of cloud&nbsp;specialisation, driven by the practical realities of running AI at scale.</span><span data-ccp-props="{&quot;335559738&quot;:240,&quot;335559739&quot;:240}">&nbsp;</span></p> </div> <div> <h2 paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{17}" paraid="50795505"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">Why AI breaks the&nbsp;</span><span data-ccp-parastyle="No Spacing">hyperscaler</span><span data-ccp-parastyle="No Spacing">&nbsp;model</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></h2> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{19}" paraid="1505316589"><span xml:lang="EN-GB" data-contrast="auto"><span data-ccp-parastyle="No Spacing">AI workloads differ fundamentally from traditional organisational compute.&nbsp;</span></span><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">They are GPU-intensive, latency-sensitive, power-</span><span data-ccp-parastyle="No Spacing">hungry</span><span data-ccp-parastyle="No Spacing">&nbsp;and capital-heavy. They also scale unevenly</span><span data-ccp-parastyle="No Spacing">,</span><span data-ccp-parastyle="No Spacing">&nbsp;</span><span data-ccp-parastyle="No Spacing">spiking for</span><span data-ccp-parastyle="No Spacing">&nbsp;model training, throttling for inference, then surging again as models are refined,&nbsp;</span><span data-ccp-parastyle="No Spacing">retrained</span><span data-ccp-parastyle="No Spacing">&nbsp;and redeployed.</span></span></p> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{19}" paraid="1505316589"><span xml:lang="EN-US" data-contrast="auto">Hyperscalers&nbsp;were designed for breadth, not the specific demands of GPU-heavy AI workloads. Their strength lies in offering general-purpose&nbsp;services on a&nbsp;global scale, abstracting complexity behind layers of managed infrastructure. For many&nbsp;organisational&nbsp;workloads, that abstraction&nbsp;remains&nbsp;a strength. For AI workloads, however, it increasingly becomes friction.</span><span data-ccp-props="{}">&nbsp;</span></p> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{25}" paraid="1405701722"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">Companies are now&nbsp;</span><span data-ccp-parastyle="No Spacing">encountering</span><span data-ccp-parastyle="No Spacing">&nbsp;three interrelated constraints that are shaping AI infrastructure decisions. Cost opacity is rising as GPU pricing becomes increasingly bundled and variable, often inflated by overprovisioning and long reservation commitments that assume steady-state usage. At the same time, supply bottlenecks are constraining access to advanced accelerators, with long lead times, regional&nbsp;</span><span data-ccp-parastyle="No Spacing">shortages</span><span data-ccp-parastyle="No Spacing">&nbsp;and limited visibility into future availability. Layered onto this are performance trade-offs, where&nbsp;</span><span data-ccp-parastyle="No Spacing">virtuali</span><span data-ccp-parastyle="No Spacing">s</span><span data-ccp-parastyle="No Spacing">ation</span><span data-ccp-parastyle="No Spacing">&nbsp;layers and shared tenancy reduce predictability for latency-sensitive training and inference workloads.</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></p> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{27}" paraid="2147015656"><span xml:lang="EN-US" data-contrast="auto">These pressures are no longer marginal. They create a market opening that&nbsp;neoclouds&nbsp;are designed to fill.</span><span data-ccp-props="{&quot;335559738&quot;:240,&quot;335559739&quot;:240}">&nbsp;</span></p> </div> <div> <h2 paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{29}" paraid="1436860"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">What&nbsp;</span><span data-ccp-parastyle="No Spacing">neoclouds</span><span data-ccp-parastyle="No Spacing">&nbsp;change</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></h2> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{31}" paraid="1922050836"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">Neoclouds</span><span data-ccp-parastyle="No Spacing">&nbsp;</span><span data-ccp-parastyle="No Spacing">speciali</span><span data-ccp-parastyle="No Spacing">se</span><span data-ccp-parastyle="No Spacing">&nbsp;in GPU-as-a-service (</span><span data-ccp-parastyle="No Spacing">GPUaaS</span><span data-ccp-parastyle="No Spacing">), delivering bare-metal performance, rapid&nbsp;</span><span data-ccp-parastyle="No Spacing">provisioning</span><span data-ccp-parastyle="No Spacing">&nbsp;and transparent consumption-based economics. Many provide cost savings of up to 60–70% compared with&nbsp;</span><span data-ccp-parastyle="No Spacing">hyperscaler</span><span data-ccp-parastyle="No Spacing">&nbsp;GPU instances, while offering near-instant access to the latest hardware generations.</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}"> </span></p> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{31}" paraid="1922050836"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">Yet the more&nbsp;</span><span data-ccp-parastyle="No Spacing">significant change</span><span data-ccp-parastyle="No Spacing">&nbsp;is architectural rather than financial.</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></p> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{37}" paraid="1075568875"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing"></span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">Neoclouds</span><span data-ccp-parastyle="No Spacing">&nbsp;</span><span data-ccp-parastyle="No Spacing">encourage</span><span data-ccp-parastyle="No Spacing">&nbsp;</span><span data-ccp-parastyle="No Spacing">organi</span><span data-ccp-parastyle="No Spacing">s</span><span data-ccp-parastyle="No Spacing">ations</span><span data-ccp-parastyle="No Spacing">&nbsp;to make explicit decisions about AI workload placement. Training, fine-tuning, inference,&nbsp;</span><span data-ccp-parastyle="No Spacing">simulation</span><span data-ccp-parastyle="No Spacing">&nbsp;and agent execution each have distinct performance,&nbsp;</span><span data-ccp-parastyle="No Spacing">cost</span><span data-ccp-parastyle="No Spacing">&nbsp;and locality requirements. Treating them as interchangeable cloud workloads is increasingly inefficient, and often unnecessarily expensive.</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></p> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{41}" paraid="652666551"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing"></span></span><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">As a result, AI infrastructure strategies are becoming inherently hybrid and&nbsp;</span><span data-ccp-parastyle="No Spacing">multicloud</span><span data-ccp-parastyle="No Spacing">&nbsp;by design</span><span data-ccp-parastyle="No Spacing">,&nbsp;</span><span data-ccp-parastyle="No Spacing">not as a by-product of vendor sprawl, but as a deliberate response to workload reality. The cloud market is&nbsp;</span><span data-ccp-parastyle="No Spacing">fragmenting along</span><span data-ccp-parastyle="No Spacing">&nbsp;functional lines, and&nbsp;</span><span data-ccp-parastyle="No Spacing">neoclouds</span><span data-ccp-parastyle="No Spacing">&nbsp;occupy a clear and growing role within that landscape.</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></p> </div> <div> <h2 paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{45}" paraid="236443624"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing"></span></span><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">Co-opetition, not disruption</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></h2> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{50}" paraid="841843236"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">The growth of&nbsp;</span><span data-ccp-parastyle="No Spacing">neoclouds</span><span data-ccp-parastyle="No Spacing">&nbsp;is not a&nbsp;</span><span data-ccp-parastyle="No Spacing">hyperscaler</span><span data-ccp-parastyle="No Spacing">&nbsp;extinction event. In fact,&nbsp;</span><span data-ccp-parastyle="No Spacing">hyperscalers</span><span data-ccp-parastyle="No Spacing">&nbsp;are among their largest customers and partners, using&nbsp;</span><span data-ccp-parastyle="No Spacing">neoclouds</span><span data-ccp-parastyle="No Spacing">&nbsp;as elastic extensions of capacity when demand spikes or accelerator supply tightens.</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></p> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{52}" paraid="764481698"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing"></span></span><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">This creates a new form of co-opetition.&nbsp;</span><span data-ccp-parastyle="No Spacing">Hyperscalers</span><span data-ccp-parastyle="No Spacing">&nbsp;retain control of platforms,&nbsp;</span><span data-ccp-parastyle="No Spacing">ecosystems</span><span data-ccp-parastyle="No Spacing">&nbsp;and</span><span data-ccp-parastyle="No Spacing">&nbsp;</span></span><span xml:lang="EN-GB" data-contrast="auto"><span data-ccp-parastyle="No Spacing">company relationships,</span><span data-ccp-parastyle="No Spacing">&nbsp;</span></span><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">while&nbsp;</span><span data-ccp-parastyle="No Spacing">neoclouds</span><span data-ccp-parastyle="No Spacing">&nbsp;</span><span data-ccp-parastyle="No Spacing">speciali</span><span data-ccp-parastyle="No Spacing">s</span><span data-ccp-parastyle="No Spacing">e</span><span data-ccp-parastyle="No Spacing">&nbsp;in raw AI performance, speed to hardware and regional capacity. Each addresses a different constraint in the AI value chain.</span></span></p> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{52}" paraid="764481698"><span xml:lang="EN-GB" data-contrast="auto"><span data-ccp-parastyle="No Spacing">For companies and&nbsp;</span><span data-ccp-parastyle="No Spacing">organi</span><span data-ccp-parastyle="No Spacing">s</span><span data-ccp-parastyle="No Spacing">ations</span><span data-ccp-parastyle="No Spacing">&nbsp;buying cloud services</span><span data-ccp-parastyle="No Spacing">,&nbsp;</span></span><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">this blurs traditional cloud categories. The question is no longer simply which cloud provider to use, but how AI workloads should be placed across environments to&nbsp;</span><span data-ccp-parastyle="No Spacing">optimi</span><span data-ccp-parastyle="No Spacing">s</span><span data-ccp-parastyle="No Spacing">e</span><span data-ccp-parastyle="No Spacing">&nbsp;cost, performance,&nbsp;</span><span data-ccp-parastyle="No Spacing">sovereignty</span><span data-ccp-parastyle="No Spacing">&nbsp;and operational risk.</span></span></p> <h2 paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{52}" paraid="764481698"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">The real risk: tactical adoption</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></h2> </div> <div> <p><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">The greatest risk for CIOs and technology leaders is treating&nbsp;</span><span data-ccp-parastyle="No Spacing">neoclouds</span><span data-ccp-parastyle="No Spacing"> as a short-term workaround for GPU shortages. </span></span><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">Neoclouds</span><span data-ccp-parastyle="No Spacing">&nbsp;introduce new considerations: integration complexity with existing platforms, dependency on specific accelerator ecosystems, energy&nbsp;</span><span data-ccp-parastyle="No Spacing">intensity</span><span data-ccp-parastyle="No Spacing">&nbsp;and vendor concentration risk. Used tactically, they can fragment architectures and increase long-term operational exposure.</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}"> </span><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">Used strategically, however, they unlock something more valuable</span><span data-ccp-parastyle="No Spacing">,&nbsp;</span><span data-ccp-parastyle="No Spacing">control:</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></p> </div> <div> <ul class="default-list"> <li paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{74}" paraid="1845249789">Control over cost visibility, through transparent, consumption-based GPU pricing that reduces overprovisioning and exposes the true economics of AI workloads</li> <li paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{74}" paraid="1845249789">Control over data locality and sovereignty, by enabling regional or sovereign deployments where regulatory or latency requirements demand it</li> <li paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{74}" paraid="1845249789">Control over workload placement, by allowing organisations to deliberately orchestrate AI training and inference across hyperscalers, neoclouds and on-premises environments based on performance, cost and compliance requirements.&nbsp;</li> </ul> <h2 paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{74}" paraid="1845249789"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">From cloud strategy to AI placement strategy</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></h2> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{82}" paraid="909641615"><span xml:lang="EN-US" data-contrast="auto"><span data-ccp-parastyle="No Spacing">Neoclouds</span><span data-ccp-parastyle="No Spacing">&nbsp;are not an alternative cloud. They are a forcing function, compelling&nbsp;</span><span data-ccp-parastyle="No Spacing">organi</span><span data-ccp-parastyle="No Spacing">sa</span><span data-ccp-parastyle="No Spacing">tions</span><span data-ccp-parastyle="No Spacing">&nbsp;to rethink infrastructure assumptions that&nbsp;</span><span data-ccp-parastyle="No Spacing">no</span><span data-ccp-parastyle="No Spacing">&nbsp;longer hold in an AI-driven world.</span></span><span data-ccp-props="{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}">&nbsp;</span></p> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{84}" paraid="1070728363"><span xml:lang="EN-US" data-contrast="auto">The new competitive advantage will come from&nbsp;AI&nbsp;placement strategy&nbsp;–&nbsp;deciding when&nbsp;hyperscalers,&nbsp;neoclouds, on-premises or edge environments are the right choice for each workload.</span><span data-ccp-props="{&quot;335559738&quot;:240,&quot;335559739&quot;:240}">&nbsp;</span></p> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{86}" paraid="1160500405"><span xml:lang="EN-US" data-contrast="auto">Over the next five years,&nbsp;IT&nbsp;leaders will be defined not by how much cloud they consume, but by how precisely they place intelligence where it creates the most value.</span><span data-ccp-props="{&quot;335559738&quot;:240,&quot;335559739&quot;:240}">&nbsp;</span></p> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{86}" paraid="1160500405"><span data-ccp-props="{&quot;335559738&quot;:240,&quot;335559739&quot;:240}"><a rel="noreferrer noopener" target="_blank" href="https://www.gartner.com/en/experts/mike-dorosh"><span xml:lang="EN-US" data-contrast="none">Mike Dorosh</span></a><span xml:lang="EN-US" data-contrast="auto"> is a senior director analyst at Gartner.</span>&nbsp;</span></p> </div> <div> <p paraeid="{d223dbf3-b08c-424e-9e49-324f8a21e757}{89}" paraid="1196796117"><span xml:lang="EN-GB" data-contrast="auto">Gartner analysts&nbsp;will further&nbsp;explore how&nbsp;neoclouds&nbsp;and AI workload placement are reshaping cloud and data strategies at the&nbsp;Gartner&nbsp;</span><a rel="noreferrer noopener" target="_blank" href="https://www.gartner.com/en/conferences/emea/symposium-spain"><span xml:lang="EN-US" data-contrast="none"><span data-ccp-charstyle="Hyperlink">IT Symposium/</span><span data-ccp-charstyle="Hyperlink">Xpo</span></span></a><span xml:lang="EN-GB" data-contrast="auto">&nbsp;in Barcelona, from 9–12 November 2026.</span><span data-ccp-props="{}">&nbsp;</span></p> </div> Neoclouds are set to change the economcs of AI in the cloud. By 2030, neocloud providers will capture around 20% of the $267bn AI cloud market https://cdn.ttgtmedia.com/visuals/German/article/cost-of-storage-2-adobe.jpg https://www.computerweekly.com/opinion/Gartner-Why-neoclouds-are-the-future-of-GPU-as-a-Service Fri, 20 Feb 2026 13:00:00 GMT <p>T-Labs, the research and development division within Deutsche Telekom, has collaborated with quantum networking firm, Qunnect on a demonstration of quantum teleportation over a commercial network.</p> <p>T-Labs deployed Qunnect’s commercially available quantum entanglement distribution hardware in Berlin in a trial to show how quantum technology can be used to tackle instabilities and interferences in existing <a href="https://www.computerweekly.com/news/252506610/BT-advances-hollow-core-fibre-research-with-worlds-first-trial-of-quantum-secure-comms">telecom infrastructure</a>. T-Labs said the trial shows how a telecommunications operator can integrate quantum teleportation capabilities into operational networks. &nbsp;</p> <p>The experiment, which was run in January using the Qunnect Carina platform, along with Deutsche Telekom’s Berlin quantum infrastructure, achieved quantum teleportation more than 30km of commercial fibre cables.</p> <p>In an optical fibre network, optical repeaters are needed every 50km or so to ensure that the photons, which are used to transmit network traffic, are able to travel longer distances. For a quantum internet, it is not possible to amplify and forward quantum information. In an <a href="https://phys.org/news/2025-11-quantum-teleportation-photons-distant-sources.html">article on phys.org</a>, university of Strasbourg scientific consultant Jutta Witte states that quantum physics allows information to be transferred from one photon to another as long as the information stays unknown. She describes the process as quantum teleportation.</p> <p>T-Labs and Qunnect said the trial teleported qubits generated by a weak coherent source over a 30km fibre loop connecting T-Lab’s Quantum Lab to a node on the Berlin fibre testbed. The Carina platform integrates an entanglement generator that produces pairs of quantum-entangled photons. According to T-Labs and Qunnect, these quantum-entangled photons are distributed over telecom fibre, along with a polarisation compensation component that counteracts environmentally induced noise in both buried and aerial fibre to support high-rate, high-fidelity transport of quantum bits between network nodes.</p> <p>T-Labs describes Quantum teleportation as a key building block for the future quantum internet, enabling the transfer of quantum information between distant locations. It does this by recreating an identical quantum state of a particle at the destination using pre-shared quantum entanglement rather than transmitting a physical particle. &nbsp;</p> <p>Abdu Mudesir, Telekom board member for product and technology, said: “In Berlin, we have now proven that quantum information can be transmitted over 30 kilometres of commercial telecom fibre optics outside of a laboratory. This is done in parallel with regular data traffic and with a very high average accuracy of 90%.”</p> <p>The trial of quantum teleportation is being positioned as a step towards <a href="https://www.computerweekly.com/news/366634683/IBM-Cisco-light-up-quantum-networking-collaboration">networking quantum computers over longer distances</a> in the future and pooling computing power in more than one location. “This will create the next generation of secure communication and a building block for Europe’s technological sovereignty,” Mudesir added.</p> <p>“Teleportation is a novel tool for moving information around networks leveraging quantum physics,” said Mael Flament, CTO at Qunnect. “We are showing the building blocks of teleportation can operate inside a <a href="https://www.computerweekly.com/feature/EU-researchers-inch-closer-to-a-viable-quantum-internet">real network</a>, in real racks, under operator control, advancing it from a laboratory experiment to something a telecommunications provider can deploy.”&nbsp;</p> <p>Qunnect, Deutsche Telekom and other partners plan to extend the experiment to multi-node teleportation configurations, expanding the distance across which they are able to transfer quantum states. Qunnect and Deutsche Telekom said this expansion will evaluate broader deployment and next generation use cases within a metro-scale carrier network infrastructure.</p> <div class="extra-info"> <div class="extra-info-inner"> <h3 class="splash-heading">Read more quantum computing stories</h3> <ul class="default-list"> <li>Government showcases <a href="https://www.computerweekly.com/news/366634198/Government-showcases-UK-quantum-computing-pledge">UK quantum computing pledge</a>: With £670m set aside, the government is keen to show how the funding is supporting quantum innovation.</li> <li><a href="https://www.techtarget.com/searchcio/tip/Attention-CISOs-Quantum-computing-security-risks-are-here">Quantum computing security risks</a> are here: The promise of quantum processors solving complex problems at extraordinary speeds offers numerous business opportunities. But what risks does this new technology present?</li> </ul> </div> </div> The research and development arm of telco Deutsche Telekom has worked with Qunnect to demo quantum networking over 30km of optical fibre https://cdn.ttgtmedia.com/visuals/ComputerWeekly/HeroImages/fibre-optics-network-4G-5G-monsitj-adobe.jpg https://www.computerweekly.com/news/366639238/T-Labs-demos-commercial-viability-of-quantum-networking Thu, 19 Feb 2026 08:00:00 GMT 60 editor@computerweekly.com