Anyone making predictions about IT and networking will inevitably come up against a major problem – the pace of development is so quick that it is difficult to make accurate estimations. There is also a prediction that seems axiomatic, in that network management will rely increasingly – if not exclusively at some point – on artificial intelligence (AI).
AI is being deployed to observe and gain insight from a host of networking operations, including, but not limited to, configuration data, messages from devices and monitoring data. Companies can rely on the fact that AI “knows” how networks should be operating and will send alerts when they do not operate as expected, as well as explaining why and suggesting ways to resolve these issues.
Inevitably, it seems that any conversation about the power of AI in networking must start with Nvidia and its CEO Jensen Huang. When it comes to predictions, the AI company’s founder has been consistently on the money – almost literally – for a long while.
At a tech conference in 2024, Huang said the era of generative AI (GenAI) had already arrived, and that enterprises must engage with “the single most consequential technology in history”, noting that what was happening was the greatest fundamental computing platform transformation in 60 years, encompassing general-purpose computing to accelerated computing.
For Huang, the key to success is making use of the vast amounts of data that enterprises generate through the deployment of AI tools and services. This means a radical shift in what IT organisations within businesses do.
“We’re sitting on a mountain of data – all of us. We’ve been collecting it in our businesses for a long time. But until now, we haven’t had the ability to refine that, then discover insight and codify it automatically into our company’s natural experience, our digital intelligence. Every company is going to be an intelligence manufacturer. Every company is built on domain-specific intelligence. For the very first time, we can now digitise that intelligence and turn it into our AI – the corporate AI,” he observed.
“AI is a lifecycle that lives forever. What we are looking to do is turn our corporate intelligence into digital intelligence. Once we do that, we connect our data and our AI flywheel so that we collect more data, harvest more insight and create better intelligence. This allows us to provide better services or to be more productive, run faster, be more efficient and do things at a larger scale.”
Making strides towards autonomous networks
Today, network management and network operations are indeed being done at a faster rate. In February 2026, Nvidia’s fourth annual State of AI in telecommunicationssurvey concluded that AI has already accelerated how AI is driving enterprise transformation, unlocking new business and revenue opportunities.
Respondents encompassed a range of industry segments, including internet service providers, independent software suppliers, network equipment providers, consulting service providers, operators and systems integrators. The study showed AI has a tangible revenue impact and return on investment (ROI). The top AI use cases cited by respondents were AI for autonomous networks (50%), improved customer service (41%) and internal process optimisation (33%).
Overall, around nine out of 10 respondents said AI was helping to increase revenue and reduce costs. Operators, representing about a quarter of the 1,000 responses in the survey, were also seeing the benefit, with 90% saying AI has had a positive impact on revenue and costs. Some 60% said their organisation was using or assessing GenAI, up from 49% in 2024, while 89% said open source models and software were important to their AI strategy.
The impact on revenue and ROI was found to be leading telecommunications companies to increase their AI budgets in 2026. Overall, 89% of respondents said their AI budget would increase in the next 12 months, up from 65% in the 2024 survey, with 35% saying their budgets would increase by more than 10% compared with 2025.
According to Nvidia, these findings signal a bold step towards autonomous networks – AI-driven, self-managing systems that can self-configure, self-heal and self-optimise with minimal human intervention. In addition, 88% of organisations reported being between levels 1-3 of autonomy, as defined by the TM Forum, and the use of GenAI and agentic AI was expected to accelerate the shift to level 5 autonomous networks.
A new era of agentic network management
According to John Burke, chief technology officer and research analyst at Nemertes Research, this era of network management is being ushered in – and redefined – by agentic AI. “AI agents are designed to exhibit goal-directed behaviour. In the context of the network, AI agents work to keep the network functioning at expected levels and maintain network configuration according to company security policies,” he says.
“In addition, agentic AI can show some level of environmental awareness, such as knowing not to restart a switch as part of routine maintenance during business hours. Like their non-agentic counterparts, agentic AI systems can create multistep plans and adapt plans to changing circumstances. But AI agents can execute those plans as well as more broadly pursue policy and behavioural objectives with minimal human intervention.”
Burke says agentic AI constantly cycles through the four stages of what is known as an OODA – observe, orient, decide and act – loop and learns as it goes. In operation, this means: observe, as in identifying what happens in the network; orient, by analysing and understanding the data based on its past learning; decide, by determining which actions it should take in response based on the data; and act, as in executing the agent’s decisions.
Improved time to value
This results in a faster ROI, as Chetan Sharma, CEO of Chetan Sharma Consulting, explains: “Autonomous networks are delivering return on investment faster than any other AI use case because they directly reduce outages, energy consumption and manual intervention. Agentic AI accelerates this by coordinating decisions across domains in real time.
“Generative AI delivered fast productivity gains, but agentic AI is where telecoms begins to see structural ROI. Autonomous agents can act across networks, IT and customer journeys, turning insights into decisions without human delay.”
Generative AI delivered fast productivity gains, but agentic AI is where telecoms begins to see structural ROI. Autonomous agents can act across networks, IT and customer journeys, turning insights into decisions without human delay
Chetan Sharma, Chetan Sharma Consulting
From an operational perspective, this will likely result in the transition of IT departments from the traditional practice of reactive troubleshooting to proactive management. This concept is being deployed by Tata Communications, which launched the IZO DC Dynamic Connectivity self-healing network platform in March 2025. The platform is designed to eliminate costly datacentre downtime and support the growing demands of AI. In this, enterprises operate across global locations and cloud environments, moving huge volumes of data in real time to support AI workloads and business needs.
Explaining the rationale for the launch, the digital ecosystem provider said that in the current digital economy, disruptions from cable cuts, route failures or sudden AI workload spikes can bring business to a standstill.
The company also warned that the networks connecting many enterprise datacentres were built for a different era – traditional datacentre links were designed for predictable workloads and stable traffic patterns, while the current reality is far more dynamic. Increasing geopolitical constraints, cable outages, route failures or sudden spikes in demand could cascade into service disruption and operational risk, leading to costly downtime. In such scenarios, the traditional response has often been reactive and manual, consuming valuable time when businesses need certainty and speed.
In contrast, the new platform deploys deterministic multipath routing to deliver predictable latency and performance. This promises to transform resilience from a reactive process into an autonomous capability, changing how enterprises connect their datacentres in an increasingly AI-driven and distributed world.
The new Tata Communications platform is smart enough to re-route traffic automatically within seconds without manual intervention during disruptions and is able to maintain very high levels of service availability across mission-critical infrastructure that supports business-critical applications. Through a unified digital interface and application programming interfaces (APIs), enterprises can monitor performance, receive proactive alerts and dynamically scale bandwidth as workloads evolve. The result is that resilience becomes an autonomous capability and a default state, not a contingency.
In a similar vein, in mid-2025, Nokia announced the launch of its Autonomous Networks Fabric, designed to accelerate full network automation in an open, cloud-native, multi-supplier environment, including trained models, integrated security and AI apps for automation workflows. The fabric was designed to enable automation at scale and address issues encountered in this endeavour – the comms tech provider said it had seen a steady increase in the number of companies moving towards implementing fully autonomous networks, yet it also found that many have been held back by legacy systems, siloed processes and fragmented data.
The Autonomous Networks Fabric looks to reduce the complexity of automation while allowing network providers to improve reliability and make operational cost savings by quickly testing new ideas and integrating those that deliver desired benefits. It combines observability, analytics, security and automation across every network domain, allowing a network to behave as one adaptive system, regardless of supplier, architecture or deployment model.
In addition, the fabric federates the use and distribution of data and AI across an organisation, monitoring the chain of custody from end to end and ensuring quality and consistency in automation. Trained large language models (LLMs) support all automation through a knowledge engine designed to give reasoning for how data is interpreted, how issues are analysed and why certain actions are recommended.
The fabric is also constructed to work with Google Cloud’s GenAI, including Vertex AI and BigQuery, to deliver agent-driven workflows for network operations. Capabilities on offer include real-time monitoring and visibility into network traffic patterns, anomaly detection, zero-touch remediation of performance issues, and support for elastic scale-out and disaster recovery to the cloud.
Adopting AI for network management
Between the likes of Nvidia, Tata Communications and Nokia, a whole host of AI-driven autonomous network management solutions are currently available. Yet there are a few fundamental assumptions at play in looking at how firms can best take advantage of AI for autonomous network management – one of which is the intrinsic robustness of company infrastructures.
April 2026 research by Cisco found that while as many as two-thirds of industrial organisations have moved to active AI deployments in live operational environments, infrastructure and organisational alignment – especially networking and security – will dictate how businesses achieve real transformation.
The resulting State of industrial AI report 2026 looks to provide a data‑driven view into how industrial organisations are adopting AI, the challenges they face as AI moves into live operations and the opportunities created as AI becomes embedded in physical systems, infrastructure and workflows.
One of the top findings is that AI organisations are harnessing AI to drive progress and overcome industry challenges, and that it is now delivering measurable operational benefits, in particular in use cases such as process automation, automated quality inspection, predictive maintenance, logistics and energy forecasting. Strong expected benefits from AI include productivity (59%), cost reduction (42%) and sustainability.
Yet just as adoption is accelerating, many firms in the survey conceded that they are struggling to sustain and expand deployments, with readiness across network infrastructure, security and skills increasingly determining whether AI can scale consistently across core physical environments.
Network readiness and security posture were cited as the primary factors shaping how quickly and safely organisations scale AI across connected assets, machines and sites. The report observes that as AI becomes embedded in machines, sensors, vision systems and autonomous operations, organisations face rising demands for reliable connectivity, wireless mobility, predictable latency, edge compute and power. This is making network readiness a gating factor for AI deployments.
Seeking network efficiency, security and scalability
Such concerns are also voiced by Gordon Thomson, president of EMEA at Cisco, who believes that in a world defined by AI, companies run the risk of being left behind if they are not leading with AI in their operations. He says that with AI, the tech industry has reached a key point as regards to infrastructure, compute, networks, security and monitoring. However, according to Thomson, the IT infrastructure organisations have relied on to date was not built for the scale and the velocity of future workloads.
“The solution isn’t about stacking tiny new products on top of each other – that just creates complexity and will slow you down. [Success] requires a platform that uses data to be more efficient, more secure and more scalable,” he says.
The bottom line is that there is simply a seismic shift underway in how networks are being managed, and the key to all of this is AI – and increasingly agentic AI. As networks become more autonomous, they will require different forms of AI – from classical algorithms to language-based systems and intelligent agents – to each contribute distinct capabilities.
Networking has now evolved far beyond moving data to moving gatherable intelligence across local and regulated infrastructure. Moreover, autonomous networks can deliver immediate ROI by eliminating human effort from repetitive, reactive workflows, with the fastest impact areas being energy management, fault prediction, configuration drift correction and capacity planning. And this will likely be the future – a future that will be autonomous and observed.
Read more about AI in network management
10 AI-driven network management tasks: AI can automate key network operations tasks, such as anomaly detection, event correlation and ticketing. This shifts network engineers toward governance and system design.
