Krot_Studio - stock.adobe.com
APAC enterprises to boost AI spend by 15% in 2026
Lenovo’s CIO Playbook 2026 reveals that 96% of ASEAN organisations are planning to invest more in AI, with a growing reliance on hybrid infrastructure to manage rising inference costs
Enterprises across Asia-Pacific (APAC) are doubling down on artificial intelligence (AI) initiatives, with 96% of organisations planning to increase their AI investments by an average of 15% in 2026, a new study has found.
According to Lenovo’s CIO playbook 2026 study conducted by IDC, the growth in AI investments is driven by changing business priorities. While improving productivity has been a key goal of AI and automation efforts, CIOs have now earmarked revenue growth as their top business priority this year.
Speaking at the launch of the report, Rakshit Ghura, vice-president and general manager of digital workplace solutions at Lenovo, highlighted the change in executive mindset.
“Increasing revenue – the top line and bottom line – is becoming the top-notch requirement for enterprises. It was at the number eight position last year; this time, it is number one,” said Ghura. “Enterprises want to completely transform with AI, redo their traditional operating models, and embrace new ways of working.”
The report, which surveyed 920 technology leaders, suggested that the hype phase of AI is giving way to real-world projects. About 66% of organisations in the region are now either piloting or systematically adopting AI.
In addition, 88% of APAC organisations expect returns from AI projects this year, with an anticipated return on investment (ROI) of $2.85 for every dollar invested.
However, the source of this investment is changing. The report found that half of non-IT departments are now funding AI initiatives, with projects increasingly originating from marketing, finance, and business units rather than the CIO’s office alone.
Scott Tease, vice-president and general manager of enterprise product line at Lenovo, noted that successful AI adoption now relies on partnerships between IT and business teams.
“If a CIO wants to pick an operating system or database, they make that decision and implement it. With AI, they likely have the AI skills, but they don’t know the business problem well enough,” Tease said. “That is why projects coming from outside the IT department are so important.”
Cost of AI inferencing vs. training
As organisations move from building models to running them, AI infrastructure strategies are also changing. The report noted that the industry is focusing more on AI inferencing, where an AI model processes data to make predictions and deliver outputs.
Sumir Bhatia, president of APAC at Lenovo’s infrastructure solutions group, warned CIOs that not planning for inference costs could be detrimental, noting that inference can cost 15 times more than training over the lifetime of a model.
“By 2030, 75% of AI compute will be dedicated to inferencing,” Bhatia said. “The value doesn’t come from building models in a lab; the value comes from how you use that model.”
To manage these costs and meet data sovereignty requirements, 86% of APAC organisations are opting for a hybrid AI approach, repatriating certain workloads from the public cloud to on-premises datacentres or edge devices.
In the ASEAN region, the demand for hybrid AI infrastructure is being driven by security and regulatory requirements, according to Nigel Lee, general manager for Singapore at Lenovo. “A lot of countries are putting guardrails around AI and looking to pass legislation around the adoption of AI,” he said.
Looking ahead, the report singled out agentic AI systems capable of independent action and decision-making as the next frontier. While interest is high, with 60% of APAC organisations exploring the technology or planning limited deployments, only 10% consider themselves ready to scale their agentic AI implementation.
Fan Ho, executive director and general manager of Lenovo’s solutions and services group in APAC, noted that generally, the uneven maturity of infrastructure and lack of quality data are some of the barriers to scaling AI initiatives. Governance is another hurdle in scaling AI deployments.
“If you are now going to manage 50,000 humans plus 50,000 agents, how do you put solid governance processes around that?” Ghura said, noting that without deep integration with enterprise systems and strict compliance guardrails, the growing adoption of agentic AI could compromise data security.
Read more about AI in APAC
- Australia’s Woolworths Group will enhance its digital shopping assistant, Olive, using Google Cloud’s newly released agentic AI platform, Gemini Enterprise for Customer Experience.
- Research from Thoughtworks reveals that while 77% of global businesses are focused on generating revenue from AI initiatives, Asian markets are leading the charge in agentic AI adoption, job creation and executive confidence.
- Microsoft has expanded its AI footprint in India, teaming up with four of the country’s largest IT services companies to deploy agentic AI capabilities across enterprises.
- Singlife has started deploying Salesforce Agentforce to assist customer service teams as part of a broader move to harness the capabilities of agentic AI across its business.
