The hunt for AI ROI in 2026

By 2026, 70% of G2000 CEOs will focus AI return on investment (ROI) on growth, driving C-suite efforts to boost revenue and reinvent business models without growing headcount, IDC has predicted. 

Industry observers agree that the emphasis on AI will be different in 2026.

Yasutaka Mizutani, APAC President, Colt Technology Services, said: "2026 will mark the shift from AI model training to AI inference becoming the dominant workload, with enterprises expecting AI systems to deliver predictions, decisions and operational outcomes in real time." 

Source: Colt Technology Services.
Mizutani.

Colt’s research has found that one in five global firms spend US$750,000 annually on AI while 95% of the respondents in a recent MIT report study see no return on their investments, Mizutani shared. "Businesses will close the gap between AI investment and tangible ROI by adopting AI maturity assessments, structured value tracking and integrated lifecycle management built around measurable business impact," he said.

In Asia Pacific and Japan (APJ), organisations are also intensifying AI investments, with over 50% reallocating funds from other areas to double down on AI, Amit Khandelwal, Regional VP and MD for Southeast Asia, UiPath predicted.

"For those planning to invest in AI agents in the next two years, close to a third (29%) already have an established investment plan. This is a deliberate, strategic decision to fund future growth by shifting away from traditional investments," he shared. 

"However, organisations now expect AI to deliver concrete, measurable results that will secure their position in a competitive future. Across APJ, C-suites are demanding a 2–4x return on investment within 12 to 18 months of deployment. This is driving a new level of rigour and accountability in every step of the AI project lifecycle."

Rajiv Ramaswami, President and CEO of Nutanix, said businesses would "move from AI-first to AI-smart" in 2026. "Many organisations dove headfirst into AI without thinking about the consequences and anticipating the real business use cases. Just like we saw with the initial rush to cloud-first adoption, enterprises are going to re-evaluate their technology stacks and truly see where AI makes sense," he said.  

"AI applications have become business-critical more quickly than any other applications we've ever seen. In 2026, we’re going to see organisations integrate AI into their enterprise IT and explore three areas: business resiliency, Day Two operations, and security." 

Source: Solace.
Funnekotter.

Edward Funnekotter, Chief AI Officer at Solace, predicted that AI will transition in 2026 "from experimental pilots to robust AI applications that have the rigour to stand up to everyday use in industries that demand real-time delivery in order to match customer, supplier, and employee expectations".

"The last two years were defined by the explosive promise of generative AI, then followed the agentic-driven rush. The year of 2026 is going to be a year of reckoning, one that truly shows which AI applications make the grade," Funnekotter said.

"We are moving away from the initial rush of excitement and getting back to real business value. While the models themselves continue to improve, the focus for enterprises is shifting from 'what can this cool demo do?' to 'how do we run this safely in production?'"

Remus Lim, Senior VP, Asia Pacific & Japan, Cloudera said: "In 2026, economic headwinds will push organisations to shift from 'AI for innovation' to 'AI for impact'. The next phase of enterprise AI will be defined by a sharper focus on return on investment, efficiency, and purpose-built deployment." 

Source: Cloudera.
Lim.

Pascal Brier, Chief Innovation Officer at Capgemini and Member of the Group Executive Committee touched on impact as well. "AI is without doubt the defining technology of the decade, but the pace of investment has outstripped the speed at which organisations have deployed and extracted value from it," he said.

"Taking stock of where some of their AI experimentations failed to deliver the expected outcomes, business leaders now understand that the issue didn’t come from the technology itself but from the business approach and methodology."

"Full-scale deployments will take time, and long-term value will not lie in isolated AI use cases but in enterprise-wide implementations. While the true growth phase begins, an AI ecosystem more rooted in operational value and enterprise architecture is emerging, starting with data foundations and infrastructure, and focusing on 'human-AI chemistry'," Brier added.

"2026 will be the moment to move from proof-of-concept to proof-of-impact, ensuring AI drives measurable outcomes, trust, and collaboration at scale, whilst laying the foundations for larger-scale transformation to follow."

Rajiv Ramaswami, President and CEO of Nutanix, predicted that businesses will move from 'AI-first' to 'AI-smart'. "Many organisations dove headfirst into AI without thinking about the consequences and anticipating the real business use cases. Just like we saw with the initial rush to cloud-first adoption, enterprises are going to re-evaluate their technology stacks and truly see where AI makes sense," he said.  

"AI applications have become business-critical more quickly than any other applications we've ever seen. In 2026, we’re going to see organisations integrate AI into their enterprise IT and explore three areas: business resiliency, day two operations, and security."

If day one is the installation or onboarding of a technology or solution, day two refers to actually making use of it.

Kai Wombacher, Product Manager, IBM Kubecost, agreed that AI ROI will become a hot topic in 2026. “In many cases, companies have been so focused on the AI horse race that model cost considerations have been deprioritised. As organisations face increasing pressure to demonstrate AI value and align investments with business outcomes, FinOps for AI will gain traction as a critical framework to understand and optimise AI and infrastructure spend," he said. 

Source: Dataiku.
Douetteau.

Waste in AI use will be a focus in 2026, Florian Douetteau, Co-founder and CEO at Dataiku said. "Companies will start hunting for 'wasted tokens', realising that their multimillion-dollar commitment to OpenAI/Anthropic/Google is excessive, and that their spend could be reduced by better application or self-hosting open-source models. It will lead to a crisis in terms of the pricing and revenue model for cloud vendors and LLM frontier models," he said. 

"CIOs and CTOs will need to build a strong business case for every AI initiative. In that case, they need to understand that not every workload needs high-end GPUs or complex models. As an analogy: if the goal is simply to get from your home to Changi Airport, a sedan will do just fine. But if you’re racing in Formula 1 where every sub-second counts, you’ll need an F1 car. The same principle applies to AI: organisations should invest according to their objectives, not the trend," said Lim.

Lim shared that according to Cloudera’s The Future of Enterprise AI Agents report, 67% of respondents in
Singapore and China cite unclear ROI as a top barrier to adoption, while 100% of respondents in Japan and Indonesia admit that AI agents’ perceived complexity delays implementation. 

"As Gartner predicts, more than 40% of AI agent projects may be scrapped by 2027 due to poor ROI," he added.

"In 2026, AI will separate the builders from the believers. Ultimately, the winners will be those that integrate AI seamlessly into their data fabric, supported by strong data foundations, standardised metrics, and sustainable governance. Meanwhile, those that don’t get their data foundations right will remain trapped in endless pilots."  

According to Forrester's 2026 predictions, the relentless push for ROI will drive enterprises to push some of their AI spend to 2027. "AI value is failing to land: only 15% of AI decision-makers reported an EBITDA lift for their organisation in the past 12 months, and fewer than one-third can tie the value of AI to P&L changes," the consultancy said in a blog post. 

"Given the high payback expectations from AI investments now, we believe that CEOs will pull more CFOs into AI decisions, leading enterprises to delay 25% of their planned spend into 2027."

Source: BytePlus.
Zhang.

The fundamentals will still matter, said Yongliang Zhang, GM, BytePlus. "Whether enterprise AI truly breaks through will depend on five practical factors," he said.

"These basics will matter more than the size of any model or the novelty of the technology."

Zhang listed:

- Controllability and accuracy: Outputs need to be reliable, aligned with intended use cases, and accurate enough to support confident decision-making.

- Cost management: AI investments need to be tracked prudently across infrastructure, integration, training, and maintenance — with clear returns and long-term scalability in mind.

- Security and risk controls: Robust safeguards are essential to protect sensitive data, comply with regulations, and mitigate risks like model misuse or hallucination. These controls are the foundation for stakeholder trust.

- Compatibility with existing systems: AI needs to integrate with current tech stacks and workflows to avoid costly reengineering and enable faster deployment.

- Data protection: Enterprise data must be centralised, pre-processed, and governed with clear access controls.

OpenAI offers ChatGPT, while Anthropic is known for the Claude AI model. Google's AI model is Gemini. AI models consume tokens when they interact with users. 

Hashtags: #2026predictions 

*Day Two operations refers to focusing on what happens after deployment, such as monitoring, maintenance, and optimisation, as opposed to Day Zero (planning), and Day One - actual deployment. EBITDA is an acronym for earnings before interest, taxes, depreciation and amortisation. G2000 refers to the top 2,000 global companies, while LLM stands for large language model. P&L refers to profit and loss.