Across the Asia Dialogues Forum series, senior executives returned repeatedly to the same concern: companies across Asia are moving fast on AI investment but far more slowly on defining the institutional problems those systems are meant to solve. The technology is accelerating faster than the thinking around it. And the data confirms it.

The FOMO Trap: Why AI Investment Strategy Is Failing

In Hyderabad, 50-plus leaders from Microsoft, Deloitte, PwC, DBS Bank, and Larsen and Toubro named the root cause of a failed AI investment strategy with unusual directness. Nearly 90% of global AI investments fail not because of weak technology but because of misplaced intent. Boards demand AI. Investors expect transformation narratives. Competitors announce automation programmes weekly. Organisations respond by implementing systems before they have decided what meaningful success looks like.

Ritesh Dogra, Gen AI Innovation Head, put it plainly: "Every model we were building had one significant gap: context. Technology without context creates no value. Nearly 90% of global AI investments are write-offs because of FOMO. Boards demand AI, investors expect it, customers ask for it, even when simpler solutions would work better."

The consequence is predictable. Forty-eight percent of organisations report fragmented AI investments. Twenty-seven percent pursue department-led initiatives rather than enterprise strategies. The activity is visible. The outcomes are not.

From Operational ROI to Strategic ROI

The New Delhi forum pushed the AI investment strategy conversation further and in a more consequential direction. Leaders there were less interested in whether AI could improve productivity and more focused on whether organisations were measuring success against the right outcomes entirely.

Anurag Jain, Chief Digital and Technology Officer at KFC India-Yum! Brands drew a distinction that reframed the room's thinking. Operational ROI is measurable and immediate: cost efficiency, speed, automation of repetitive tasks. But strategic ROI is what actually determines whether an organisation survives the decade: the ability to attract the best talent because you are a technologically ambitious organisation, the ability to build competitive moats that take years to replicate, and the ability to remain genuinely relevant in 2035. As Jain observed: "For organisations to withstand disruption and remain successful, their business strategy, technology strategy, and talent strategy must go hand in hand. Technology is most effective when it disappears into day-to-day work."

That last insight is among the most practically important to emerge from the forum series. The best AI integrations are invisible. When a kitchen display system optimises order sequencing without anyone managing it, when a credit algorithm assesses a loan in the background, when a CRM auto-populates with customer context before a call begins: in each case, technology has dissolved into the workflow. It is not a tool being used. It is infrastructure being inhabited.

The Infrastructure Trap: Where AI Wealth Is Really Created

Gaurav Dalmia of Dalmia Group brought a perspective to the Delhi forum that no other city in the series could offer: forty years of investment experience across every major wave of disruption India has navigated. His most provocative argument cut against the current consensus on where AI investment creates the most value.

"AI infrastructure, like data centres, is being treated as a safe bet. But remember the telecom capex boom of the early 2000s. The infrastructure providers did not create the wealth. Google did. The verticals always create more value than the rails."

The implication for Asia's AI investment strategy is significant. The organisations that will define the next decade are unlikely to be those that built the most compute capacity or the largest data centre footprints. They will be those that built the most extraordinary applications on top of that infrastructure: India-specific AI tools, vernacular products, sector-specific intelligence systems that no global platform has bothered to design.

Tarun Anand of Jakson Group grounded this in an important historical frame: "AI is not a new concept. It dates to 1955. What has changed is the data." What that change demands of leaders is not more technology investment. It is clearer thinking about what problem the technology is being asked to solve.

The Clearest Thinkers Will Win

The defining divide in Asia's AI economy may not be between organisations with large budgets and those without. It may be between organisations that started with a clear problem and those that started with a technology announcement.

As Rupessh Goel, Managing Director at First Citizens India, observed in Hyderabad: "A lot of AI and ML conversations happen to reduce cost. But that is the problem. It should be about decision velocity: how many decision points have you been able to reduce? The problem with India-related GCCs is running after cost rather than reducing decision nodes."

The organisations that will look back on 2026 as their inflection point are those that resisted the pressure to move fast and chose instead to move with intent. Because in the AI economy, the greatest competitive advantage may not be the organisation with the most compute. It may be the one that asked the right question first.