Asia Could Win the Next AI Race Without the Biggest Models: DBS Report

You have probably heard the argument that whoever builds the largest language model wins the AI race. DBS Research has published a different view. In Intentional AI: Connecting Intelligence, Capital and Outcomes, the bank's research arm argues that access to advanced AI capabilities alone will become a less sustainable competitive advantage. Instead, organisations will stand out through the problems they address and their ability to translate AI capabilities into practical outcomes. While the first wave of AI development focused on technological capabilities, DBS Research expects the next phase to place greater emphasis on real-world applications.

The shift matters because the geography of AI investment and the geography of AI deployment do not overlap. AI companies raised approximately US$270 billion in venture capital funding in 2025, according to the report. Nearly 80 per cent went to North America, while Asia received just 5 per cent. Despite attracting a smaller share of investment, Asia accounted for 74 per cent of global industrial robot installations in 2024. DBS Research sees this as an advantage for Asia as AI moves beyond prediction towards applications that can perform actions. The region's manufacturing capabilities, physical infrastructure and experience in technology deployment could support broader AI adoption. Rather than focusing solely on developing increasingly advanced models, the report suggests Asia could create value by applying AI in sectors where it can deliver measurable improvements.

The problems that create opportunities

DBS Research identifies ageing populations, healthcare access, climate resilience and financial inclusion as major challenges that could also create opportunities for businesses and investors. In financial services, the report argues that expanding access only creates lasting value when it improves financial health and people's ability to make economic decisions. Similarly, more accurate weather forecasts become valuable when they help communities and businesses respond more effectively to climate risks. These examples reflect the report's broader argument that the benefits of AI depend on how organisations put the technology into practice.

The framing inverts the usual venture narrative. Instead of asking what a model can do, the report asks what a sector needs and whether AI can deliver it. That reframing opens space for organisations that do not build foundation models but understand the operational context in which AI must perform. A bank that deploys thousands of models across hundreds of use cases learns different things from a lab that trains one very large model. DBS itself deployed more than 2,000 AI models across over 430 use cases in 2025, generating approximately S$1 billion in economic value that year, according to its annual report.

Singapore as integration hub

DBS Research also identifies an opportunity for Singapore to connect the region's strengths in innovation, manufacturing, technology deployment, finance and governance. Rather than competing on the scale of AI development, Singapore could help businesses bring together these capabilities and support the adoption of AI solutions across borders. The report suggests this role could help make Asian innovation more trusted, interoperable and attractive to investors.

The proposal is not about building the next OpenAI. It is about creating the conditions under which organisations can move AI from prototype to production, across regulatory environments, languages and industries. That requires infrastructure, yes, but also the kind of institutional knowledge that comes from deploying technology in messy, regulated, high-stakes environments. Banks, hospitals, manufacturers and governments all face constraints that research labs do not. The organisations that learn to navigate those constraints at scale build a different kind of moat.

What this means for L&D

If the next phase of AI competition turns on deployment rather than model size, then learning and development leaders need to prepare their organisations for a different set of capabilities. Technical literacy remains important, but so does the ability to identify high-value problems, translate business requirements into AI specifications and manage the organisational change that follows deployment. The report's emphasis on outcomes over capabilities suggests that training programmes should focus less on what AI can do in theory and more on how to integrate it into workflows, measure its impact and iterate based on results.

For organisations in Asia, the report offers a practical argument: you do not need to outspend Silicon Valley to compete. You need to outperform on deployment, application and the translation of AI into measurable business value. That is a race L&D can help win, by building the skills that turn intelligence into outcomes.

Sources:

• Fintechnews Singapore