Singapore's SWITCH 2026 AI push: 24 'Champions of AI', S$150M for AI materials discovery, and agentic AI for small business

Singapore used the opening of its flagship tech festival on October 27 to lay out the next phase of its AI strategy — and the through-line was not bigger models, but wider use. Speaking at SWITCH 2026, Deputy Prime Minister Gan Kim Yong announced a first cohort of 24 companies for the country's Champions of AI programme, a S$150 million platform for AI-led materials discovery, and wider practical support to get small businesses from AI experiments to daily operations.

The package is one argument made three ways: frontier research should reach the factory floor, large companies should transform whole businesses rather than run pilots, and the smallest firms should not be left to figure out agentic AI alone. It lands in a year when Singapore has already staked out four sectoral AI missions — in advanced manufacturing, finance, healthcare and connectivity — announced earlier in 2026 by Prime Minister Lawrence Wong.

MARVEL-AI: S$150 million to close the lab-to-factory gap#

The most concrete cheque in the package goes to materials science. The Materials Accelerator for Research and Venture Enterprise Launchpad — MARVEL-AI — will start with S$150 million in government funding, led by the Agency for Science, Technology and Research (A*STAR) and backed by the trade ministry, Enterprise Singapore, the Economic Development Board and the National Research Foundation.

The design pairs two halves that usually sit apart: AI models identify promising materials and design the experiments, while self-driving laboratories synthesise and test them quickly. Gan cited Google's Project GNoME and Microsoft's MatterGen as evidence that AI can already speed up discovery — and pointed to the harder problem that follows, turning a promising material into something that can be made, validated and sold at scale for semiconductors, energy and chemicals. MARVEL-AI is meant to close that gap by connecting researchers with companies, including startups and smaller firms.

Champions of AI: 24 companies signed up for the deep version#

The second announcement is about the demand side. After six months of engagement with companies and industry experts, Enterprise Singapore has named the first 24 companies in its Champions of AI initiative, spanning several sectors. Participants reported by Enterprise Singapore include DBS, OCBC, UOB, ST Engineering and Singapore Airlines, with more companies expected to join progressively.

The pitch is deliberately demanding. A pilot or a single AI tool is, in Gan's framing, a useful starting point but not the point: the Champions are committing to enterprise-wide transformation, rethinking business processes and building new capabilities — and to the workforce side, developing staff, creating new jobs, redesigning roles and building AI skills. They are also expected to support their smaller suppliers and share what they learn, so firms further down the supply chain can see what real transformation involves. Whether that diffusion happens is the initiative's real test — 24 transformed companies are a showcase, not an economy.

The other end of the ladder: agentic AI for small firms#

The third strand aims at firms that cannot hire an AI team at all. Through Catalytic AI Solutions, Enterprise Singapore is funding shared solutions for problems that recur across an industry, so companies with similar needs can adopt the same answer rather than each building their own. In logistics, AI can allocate jobs across fleets and optimise routing; in wholesale trade, it can help rice traders make better procurement and inventory decisions — a commercial gain that also touches Singapore's food security.

Alongside those shared solutions, Enterprise Singapore's hands-on agentic AI workshops have helped close to 300 micro-enterprises and small businesses build working prototypes, using tools including ChatGPT, Claude and Manus. More than 100 AI tools have been added to the SMEs Go Digital platform, and a new Pre-approved Solutions Package Builder will let smaller firms configure solutions to their requirements. More workshops are planned for November, with further expansion in 2027, and Enterprise Singapore says it will explore support for the final stages of deployment — the step where prototypes most often stall.

Why it matters, and what to watch#

Step back, and the strategy is a bet that the AI race will be won as much on adoption as on invention. Singapore cannot outspend the largest labs or states on frontier models, and it is not trying to: the SWITCH announcements assume the models largely exist already, and focus public money on the unglamorous middle — validation, integration, workforce redesign and shared infrastructure — where productivity is actually decided.

Three things are worth watching. First, whether the Champions cohort grows beyond its first 24 and produces measurable transformation rather than announcements. Second, whether MARVEL-AI produces validated materials and new ventures, or just papers. And third, whether the small-business push crosses the last mile from prototype to deployment, the gap Enterprise Singapore itself flagged when it cited shortages of technical expertise and time. Singapore's bet is that a country can lead in AI by being the place where AI actually gets used — and the next year will show whether diffusion, unlike a model launch, can be delivered on schedule.

Sources: Singapore's Ministry of Trade and Industry — Deputy Prime Minister Gan Kim Yong's delivered speech at the SWITCH 2026 opening ceremony; ASEAN Coverage's report on the announcements; and Tech in Asia's report on the MARVEL-AI funding. Photographs via Wikimedia Commons, credited in the captions above.