Singapore: A US–China AI Transit Hub

Neutral ground, big ambitions

Singapore is positioning itself as a neutral, well-funded transit hub for artificial intelligence between the United States and China. Oxford statistician Zheng Yuhuai (郑宇怀) told a Singapore audience that when geopolitics grows turbulent, the city-state can offer “a neutral and stable haven” for AI development. According to The Observer’s global AI index, Singapore ranks third worldwide on talent, infrastructure, research and government strategy—behind only the US and China. Can a small city-state sit between two superpowers and stay useful without taking sides?

Money, startups and relocations

The government is backing that ambition with cash: Singapore’s Ministry of Communications and Information announced plans to invest more than S$1 billion (about US$780 million) in public AI research from 2026–2030, funding education, teacher training and industrial applications. Industry data from Google, Temasek (淡马锡) and Bain show Southeast Asia hosted roughly 680 AI startups and attracted over US$2.3 billion between mid‑2024 and mid‑2025; Singapore alone reportedly has 495 AI startups with about US$1.31 billion in financing. It has been reported that the AI startup Manus was sold to Meta and moved to Singapore, and that MiroMind, an AI lab affiliated with Shanda (盛大), asked some Shanghai staff to relocate to the city-state. Separately, it has been reported that Micron Technology (美光科技) plans a US$24 billion expansion of its Singapore memory facilities over the next decade.

Talent gaps and a conservative culture

But money and companies don’t erase a different bottleneck: local talent. Entrepreneurs interviewed for this report — including Markus, a Silicon Valley founder and former ByteDance (字节跳动) team lead; Liu Yan (刘岩), serial entrepreneur; and investor Lü Weisheng (吕伟胜) — all point to a shortage of home‑grown algorithm engineers and a culture that favors stability over the aggressive risk‑taking that built major AI clusters in Shenzhen or Silicon Valley. “You can recruit algorithm engineers, but most are brought in from abroad or from multinational offices,” Liu said. Singapore excels at “AI+” deployments—using AI to augment existing services—but it has fewer genuine AI‑native startups or locally trained model builders to push frontier research.

A deliberate, non‑aligned strategy

There is a geopolitical logic to Singapore’s approach. It wants access to both American and Chinese AI ecosystems without becoming a battleground. It has been reported that Singapore politely rejected proposals to become a national champion in large‑model development, preferring to remain a user and integrator rather than pick a side in an AI arms‑race. The result: a bustling, well‑funded middle layer that can move personnel, capital and products between East and West—but not yet the birthplace of many global breakthrough models. The question now is whether Singapore will remain a durable “buffer” and transit hub, or whether it can build the local depth needed to graduate from middleman to originator.