Singapore positions itself as neutral ground for AI firms
TL;DR:
- Anthropic, OpenAI, Meta’s Superintelligence Labs and Google DeepMind have all built or planned Singapore offices, joining a wave of Chinese AI startups distancing themselves from Beijing’s scrutiny.
- Singapore is offering AI-talent visas, employment passes inside three days in some cases, and tax breaks for IP registration; about 50 Chinese AI-related firms have set up via one corporate-services provider since 2024.
- For UK firms watching the Sino-US AI rivalry, Singapore’s positioning is a useful comparator — and a reminder that “neutral hub” pitches will increasingly compete with the UK’s own.
Singapore is rapidly transitioning from East-West gateway to declared neutral ground for AI firms caught between Washington and Beijing, according to Reuters. The story details how Chinese startups are using Singapore residency to escape government attention and political-bias perceptions, while US labs are using it to hire foreign talent without H-1B visa friction in the second Trump administration.
The visible pattern is hyperscaler concentration. Anthropic, whose recent $30 billion fundraising was led by Singapore sovereign-wealth fund GIC, plans to open a Singapore office. Anthropic joins OpenAI, Meta’s Superintelligence Labs and Google’s DeepMind in the country, plus operating AI companies including automation platform Workato, wealth-management developer Addepar and note-taking device maker Plaud AI. Legal-AI platform Harvey AI is set to follow in June.
Chinese AI’s quiet relocation
The China-side movement is more discreet. Two former Alibaba executives establishing the AI-video business Topview were advised that locating in Singapore would reassure international clients wary of Chinese government oversight; Topview has received over $8 million from Kamet Capital since 2024. About 50 Chinese AI-related firms have used one Singapore corporate-services provider, Link-da, to set up since 2024. Beijing has pushed back: founders of Manus, an AI startup that relocated from China to Singapore before being acquired by Meta, were reportedly subjected to a travel ban; Shanda’s MiroMind was told not to send talent abroad after expanding into Singapore, Japan and the US.
The trade-off, according to Insignia Ventures’ Tan Yinglan, is genuine: a Chinese founder genuinely escaping Beijing’s reach in Singapore must drop their Chinese passport, employ no engineers in China, and keep revenue, data and headquarters out of China. Singapore is also exposed to retaliation: NUS political scientist Chong Ja Ian warned that as Washington and Beijing demand cleaner separation between their tech stacks, Singapore “is seen as a grey space for technology transfers” and “could result in restrictions being placed on Singapore.”
Looking forward
For the UK, Singapore is the live test case for whether a small, well-regulated, English-speaking jurisdiction can pitch itself as the natural home for AI firms hedging Sino-US risk. Some of the same firms now opening in Singapore — Anthropic, OpenAI — already have major London footprints and dual-hyperscaler funding. The competitive question for UK policymakers, as ministers reportedly resist EU-rule alignment specifically to keep AI investment, is what concrete edges (talent visa speed, IP-tax treatment, public-procurement friction) the UK is willing to match. Both jurisdictions can credibly claim to be neutral — but only one tends to win each individual lab’s office decision.