Manus’s Singapore Shift: US Venture Money and the Splintering of China’s AI Scene

Manus rose quickly this year on Chinese social media and state television coverage, touting an agent capable of executing multi-step tasks and drawing on models such as Anthropic’s Claude and Alibaba’s Qwen. That technical mix particularly reliance on U.S. models - clashed with China’s tightening rules on generative AI services, complicating registration and content compliance. With U.S. venture money now in the cap table, the founders pivoted to Singapore, trimmed China staffing, and scrubbed public ties that could complicate operating in either jurisdiction.

Benchmark’s participation became a lightning rod for Capitol Hill and a test of the Biden-era framework that screens or restricts American capital in sensitive Chinese technologies. Treasury’s review of the Manus financing, while not itself a sanction, has had a chilling effect in Sand Hill Road as funds weigh whether governance rights, information access and IP location could trigger compliance issues. For start-ups, the episode spotlights a new diligence checklist: data flows, model provenance, compute sourcing, and the legal domicile of code and customers.

The longer-term consequence may be a two-track AI ecosystem in which companies choose either China-first or U.S.-aligned footprints early, with Singapore and other neutral hubs catching the spillover. Manus’s trajectory - fast local traction, foreign capital, then a hurried redomicile, suggests the window for straddling both systems is narrowing. As national security concerns bleed into venture financing, deals that once hinged on product-market fit now rise or fall on regulatory fit across two increasingly incompatible regimes.