Beijing is moving to require government approval before leading Chinese AI startups can accept US capital, a direct response to Meta's $2.5 billion acquisition of Manus that has shaken the assumptions of every foreign investor active in China's technology sector.
The policy shift, first reported by Bloomberg on Friday, is sweeping in scope and represents a significant escalation in the ongoing tech cold war between Washington and Beijing. Chinese regulators including the National Development and Reform Commission have already begun instructing specific private tech companies to reject US investment in active funding rounds unless they receive explicit government clearance. This is not a subtle shift in bureaucratic posture. It is a firm wall going up around some of the country's most promising technology companies. AI startups Moonshot AI and StepFun are among those that received the guidance directly, effectively putting their current funding ambitions on hold. ByteDance, the owner of TikTok, has been told that secondary share sales to US investors also require approval, adding another layer of friction to an already complicated ownership structure. The overarching intent, according to people familiar with the matter, is to prevent US investors from building stakes in sectors Beijing considers sensitive to national security. This approach mirrors similar moves in Washington, where American officials have aggressively worked to block Chinese capital from accessing domestic AI advancements. The global investment landscape for artificial intelligence is effectively splitting into two highly guarded, heavily regulated zones.
The trigger for this aggressive financial lockdown is Meta's acquisition of Manus, the AI agent startup founded by Chinese engineers and originally headquartered in Singapore. Meta announced the deal in December 2025 at roughly $2.5 billion, making it one of the largest acquisitions of a China-linked AI company by a US tech firm. The deal caught the attention of executives in Silicon Valley and immediately set off alarm bells in Zhongnanhai. Manus had drawn global attention in March 2025 when it launched an autonomous AI agent capable of building websites and handling coding tasks independently, proving that cutting edge innovation was not solely the domain of American tech giants. By December it had crossed $100 million in annual recurring revenue. The speed of its rise and the terms of its sale alarmed Beijing, exposing a massive gap in China's regulatory framework. The Chinese government clearly did not anticipate a domestic startup scaling so rapidly and selling to an American competitor before regulators could even assess the strategic implications.
China's response has been methodical and increasingly pointed since the magnitude of the deal became clear. Since January 2026, the Ministry of Commerce has been reviewing whether the Manus deal violates Chinese technology export control laws, which require government approval for the export of interactive AI systems. Those same controls are exactly what Beijing used to claim oversight over TikTok's US operations during Donald Trump's first term. Beijing is now repurposing a familiar legal playbook to assert sovereignty over advanced algorithms and AI models. In March, regulators summoned Manus co-founders Xiao Hong and Ji Yichao to a meeting with NDRC officials in Beijing. Following that session, both were informed they should not leave the country while the review continues. As Reuters confirmed, the restriction was framed as guidance rather than a formal travel ban, but the distinction is largely academic when your passport is effectively frozen and your movements are entirely restricted by the state.
The New York Times reported that the Chinese government's scrutiny may serve as a deliberate signal to the broader AI startup community: relocate your operations abroad without regulatory clearance and face consequences that go beyond fines. One analyst cited in the coverage described the measures as creating a chilling effect on cross-border deals, making founders think twice before accepting foreign money or attempting to set up offshore holding companies. For venture capitalists outside of China, the message is equally blunt. Investing in Chinese AI talent now carries a level of geopolitical risk that few private equity funds are willing to underwrite. The era of open cross-border startup building is over.
Also read: DeepSeek V4 arrives as a trillion-parameter open-source challenge to Silicon Valley's AI elite • Starlink's billions are quietly bankrolling the AI race from orbit • A 21-year-old cyberweapon called Fast16 just rewrote the history of state-sponsored sabotage
