China flexes AI and auto power—Meta’s Manus deal blocked as DeepSeek’s sequel turns open models into leverage

China’s AI and industrial competition narrative is tightening on multiple fronts after a sequence of high-signal moves. On Monday, Beijing blocked Meta’s $2 billion acquisition of Manus, an AI start-up founded in China, and the intervention followed earlier government flags about an investigation. The central government also barred Manus’s two founders from leaving the country, signaling that the review was not merely procedural. Separately, the New York Times reports that China has made the inner workings of a DeepSeek AI model open to all, framing it as a potential soft-power win. Strategically, the Manus block reads like a targeted assertion of control over frontier AI assets at the exact moment global tech firms are racing to secure model capabilities and talent. By stopping a major US-linked buyer from consolidating a Chinese-founded AI company, Beijing is effectively drawing a line around what it considers sensitive technology and cross-border ownership. This also sets a tone ahead of the Xi–Trump summit referenced by SCMP, where “shows of strength” can be used to shape bargaining space on investment screening, export controls, and AI governance. The auto-industry angle—foreign carmakers trying to remake themselves in the image of ascendant Chinese competitors—adds a parallel industrial contest in which China’s scale, supply chains, and software-defined vehicle approach can pressure Western margins and market share. Market and economic implications are likely to concentrate in AI infrastructure, cross-border M&A, and semiconductors tied to model training and inference. A blocked $2 billion deal can reverberate through AI venture funding expectations, deal certainty, and the risk premium applied to China-linked AI assets; it also reinforces the probability of tighter regulatory friction for US tech investors. In equities, the most visible sensitivity would be in AI software and cloud-adjacent names, plus semiconductor and networking suppliers exposed to China demand, where sentiment can swing on policy headlines. On the industrial side, the push by foreign automakers to emulate Chinese strategies can intensify price and feature competition, pressuring European and US auto suppliers while potentially benefiting Chinese component ecosystems and battery/material supply chains. What to watch next is whether Beijing expands the Manus-style review into broader AI M&A and founder-exit restrictions, and whether Meta or other foreign buyers adjust their deal structures accordingly. For the Xi–Trump track, monitor summit messaging for concrete language on investment screening, AI model governance, and reciprocal access for technology acquisitions. In parallel, track how widely DeepSeek’s open-model approach is adopted by developers and whether it triggers new export-control or licensing debates in Washington. Trigger points include additional founder travel bans, follow-on regulatory actions against other AI transactions, and any policy announcements that link AI openness to national security or industrial policy objectives.

Geopolitical Implications

• 01

  China is using regulatory control to shape ownership and flow of frontier AI capabilities, limiting US corporate consolidation.

• 02

  Open model internals can expand China’s technical influence while complicating Western governance and security efforts.

• 03

  AI and auto signals together point to a broader industrial contest where China converts scale and software advantages into leverage.

Key Signals

• —More founder-exit bans or M&A blocks involving Chinese AI start-ups.
• —Summit language on investment screening and AI governance enforcement.
• —Developer adoption of DeepSeek’s open releases and any resulting US licensing/export-control debates.
• —Auto policy and procurement moves that intensify pressure on foreign OEMs.