When a government decides a deal won’t go through, a certain silence descends. There won’t be a press conference or an outright ban; instead, the regulator will wait a few weeks, ask for additional documents, and then wait a few more weeks. That’s how Beijing handled Meta’s $2 billion offer for Manus, the Chinese agentic AI startup that was one of the most anticipated names in the industry until recently. The agreement didn’t fall through in court. It was just left to suffocate.
It’s difficult to ignore how well this aligns with Beijing’s other recent actions, which include discussing protecting the Chinese economy from outside shocks virtually every day. These appear to be two distinct stories, one about macroeconomics and the other about technology. They’re not. They are two distinct registers of the same posture.
| Subject Profile | Details |
|---|---|
| Topic | China’s parallel strategy on AI deals and economic resilience |
| Central Event | Beijing blocks Meta’s $2 billion acquisition of Manus, a Chinese agentic AI startup |
| Key U.S. Response | Trump administration memo (April 24, 2026) accusing Chinese firms of “distilling” American AI models |
| U.S. Official Cited | Michael Kratsios, White House director of Science and Technology Policy |
| Chinese Position | Opposes “unjustified suppression” — statement from China’s embassy in Washington |
| Performance Gap | U.S.-China gap on top AI models has effectively closed, per Stanford’s HAI report |
| Strategic Frame | Export controls on chips slowing but not stopping Chinese progress |
| Broader Context | Both governments now exploring AI safety guardrails despite open rivalry |
| Year of Escalation | 2026 — the year the rivalry shifted from chips to entire deal flows |
What China won’t give up is revealed in the Manus block. Beijing has determined that it is unacceptable, regardless of cost, for an American company to own a Chinese frontier-stage builder of agentic AI, which is software that can plan, act, and carry out tasks without constant human prompting. Speaking with those who adhere to Chinese tech policy, there is a feeling that the cost nearly made matters worse. Two billion dollars suggested that the asset was genuine, valuable, and marketable. Beijing responded, “Not this one.”
The second front comes next. Recent communiqués from the Politburo have been remarkably candid about getting ready for “external shocks,” a phrase that seldom occurs by accident in Chinese policy circles. Tools for stimulation are being arranged. There is a covert restructuring of local government debt. Subsidies for consumption are being expanded. The leadership appears to be preparing for the possibility that trade and technology relations with Washington will deteriorate before improving.
Investors appear to think that both actions are defensive, but that interpretation is incomplete. In addition to being defensive, blocking Manus is an offensive wager that China can create these systems on its own schedule and without Western funding diluting local champions. Beijing appears to have gained confidence that it did not have eighteen months ago after Stanford’s recent HAI report revealed that the top-model performance gap between the United States and China has essentially closed.
The response from the Trump administration, a memo written by Michael Kratsios accusing Chinese companies of “distillation” of American models on an industrial scale, is a reflection of the same fear. Washington wants to prevent the leakage of capability. Beijing aims to prevent ownership from seeping in. Just from different sides of the same wall, both governments are tightening the perimeter.
Observing this unfold, it’s remarkable how infrequently either side now uses the more traditional language of cooperation. Chris McGuire of the CFR has contended that loopholes, smuggling, and cloud rentals have caused China to slow down rather than stop chip export regulations. Beijing learned not to give up from that. Its purpose was to ensure that its own crown jewels were never lost in the first place.
When a real downturn tests the economic cushion, it’s still unclear if this strategy will hold. It is inexpensive to protect a startup. It is another matter entirely to defend a slowing economy while simultaneously repelling Western investment. However, Beijing seems to have concluded that economic resilience and AI sovereignty are not mutually exclusive objectives. They are identical, but they have been written twice.
