Trump Defends Open AI While Accusing China of Stealing It

The Trump administration this week accomplished what few political operations manage: staking out two completely irreconcilable positions on the same issue and calling it strategy. According to Axios, the White House accused Chinese companies of stealing U.S. AI technology while simultaneously defending open AI development—a rhetorical maneuver so transparently self-serving it deserves its own place in the Smithsonian alongside other historical artifacts of cognitive dissonance. The distinction, we're told, could become the foundation for U.S. action against Chinese AI companies, which is a polite way of saying: the rules are whatever we decide they are on any given Tuesday.

Let's parse the actual logic here, because it's genuinely instructive in understanding how American AI policy has become indistinguishable from a mid-budget heist film. The administration wants to preserve an "open" AI ecosystem—presumably the one where Meta, OpenAI, and Google release model weights, allow researchers to tinker, and democratize access to foundational technology. Simultaneously, it wants to prosecute Chinese companies for accessing that same open technology and building competitive systems from it. This is like leaving your house unlocked, publishing your security code in the New York Times, then calling the police when your neighbor uses it to enter.

The historical precedent here is worth examining. American tech policy has long oscillated between "open source is the future" and "only American companies should profit from openness." We've seen this exact dance before with semiconductors, 5G infrastructure, and cloud computing—sectors where U.S. dominance required both the appearance of freedom and the reality of gatekeeping. The Trump administration isn't inventing hypocrisy; it's simply being more honest about what U.S. tech policy has always been: a distinction between allies who get to compete (within limits) and adversaries who get sanctioned for doing the same thing.

What makes this particular moment satirically rich is the administration's faith that geopolitical theater can substitute for actual policy coherence. "IP theft" is a real problem; Chinese companies have indeed engaged in systematic IP acquisition. But the White House seems to believe it can simply redraw the definition of theft along nationalist lines: if Americans do open AI development, it's innovation; if Chinese companies do it, it's espionage. The distinction could become the foundation for action, as Axios notes—which is a delicate way of saying: we'll use whatever legal language we need to punish companies we've already decided to punish.

The venture capital and M&A implications are straightforward and grim. U.S. AI companies will enjoy a protected domestic market, regulatory tailwinds, and the implicit promise that Chinese competitors will be regulated out of contention. This creates moral hazard on a spectacular scale: why build genuinely defensible AI products when government policy guarantees market protection? The short-term funding environment for AI startups will remain robust, but only for the right zip codes and the right backgrounds. Innovation moves to whoever can afford to wait for policy to render the playing field uneven.

This is the future of tech policy in a decoupling era: not coherent principle, but coherent tribalism dressed up as principle. The administration will announce protections for American AI innovation while China does the same thing. Both will use open-source rhetoric when convenient and nationalist rhetoric when profitable. Meanwhile, actual engineers and founders will optimize for whichever government is most willing to subsidize their ambitions, and we'll all pretend this is about competition rather than the mundane reality that power writes the rules and calls it justice.