Chinese AI caught up so fast that Anthropic wants government to step in

This is the part of the AI race that gets interesting.

The argument behind many U.S. AI valuations has been that American companies have a huge technology lead.

But now Chinese AI companies are challenging that assumption.

Chinese model Zhipu AI GLM 5.2 reportedly matched or exceeded Anthropic’s Claude/Mythos on certain cybersecurity benchmarks, including finding software bugs. Another Chinese cybersecurity AI tool from 360 Security made similar claims.

The bigger shock is how these models are being distributed.

Many Chinese open weight models can be downloaded and run by developers instead of requiring expensive subscriptions to closed systems.

That changes the competition.

If another country can build powerful AI models and make them widely available at lower cost, the “moat” protecting premium U.S. AI valuations becomes much harder to defend.

And this is where the controversy starts.

Anthropic CEO Dario Amodei has warned lawmakers that open-source AI could be dangerous and has pushed for restrictions.

That creates an uncomfortable question.

Is the government being asked to protect national security?

Or is it being asked to slow down a competitor that caught up faster than expected?

The answer depends on who you ask.

But the market question is much simpler.

If Chinese AI keeps improving while offering cheaper alternatives, how much are investors willing to pay for a U.S. AI advantage that may not stay exclusive?

The AI race may not be decided by who has the biggest model.

It may be decided by who can make powerful AI cheap and available to everyone.