Washington export curbs on Anthropic models lift competitive pressure from Chinese AI

Traders Union
All News
Financial News
Washington export curbs on Anthropic and Chinese AI

Highlights

U.S. export controls abruptly cut off Anthropic’s latest AI models globally, heightening commercial risk for companies relying on American AI providers.
Chinese AI models like DeepSeek’s V4 Pro offer roughly 60 times lower pricing than Anthropic and captured four of the five top OpenRouter spots in early June.
Increased adoption of Chinese AI models may shift global technical standards and raises valuation concerns for Anthropic, OpenAI, and SpaceX due to reduced international growth prospects.

Export ban raises reliability concerns

As the Financial Times reports, the Trump administration imposed export controls on Anthropic’s latest frontier AI models last Friday, forcing the company to cut off global access after a notice sent at 5.21pm that Anthropic says stems from a misunderstanding over security concerns.

That abrupt process creates a commercial warning for companies deciding which AI systems to adopt. If access to a leading U.S. model can be withdrawn on short notice, businesses may see greater risk in building products and infrastructure around American providers.

The article argues that export controls are a poor fit if the concern is misuse by foreign users, because nationality does not by itself determine risk. It also notes that such restrictions are most effective when the affected product is difficult to replace, while AI models can often be substituted more easily than tightly controlled chipmaking equipment.

Chinese rivals gain pricing and scale advantage

Chinese models including DeepSeek’s V4 Pro, GLM, Kimi, MiniMax and Qwen are increasingly approaching the performance of U.S. competitors, while also offering practical advantages for customers abroad. Many are self-hostable, allowing companies to run them on their own servers, and DeepSeek’s flagship model costs 87 cents per 1 million output tokens, about 60 times cheaper than Anthropic’s Fable 5.

Adoption data cited in the article suggests the shift is already visible. In the first two weeks of June, four of the five most popular models on OpenRouter were Chinese, and among the global top 20, Chinese models processed twice as many tokens as U.S.-developed rivals.

Broader use of Chinese AI could help Beijing influence global technical standards as developers build tools and workflows around those systems. The episode also raises valuation questions for Anthropic, OpenAI and xAI parent SpaceX, because limits on overseas sales would weaken the international growth assumptions supporting their high market worth.

Our earlier report on the U.S. export-control order on Anthropic’s Fable 5 and Mythos 5 explained that the company was instructed to suspend access for foreign nationals on national-security grounds, prompting plans to disable availability more broadly to ensure compliance. We also noted the government’s concern about a potential jailbreak risk in the models’ safeguards and Anthropic’s efforts to engage U.S. officials to resolve the restrictions.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.