Washington's move to restrict foreign access to Anthropic's Claude models backfired within days. The export control, designed to preserve America's AI dominance, instead accelerated investment in competing systems elsewhere.

Cohere reports surging government interest following the restrictions. DeepSeek closed a $7.4 billion funding round, one of the largest ever for a Chinese AI lab. Chinese competitors simultaneously slashed token prices by up to 99 percent, making their models far cheaper than American alternatives for international customers.

The policy targeted Anthropic specifically by blocking overseas users from accessing its most capable models. Officials intended this to prevent adversarial nations from accessing frontier AI capabilities. The outcome contradicts that logic. Rather than consolidating America's advantage, the restrictions pushed developers and companies toward Chinese options that remain openly available.

DeepSeek's funding round signals Beijing's commitment to closing the AI gap with American labs. The company now possesses capital comparable to what OpenAI and Anthropic raised in recent rounds. Aggressive pricing undercuts American providers on the international market, where most potential customers live outside the U.S.

A separate supply chain attack highlighted another vulnerability in AI infrastructure. Researchers discovered 144 poisoned npm packages that harvested credentials and API keys from developers. These packages targeted the AI ecosystem specifically, exploiting trust in open-source repositories to steal access tokens. The attack demonstrates how restrictions on one layer of AI development expose weaknesses in another.

The timing compounds the problem. As Washington attempts to control advanced models at the top, attackers compromise the foundational tools developers rely on. Poisoned packages can inject backdoors into AI systems before they reach production, potentially affecting both American and international deployments.

Export controls on finished AI models may prove impossible to enforce effectively given China's rapid progress and open-source alternatives. The strategy assumes America maintains an insurmountable technical lead. Current trends