OpenAI, Google Sell Advanced AI Models to Blacklisted Chinese Tech Giants

Silicon Valley heavyweights OpenAI and Google have provided advanced artificial intelligence models to Chinese technology conglomerates blacklisted by the Pentagon.

The transactions expose critical structural gaps in Washington’s ongoing geopolitical initiative to decelerate Beijing's strategic computing and defense development.

Both U.S. corporations confirmed they have delivered artificial intelligence services to the Singapore-registered subsidiaries of Alibaba, Baidu, and Tencent, according to the Financial Times.

The U.S. government has previously accused these specific entities of collaborating directly with the Chinese military.

Regulatory Deficiencies and Blacklist Limitations

While the transactions remain entirely legal under existing frameworks, the developments have reignited intense pressure for tighter U.S. regulation on software models.

Proponents argue these should mirror current export controls placed on advanced semiconductor hardware used to train powerful computing systems.

The U.S. government has actively moved to control direct access to specific frontier models, such as Anthropic’s Mythos and Fable, as well as OpenAI’s GPT-5.6.

However, Washington has stopped short of a comprehensive ban on the use of cutting-edge software by Chinese-headquartered organizations.

This includes those explicitly named on the congressionally mandated 1260H blacklist, which designates Chinese enterprises with suspected ties to the People’s Liberation Army.

Distillation Violations and Strategic Access

OpenAI recently suspended access to its application programming interface (API) for Alibaba-affiliated users following concerns over illicit activity.

The system interface allows developers to utilize AI models remotely.

The platform flagged suspected "distillation" to the U.S. government, a prohibited practice where third-party developers extract outputs from premier models to systematically upgrade competing national systems.

OpenAI confirmed that while access is barred in China, it permits select Chinese-owned entities to use its services in foreign jurisdictions where safety monitoring and distillation safeguards can be enforced.

The company defended its stance by stating a preference for global AI adoption to be shaped by democratic values rather than autocratic governance, noting that nationality alone should not dictate access.

Enforcement Fractures and Corporate Responses

Google stated its artificial intelligence tools remain available to corporate clients in Hong Kong and Singapore under strict usage policies that forbid distillation.

However, the tech giant conceded that geographic sales barriers are insufficient to eliminate distillation risks, noting that sophisticated attackers can easily circumvent regional restrictions.

In contrast to its industry peers, Anthropic has instituted a blanket prohibition preventing Chinese companies and their foreign-owned subsidiaries from deploying its models.

Even so, enforcing these restrictions has proven highly complex.

Anthropic recently moved to seal loopholes used by Chinese entities to bypass unauthorized usage barriers, having previously accused Chinese labs DeepSeek, Moonshot, and MiniMax of distillation.

Anthropic recently informed Congress that Alibaba deployed 25,000 fraudulent accounts to execute more than 28.8 million exchanges with its Claude model, violating its terms of service.

Baidu declined to comment on the developments, while Tencent and Alibaba did not respond to queries.

Alibaba previously petitioned a U.S. court to remove it from the 1260H blacklist, labeling the Pentagon's classification arbitrary.

Experts warn that allowing Chinese laboratories to systematically extract frontier capabilities without absorbing the immense engineering, safety, and computing costs undermines the economic baseline of U.S. technology leadership.