TL;DR
Anthropic accused Alibaba’s Gwen lab of using 25,000 fake accounts to exchange nearly 29 million Claudes, the largest campaign to date.
Anthropic has accused Alibaba of running its biggest-ever smear campaign against a US AI companytold senators and White House officials that operators associated with Alibaba’s Gwen AI lab used about 25,000 fake accounts to exploit Claude’s capabilities between April and June. The letter, a copy of which was seen by Bloomberg, described nearly 29 million exchanges with Claude, targeting software engineering and agent reasoning, the model’s most valuable commercial skills.
The indictment is the first time Anthropic has named a large Chinese tech conglomerate as the source of a distillation attack. Earlier allegations in February targeted smaller Chinese AI startups including DeepSeek, MiniMax and Moonshot AI, which Anthropic said collectively generated more than 16 million exchanges through about 24,000 fake accounts. The Alibaba campaign alone surpassed the combined volume of the previous three efforts.
Distilled is the practice of feeding carefully constructed queries to a frontier AI model, collecting its responses, and using those responses to develop a cheaper competing system that is close to the capabilities of the original. The White House flagged the technique as a national security concern in AprilWhen OSTP director Michael Kratsios released a memo committing his government to sharing intelligence on foreign distillation campaigns with US AI labs. Anthropic said in its letter that Alibaba’s campaign followed the Kratsios memo, ignoring the administration’s warnings.
💜 of EU technology
The latest rumblings from the EU tech scene, a story from our wise founder Boris and some questionable AI art. Free in your inbox every week. Register now!
Alibaba has not commented on the allegations. An Anthropic spokesman declined to discuss specifics, but stressed the importance of combating distillation through coordinated action between government and industry.
In midday trading on Wednesday, Alibaba’s American depositary receipts fell more than three percent to below $100 on the news. The stock drop adds to a difficult period for the company, which has faced pressure on multiple fronts in Washington.
The Pentagon added Alibaba to its blacklist of Chinese military companies on June 8Anthropic designation referred to in his letter. Alibaba sued the Department of Defense this week winning delisting, calling the label baseless and claiming no military affiliation. The distillery accusation now opens a second front, framing Alibaba not only as a company with alleged military ties, but also as an active participant in Anthropic’s systematic theft of America’s AI capabilities.
In its letter, Anthropic warned that rival distillation allows Chinese labs to replicate frontier AI at a fraction of the training cost, and that models built this way often lack safeguards. The company has called on the Trump administration to clarify antitrust rules so US labs can share more information about distillation attempts, reiterated its support for export controls on advanced AI chips and called for penalties against firms that use the technique.
Legislators act in parallel. Senators Bill Hagerty and Andy Kim plan to introduce an amendment to the pending defense legislation that would blacklist or sanction any Chinese firm found to have improperly accessed access to a US artificial intelligence model. A related bipartisan bill, sponsored by Reps. Bill Huizenga and Sydney Kamlager-Dove, is also under consideration in the House, though it’s unclear whether either proposal will survive the final version of the defense bill.
Timing is also sensitive to Anthropic. The company, valued at $965 billion after a $65 billion Series H, filed privately for an IPO this month and is preparing for a list that could come this fall. U.S. officials have estimated that unauthorized distillation is costing Silicon Valley labs billions of dollars and that cheaper imitation products from China threaten to crowd out customers, posing a financial risk to the company going public.
Anthropic’s calls for government support may not find a full audience, given that the company entered into a separate dispute with the Trump administration less than two weeks ago over export controls imposed on the Fable 5 and Mythos 5 models. Commerce Secretary Howard Lutnick signed an order barring foreign nationals from accessing these models, citing security concerns, and Anthropic stopped complying with them. Even after meetings between the company’s technicians and White House officials, little progress has been made toward restoring service.
The result is a company caught between two fronts. Anthropic needs the government to crack down on Chinese labs that extract its technology, but is also fighting the same government’s decision to restrict its own products. The letter to the senators is an attempt to separate the two issues, arguing that protecting US designs from distillation and allowing commercial use of those designs are complementary, not conflicting, goals.
Whether Washington agrees will shape both the US regulatory environment for AI companies and the competitive dynamics of the industry’s most consistent competition. Anthropic has now named four Chinese labs as distillers of its technology, with Alibaba accusing it of being the largest scale. If the legislative proposals take hold, the implications could extend well beyond Anthropic’s models to the broader question of how the U.S. enforces the intellectual property boundary around AI systems that exist as software rather than hardware and can be copied over the internet.
