Anthropic sent a letter on June 10 to the U.S. Senate Committee on Banking, Housing, and Urban Affairs – addressed to senators Tim Scott and Mark Warner – accusing operators affiliated with Alibaba and its AI lab Qwen of conducting what the company calls the largest known distillation attack on Anthropic to date. The campaign, which Anthropic says ran between April 22 and June 5, generated approximately 28.8 million exchanges with Claude through nearly 25,000 fraudulent accounts. Distillation is the practice of training a less capable AI model on the outputs of a stronger one, allowing a competitor to harvest reasoning patterns, coding capabilities, and agentic skills at a fraction of the cost required to develop them independently. Alibaba, which was added to the Pentagon’s Chinese military companies list two days before Anthropic sent the letter, has sued the Defense Department seeking removal from that designation and has not publicly responded to the Anthropic accusation.
The scale of the Alibaba campaign relative to what came before it is the specific number that provides the story’s analytical frame. In February 2026, Anthropic disclosed that DeepSeek, Moonshot AI, and MiniMax had conducted distillation campaigns involving roughly 16 million exchanges across approximately 24,000 fraudulent accounts. The Alibaba campaign exceeded the combined February total in a single six-week period: 28.8 million exchanges through 24,939 accounts, targeting specifically Claude’s software engineering and agentic reasoning capabilities. Anthropic said in the letter that the attack was designed to help accelerate China’s ability to reach the capabilities of its most advanced Mythos Preview model. The comparative scale against the February campaigns is what NewsTrackerToday opens on as the data point that shifts this from an ongoing compliance problem into a discrete escalation.
Daniel Wu, who covers geopolitics and energy, places the distillation dispute in a strategic historical sequence: “Every major technology transfer dispute in the post-World War Two era has had a legitimate research version and an industrial-scale theft version. The FISA court intercepts that exposed Soviet semiconductor reverse-engineering in the 1980s looked exactly like this: technically legal at each individual step, operationally a coordinated extraction program. What is different now is the transparency. Anthropic can document every exchange, every account, every timestamp. That documentation is why this went to a Senate committee rather than a company blog post.” The White House Office of Science and Technology Policy issued a memorandum in April 2026 pledging to help AI companies detect and coordinate against industrial-scale distillation. Anthropic wrote that Alibaba “ignored the Trump Administration’s warnings” in proceeding with the campaign after that memo.
Sophie Leclerc, who covers the technology sector, reads the technical and legal complexity: “Distillation as a technique is not inherently illegitimate. AI labs, including American ones, routinely use model distillation in research and product development. What Anthropic is characterizing as illicit is the specific combination of industrial scale, systematic use of fraudulent accounts to bypass rate limits and safety protocols, and the targeted extraction of commercially sensitive capabilities. The hypocrisy accusation that Elon Musk made – that Anthropic itself trained on web-scraped data contested in copyright suits – complicates the moral framing but does not address the specific conduct described: nearly 25,000 fake accounts running 28.8 million queries in six weeks is not a research program.” The timing and the IPO context are what NewsTrackerToday traces as the letter’s commercial dimension.
Alibaba’s American depositary receipts fell more than 3% on the news, dropping below $100 in afternoon trading. Senators Bill Hagerty and Andy Kim are preparing an amendment to must-pass defense legislation that would authorize sanctions or blacklisting for Chinese firms found to be improperly accessing U.S. AI model outputs. A related bipartisan House bill from Representatives Bill Huizenga and Sydney Kamlager-Dove is also in consideration. The legislative momentum sits alongside the irony that Anthropic, which is asking Washington to protect its AI capabilities, is simultaneously in a dispute with the same administration over the export control order that restricted access to Fable 5 and Mythos 5. Asking the government to enforce against Chinese distillation while fighting a different government enforcement action is the specific contradiction that News Tracker Today stays with as the contextual frame for Anthropic’s Senate letter.
Three things to watch as the distillation dispute develops: whether the Hagerty-Kim amendment survives the legislative process and reaches the president’s desk, which would create a formal legal mechanism for sanctions against future distillation campaigns; whether Alibaba formally responds to Anthropic’s letter, either denying the conduct or offering a characterization of the exchanges as legitimate research that Anthropic’s systems authorized; and whether the export control order on Fable 5 and Mythos 5, which Anthropic has been negotiating to modify, gets resolved in a way that affects the company’s credibility as a petitioner for government protection against Chinese distillation – since it is difficult to argue the government should protect your models while simultaneously disputing that same government’s restrictions on those models. The legislative question is what NewsTrackerToday puts as the variable the Senate committee will answer.
