Anthropic told the Senate Banking Committee in a June 10 letter that roughly 25,000 fraudulent accounts tied to Alibaba’s Qwen lab ran 28.8 million queries against Claude between April 22 and June 5, a distillation campaign the company describes as the largest it has ever caught. The letter, addressed to Chairman Tim Scott and Ranking Member Elizabeth Warren and first reported by Bloomberg on June 24, frames the activity as state-adjacent extraction rather than ordinary terms-of-service abuse.
The numbers do the structural work. In February, Anthropic disclosed campaigns by DeepSeek, Moonshot AI and MiniMax that together accounted for roughly 24,000 accounts and 16 million exchanges. The Alibaba operation alone, as Anthropic tells it, exceeds the combined total of those three on both axes. Operators allegedly used commercial proxy services to defeat the geographic restrictions that bar Chinese entities from Claude, and concentrated their queries on the model’s most commercially valuable skills: agentic reasoning, software engineering, and long-horizon task completion.
Sarah Heck, Anthropic’s head of policy, wrote that the campaign was “carried out illicitly, systematically, and at industrial scale to harvest U.S. AI capabilities across frontier labs and repackage them as their own without incurring the training and R&D costs required to train U.S. frontier models.” The letter argues the practice converts “billions of dollars in American investment and R&D into a massive subsidy for our geopolitical competitors,” and accuses Alibaba of having “ignored the Trump Administration’s warnings.”
The political scaffolding was already in place. On April 24, the White House Office of Science and Technology Policy, under director Michael Kratsios, issued a memorandum pledging to share foreign-distillation intelligence with US labs. The Pentagon added Alibaba to its Chinese military companies list on June 8; Alibaba has sued the Defense Department to be removed, calling the designation baseless. Two days after Anthropic’s letter, on June 12, Commerce Secretary Howard Lutnick’s department restricted access to Anthropic’s Mythos and Fable models, including by foreign nationals on staff. Anthropic disabled both globally to comply.
Legislative momentum is following. Senators Bill Hagerty and Andy Kim are preparing an amendment to must-pass defense legislation that would sanction or blacklist Chinese firms improperly accessing US AI outputs, with a bipartisan House companion from Bill Huizenga and Sydney Kamlager-Dove moving in parallel. Anthropic, for its part, is asking the administration to clarify antitrust rules so labs can share distillation intelligence with one another. Alibaba didn’t respond to requests for comment.
Read against Anthropic’s confidential IPO filing and a reported $965 billion valuation off the $65 billion Series H, the letter is also a positioning document. It reframes a competitive problem, Chinese open-weights models rapidly closing the capability gap, as a national-security one, with the Banking Committee as the venue and export controls as the remedy. The Cold War economic statecraft of the 1980s Toshiba-Kongsberg episode worked the same way: a commercial leak became a sanctions case once the right committee took the file.
Sources
- https://www.cnbc.com/2026/06/24/anthropic-alibaba-distillation-campaign.html
- https://thenextweb.com/news/anthropic-accuses-alibaba-distillation-claude-qwen
- https://www.techspot.com/news/112897-anthropic-accuses-china-alibaba-stealing-claude-ai-capabilities.html
- https://www.techtimes.com/articles/319105/20260625/alibaba-ran-largest-known-ai-theft-campaign-against-claude-anthropic-tells-senate.htm
- https://cybernews.com/news/anthropic-alibaba-industrial-scale-ai-extraction-mythos-rival/