Anthropic told the Senate Banking Committee on June 10 that operators linked to Alibaba’s Qwen artificial intelligence laboratory ran 28.8 million unauthorized exchanges against Claude through roughly 25,000 fraudulent accounts over six weeks, beginning April 22 and ending June 5. The company described it as the largest distillation campaign yet observed against a frontier American model, and framed the letter, addressed to Senators Tim Scott and Elizabeth Warren, as evidence that adversarial extraction has graduated from opportunistic scraping to something resembling state-adjacent industrial policy.

Sarah Heck, Anthropic’s head of policy, wrote that the operation was “carried out illicitly, systematically, and at industrial scale to harvest U.S. AI capabilities across frontier labs and repackage them as their own.” Per the letter, the campaign concentrated on agentic reasoning, software engineering and long-horizon task completion, with traffic routed through commercial proxy services to bypass restrictions barring Chinese entities from accessing Claude.

The sequencing matters. The April 22 start date sits roughly two weeks before the White House Office of Science and Technology Policy issued its April 24 memorandum committing the government to share intelligence with American laboratories about foreign distillation. February’s disclosure, naming DeepSeek, Moonshot and MiniMax across three earlier campaigns totaling about 16 million exchanges and some 24,000 accounts, now reads as the smaller precedent. Six weeks later, Anthropic was already documenting a campaign nearly twice the size.

The market read it quickly. Alibaba’s American depositary receipts fell more than three percent on Wednesday after Bloomberg first reported the letter, dropping below $100. Reuters and CNBC obtained copies shortly after.

The political machinery is already moving. Alibaba was added to the Defense Department’s list of Chinese military companies on June 8, and is now suing for removal. Senators Bill Hagerty and Andy Kim plan to introduce an amendment to defense legislation; Representatives Bill Huizenga and Sydney Kamlager-Dove back a related bipartisan House bill. Less than two weeks ago, Commerce Secretary Howard Lutnick signed an order compelling Anthropic to disable foreign access to its Fable 5 and Mythos 5 models on national security grounds.

Anthropic filed confidentially for an IPO on June 1, five days before the campaign ended, and sits at a roughly $965 billion valuation following its $65 billion round. The letter functions on several registers at once: a national-security disclosure, a regulatory request, and a pre-listing document that locates the company’s moat in Washington as much as in the model weights. The 2019 Huawei entity-listing template is visible in the policy response; what’s new is that the asset being walled off is inference traffic.

Sources