Anthropic told the Senate Committee on Banking, Housing, and Urban Affairs that operators tied to Alibaba’s Qwen AI lab ran 28.8 million unauthorized exchanges with Claude through roughly 25,000 fraudulent accounts between April 22 and June 5, in what the company describes as the largest distillation campaign yet detected against its models. The June 10 letter, addressed to Sen. Tim Scott and Sen. Elizabeth Warren, was obtained by Bloomberg and CNBC on June 24, and it landed in Washington with the force of a policy artifact more than a corporate complaint.
The numbers themselves are the argument. By Anthropic’s accounting, this single alleged campaign exceeds the combined volume of three earlier ones it disclosed in February involving DeepSeek, MiniMax, and Moonshot AI, which together ran more than 16 million exchanges across roughly 24,000 accounts. One Chinese lab, in six weeks, now allegedly outpaces the entire prior cohort.
Markets read the subtext quickly. Alibaba’s Hong Kong shares fell as much as 4.9% on June 25 to a 16-month low, extending a year-to-date decline of 33%. Xiaomi and Baidu both dropped more than 3%, a sector-wide flinch that registered the letter less as a discrete accusation than as confirmation of a regulatory direction already in motion.
That direction isn’t subtle. The Pentagon added Alibaba to its Chinese military companies blacklist on June 8, two days before Anthropic’s letter; Alibaba is now suing the Defense Department to be removed. In April, White House Office of Science and Technology Policy Director Michael Kratsios issued a memorandum pledging to share intelligence with US labs about foreign distillation campaigns, an arrangement the current disclosure functionally activates in public. And the Trump administration has separately issued an export-control directive ordering Anthropic to suspend access to its Fable 5 and Mythos 5 models for any foreign national, citing “national security authorities.”
The commercial context sharpens the framing. Anthropic, valued at $965 billion after its $65 billion Series H, filed confidentially for an IPO this month. A letter detailing foreign exfiltration risk, delivered to the banking committee that oversees securities markets, is also a prospectus narrative: the moat is defensible, the threat is legible, the state is paying attention.
Distillation, as a technique, is what makes the numbers matter. Querying a frontier model at scale and using the outputs to train a competitor compresses years of capital expenditure into an API bill. It’s the reason the CHIPS-era vocabulary of export control has migrated, with surprising speed, from lithography tools to token streams. The 2019 Huawei entity-listing template, applied to a software substrate, is roughly the shape of what’s now being built.
What the letter reveals about elite psychology is the more durable signal. A year ago, US labs framed Chinese competitors as fast followers. The current frame is adversaries, and the venue is the Senate.
Sources
- https://www.bloomberg.com/news/articles/2026-06-24/anthropic-accuses-alibaba-of-illicitly-accessing-its-ai-models
- https://www.cnbc.com/2026/06/24/anthropic-alibaba-distillation-campaign.html
- https://thenextweb.com/news/anthropic-accuses-alibaba-distillation-claude-qwen
- https://www.bloomberg.com/news/articles/2026-06-25/alibaba-drops-after-anthropic-accuses-firm-of-accessing-ai-model
- https://www.bloomberg.com/news/articles/2026-06-26/what-is-ai-distillation-and-why-is-it-a-worry-for-the-industry