Anthropic Accuses Alibaba of AI Theft as Google Talent Defects to Rival

US artificial‑intelligence firm Anthropic sent a letter to the Senate Banking Committee alleging that operators linked to Alibaba’s Qwen lab illicitly extracted capabilities from its Claude model. Anthropic says the campaign involved about 25,000 fraudulent accounts that generated roughly 28.8 million interactions with Claude between April 22 and June 5, 2026 – the largest known “distillation” attack on its technology. The company warned that the theft could accelerate Chinese AI development and urged the US government to take coordinated action, supporting export controls and industry threat‑intelligence sharing.

In response, Alibaba filed a lawsuit in the U.S. District Court for the Northern District of California seeking removal from the Pentagon’s 1260H list, which designates it a “Chinese military company.” The Chinese e‑commerce giant argues the designation, announced on June 8, lacks any factual or legal basis and harms its ability to do business with U.S. partners. The Pentagon maintains that Alibaba’s ties to China’s State Administration of Supervision and the Ministry of Industry and Information Technology constitute indirect support for the People’s Liberation Army’s defense industrial base.

At the same time, Google’s Alphabet unit is losing senior AI researchers to Anthropic. Jonas Adler and Alexander Pritzel, key contributors to Google’s Gemini model, have announced moves to Anthropic, joining earlier defections of Nobel laureate John Jumper (to Anthropic) and Noam Shazeer (to OpenAI). The talent outflow, occurring as Anthropic prepares for an IPO, has raised investor concerns about Google’s ability to compete in the fast‑moving AI field.