Brokers listing Alibaba as a single-stock CFD face a two-session gap and an open-ended headline overhang after Anthropic accused the Chinese group of the largest illicit campaign to extract its Claude models on record.
In a 10 June letter to the Senate Banking Committee, first reported by Bloomberg and since confirmed by CNBC and Reuters, Anthropic alleged that operators affiliated with Alibaba and its Qwen AI lab generated about 28.8 million exchanges with Claude through roughly 25,000 fraudulent accounts between 22 April and 5 June. It called this the largest known distillation attack on the company to date, following similar February claims against DeepSeek, Moonshot and MiniMax.
The move split by listing. US-listed BABA closed around 3% lower on Wednesday; the Hong Kong line (9988) then fell roughly 5% to about HK$94.55 on Thursday, a sixteen-month low. That overnight gap between New York close and Hong Kong open is the live exposure point for desks carrying client positions.
The allegation is unanswered. Alibaba has not responded, and Anthropic has not published the attribution method behind the account figure, leaving the record one-sided. Separately, Alibaba is suing the US Department of Defense over its 8 June “Chinese military company” listing, calling the designation baseless.
The catalyst lands on an already pressured name. Alibaba swung to an operating loss of RMB848 million in the March quarter, against a RMB28.5 billion profit a year earlier, with adjusted EBITA down 84% year on year on heavy AI, cloud and quick-commerce spending. The commercial read cuts both ways. Sustained volatility on a heavily traded single name lifts client flow and commission, but it raises gap and concentration risk and puts a reputational question against promoting a name under twin clouds. With no resolution in sight, this is a positioning call for product and risk teams, not a one-day story.
