Chinese AI Stocks Slide After Report of Probe into DeepSeek, Moonshot AI

September 23, 2026 at 10:00 AM2 min

This digest was compiled by AI from multiple sources — links to the originals are below.

Shares of Chinese AI developers fell in Hong Kong on September 23, 2026, after The Information reported that China's cyberspace regulator opened a probe into DeepSeek and Moonshot AI over data security concerns. Z. AI Co. dropped as much as 12% and MiniMax Group Inc. 6.8%, while Alibaba, Xiaomi and Tencent also declined. The probe follows Anthropic's accusations that the startups routed user queries through its Claude models.

Key Facts

Z. AI Co. shares fell as much as 12% in Hong Kong trading on September 23, 2026, after the report of the probe.
MiniMax Group Inc. declined 6.8%, Alibaba Group Holding Ltd. about 5%, Xiaomi Corp. 3.8%, and Tencent Holdings Ltd. about 3%.
The probe was opened by the Cyberspace Administration of China, the country's main internet regulator.
Anthropic's 154-page report accused Moonshot AI of sending over 23 million user queries to Claude between May and July, and DeepSeek of sending 12.1 million queries over 14 days in July.
The Information reported that the probe began after Anthropic's accusations that DeepSeek and Moonshot AI routed sensitive data through Claude models.

Market Reaction

Shares of Chinese AI model developers fell in Hong Kong on September 23, 2026, after The Information reported the regulator's probe into DeepSeek and Moonshot AI. Z. AI Co. dropped as much as 12% and MiniMax Group Inc. 6.8% in Hong Kong trading, according to Bloomberg. Alibaba Group Holding Ltd., which develops the Qwen model, slid about 5%, while Xiaomi Corp. fell 3.8% and Tencent Holdings Ltd. about 3%. The declines add to investor concerns that Chinese AI firms may face intensified scrutiny amid the growing debate over AI safety.

Regulator Probe

The Cyberspace Administration of China, the country's main internet regulator, is conducting the probe into DeepSeek and Moonshot AI. The regulator summoned executives from leading Chinese AI labs for questioning, but focused its attention on DeepSeek and Moonshot AI, the creator of the open-source model Kimi. The probe is checking whether Moonshot and DeepSeek could have leaked confidential information to a third party. The Information reported that the probe began after Anthropic's accusations that the startups had been routing sensitive data through its Claude models.

Anthropic Accusations

Anthropic's 154-page report, titled 'Identifying and Preventing AI Misuse', accused the Chinese companies of transmitting dialogues between their models and users to Claude. Anthropic stated that Moonshot sent over 23 million user queries to Claude between May and July, and DeepSeek sent 12.1 million queries over 14 days in July. The Chinese authorities are concerned that this could have led to the transfer of 'sensitive information' to the United States. The potential punishment for DeepSeek and Moonshot AI remains unknown.