US data brokers sell RL environments to Chinese AI labs, report says
Original title: US data brokers are selling half a billion dollars in RL training environments to Chinese AI labs
ChinaTalk names at least nine US firms and about $500 million in sales; export controls largely miss these environments.
A ChinaTalk investigation says US data firms including Mercor, Surge AI, AfterQuery, and Turing sell reinforcement-learning environments and other training data to Chinese AI labs. The report says labs buy products previously sold to Anthropic or OpenAI and distill outputs from models such as GPT and Claude. Coding and biology environments raise dual-use concerns, and the authors argue the Bureau of Industry and Security should apply export controls. They also say Chinese labs still trail US peers by months.
Unconfirmed: Sales relationships and the half-billion-dollar figure come from ChinaTalk and were not independently verified here.
- At least nine US data firms were identified as selling to Chinese customers.
- ChinaTalk could not independently verify reports of about $100 million in AfterQuery ARR.
- ByteDance procurement offices in Singapore and San Jose are cited.
- Indian subcontractors sometimes posted client names despite code-name rules.
- Sensitive biology, chemistry, and cybersecurity environments are increasingly built in-house.
From the original article
2 min read at the-decoder.com
US data brokers like Mercor, Surge AI, and AfterQuery are selling roughly half a billion dollars' worth of AI training data to Chinese labs, according to a ChinaTalk investigation. That includes the same RL environments previously built for OpenAI or Anthropic. At least nine firms are involved, and export controls barely apply.
By Maximilian Schreiner · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.