Former Google Engineer Convicted for Spying for Chinese Companies, Stealing Secret AI Data

A U.S. court has convicted Linwei Ding, a former Google engineer, for stealing trade secrets and conducting economic espionage for Chinese companies. Ding was found guilty of seven counts of economic espionage and seven counts of theft of trade secrets, involving thousands of pages of confidential data from Google’s AI business.

Espionage for Chinese Firms
According to prosecutors, Ding secretly worked for two Chinese companies while employed at Google. He allegedly stole sensitive information related to Google’s AI operations and shared it with these firms. The 38-year-old Chinese national was convicted by a federal jury in San Francisco after an 11-day trial.

The U.S. Department of Justice reported that Ding’s theft involved critical hardware infrastructure and software platforms used by Google’s supercomputing data centers to train large AI models. Some of the stolen chip blueprints were reportedly intended to help the Chinese firms gain a competitive edge against cloud rivals like Amazon and Microsoft and reduce Google’s dependence on Nvidia chips.

Potential Penalties
Economic espionage charges carry a maximum penalty of 15 years in prison and a fine of $5 million per count. Theft of trade secrets can result in up to 10 years in prison and fines of $250,000 per count. Ding is scheduled to appear for a status conference on February 3. His lawyer has not commented on the conviction.

Background
Ding joined Google in May 2019 and began stealing confidential information three years later, reportedly after being recruited by an early-stage Chinese technology company. The case was coordinated by the Disruptive Technology Strike Force, an inter-agency unit established under the Biden administration in 2023. Google has not been accused of wrongdoing and has confirmed its cooperation with law enforcement.

Supercomputers and AI Training
Supercomputing data centers are powerful machines used to train AI models that can think and learn like humans. The stolen chip blueprints detailed designs for specialized processors that enhance the speed and performance of these systems.