A former Google engineer was convicted of stealing the company’s artificial intelligence (AI) trade secrets by a federal court in the U.S. to benefit Chinese firms.
On Thursday (Jan. 29), a jury found Linwei Ding, also known as Leon Ding, guilty on 14 federal charges, including economic espionage and theft of trade secrets.
The conviction marks the first U.S. prosecution on AI-related economic espionage charges and signals sustained government enforcement of AI intellectual property (IP) protection.
It came barely eight weeks after Ding was indicted.
“The speed alone signals something: this wasn’t exploratory prosecution,” wrote The Meridiem.
“This was the government confirming it had the evidence, the legal framework, and the stomach for sustained enforcement against AI-related IP theft.”
Ding, a Chinese national, joined Google in 2019.
He stole over 2,000 pages of Google’s most sensitive AI architecture between 2022 and 2023.
They included schematics for custom Tensor Processing Unit chips, the specialized graphics processing hardware that powers Google’s AI training infrastructure, and detailed specifications for SmartNIC, Google’s proprietary network interface card that orchestrates high-speed communication across AI supercomputers.
He uploaded the documents to his personal Google Cloud account.
Ding was secretly working with two Chinese tech firms around that time and made plans to launch his own AI startup in China.
He claimed he could build a supercomputer by rebranding Google’s technology as his own.
The FBI arrested him in 2024.
Ding now faces decades in prison on those charges.
