Last Updated:
Former Google engineer Linwei Ding has been sentenced to nearly a year in prison for stealing confidential AI chip technology and passing it to Chinese firms.
Linwei Ding was sentenced to 12 months minus one day for stealing Google’s confidential AI chip technology.
A US court has sentenced former Google engineer Linwei Ding to 12 months minus one day in prison. He was found guilty of stealing confidential details about Google’s AI chip technology and passing them along to Chinese companies.
US District Judge Vince Chhabria handed down the sentence on September 1 in San Francisco. Ding will also serve two years of supervised release, pay over $189,000 in restitution, and a $25,000 fine.
What He Stole
The files Ding transferred covered Google’s hardware and software for training large AI models. That includes details on Google’s Tensor Processing Units, or TPUs, its custom AI chips, along with GPU systems, and the software that lets thousands of chips work together as one supercomputer.
These aren’t small details. Google’s TPUs let the company train massive AI models without depending on outside chipmakers like Nvidia. Experts say documents like these could let someone skip years of design work Google spent millions of dollars and a decade building. Judge Chhabria didn’t hold back describing the crime, calling it “a systematic, brazen effort to steal Google’s property.”
Who Is Linwei Ding?
Ding, also known as Leon Ding, is a Chinese national. He came to the US in 2010 for a master’s degree, worked at a few Silicon Valley companies, then joined Google as a software engineer in 2019. He became a lawful permanent resident two years later, in 2021.
His job gave him access to the confidential systems inside Google’s AI supercomputing data centres. A San Francisco jury convicted him in January after an 11-day trial, on seven counts of trade secret theft and seven counts of economic espionage. The judge later set aside the espionage counts.
According to prosecutors, Ding started uploading confidential files from Google’s internal network to his personal Google Cloud account on May 21, 2022. That continued until May 2023.
At trial, prosecutors said he uploaded more than 500 confidential files over roughly a year. A later, expanded indictment put the number even higher, over 1,000 unique files, spanning more than 2,000 pages of confidential material. Court records also show that in December 2023, just under two weeks before leaving Google, Ding downloaded material onto his own personal computer.
Beyond TPUs and GPUs, the files also covered software that helps chips communicate with each other, and details of Google’s custom SmartNIC, a networking part used in its AI supercomputers and cloud systems. Prosecutors said some of this information was meant to give Google an edge over rival cloud companies, and reduce how much it depends on Nvidia chips.
Months after Ding started uploading files, an early-stage Chinese company called Rongshu offered him a chief technology officer role. He visited the company in late 2022 and joined investor meetings from around December that year.
In 2023, Ding started his own company, Zhisuan, based in China, aimed at training large AI models. One internal document he shared read: “We have experience with Google’s ten-thousand-card computational power platform; we just need to replicate and upgrade it and then further develop a computational power platform suited to China’s national conditions.”
Ding also applied to a Chinese government programme designed to bring tech talent to China. Prosecutors say he never told Google about either company, and while he was in China pitching Zhisuan, he had an intern swipe his Google ID badge multiple times, to make it look like he was still working from his desk. The judge later found the evidence wasn’t strong enough to prove Ding knew or intended, during that specific May 2022 to April 2023 window, that his actions would benefit the Chinese government.
September 03, 2026, 13:29 IST
Read More
