SAN FRANCISCO, September 2 — How does a software engineer walk out of one of the world’s most closely watched technology companies with hundreds of confidential files? The case of Linwei Ding offers a rare glimpse into the mechanics of alleged trade-secret theft at the heart of the global artificial intelligence race.
Ding, a former Google engineer, was sentenced to just under a year in prison on Tuesday after a jury convicted him of stealing proprietary information tied to the company’s AI operations. US District Judge Vince Chhabria handed down the sentence in San Francisco and did not mince words about the nature of the crime. Chhabria called Ding’s conduct a “systematic, brazen effort to steal Google’s property.” The judge added: “He knew very well what he was doing.
He hurt people along the way.” The sentencing followed a January jury verdict that found Ding, then 38, guilty on seven counts involving theft of proprietary information connected to Google’s AI operations. The prosecution’s case centred on how Ding allegedly removed confidential files from Google’s systems over an extended period.
At trial, prosecutors said he transferred more than 500 confidential files from Google’s network to his personal Google Cloud account across roughly a year. The science behind what Ding allegedly took is what makes this case significant. The files contained trade secrets concerning Google’s hardware infrastructure and software used in supercomputing data centres for training large AI models — the backbone of modern artificial intelligence development.
These are not trivial specifications; they represent years of engineering investment in the specialised systems that allow tech giants to build increasingly powerful machine-learning systems. Prosecutors also presented evidence about Ding’s connections with technology companies in China that were working on AI-related projects.
They said he copied and took confidential information from Google with the intention of helping Chinese interests. Several months after Ding started moving files, his activities drew the attention of investigators. Intriguingly, the case took a notable turn in August.
Chhabria overturned Ding’s conviction on economic espionage charges, ruling that the evidence was insufficient for a reasonable jury to reach a guilty verdict on that specific charge. The trade-secret theft convictions stood, but the more serious national-security allegation — that he intended to benefit China through the stolen information — did not survive judicial review.
What researchers found in cases like this often points to a broader pattern. The US Department of Justice has pursued multiple prosecutions involving Chinese economic espionage and trade-secret theft in recent years, though courts have not always affirmed the government’s most aggressive theories. Ding’s case illustrates the tension between protecting American technological advantage and proving criminal intent beyond reasonable doubt.
What to watch next: whether Ding serves his full sentence, whether Google pursues parallel civil litigation, and how this verdict shapes corporate security practices in an era when AI talent moves fluidly across borders. The mechanics of how companies safeguard their most valuable technical secrets — and how prosecutors build cases when those safeguards fail — remain live questions as competition between American and Chinese technology sectors intensifies.
