Former Google engineer Linwei Ding convicted of stealing AI secrets, found guilty of economic espionage for China
A federal jury in San Francisco convicted former Google engineer Linwei Ding on Jan. 29, 2026, of stealing confidential AI technology, finding him guilty of economic espionage and theft of trade secrets tied to China-linked entities.
Key takeaways
- Defendant: Linwei (Leon) Ding, 38, former Google software engineer.
- Verdict: Guilty on all 14 counts — seven counts of theft of trade secrets and seven counts of economic espionage.
- Scope of theft: Prosecutors say he copied more than 1,000 unique files, including technical details about Google’s AI hardware and software.
- First-of-its-kind: This marks the Justice Department’s first conviction in an AI-related economic espionage case, underscoring national security concerns.
Main coverage
The case and verdict
Federal prosecutors told an 11-day trial in the U.S. District Court in San Francisco that Linwei Ding began copying sensitive Google documents in May 2022 and moved them to personal cloud accounts while evading security monitoring. The jury returned a guilty verdict on January 29, 2026, convicting Ding on all 14 counts — seven counts of theft of trade secrets and seven counts of economic espionage.
Why it matters: Prosecutors characterized the stolen material as central to the performance and efficiency of Google’s AI infrastructure, and the verdict is being framed as a protection of U.S. innovation and national security.
Evidence presented
Prosecutors introduced court filings and trial exhibits showing more than 1,000 unique files — roughly 14,000 pages with 105 core documents — containing technical details about Google’s Tensor Processing Units (TPUs), GPUs and SmartNIC network interface cards. They argued these files were critical to optimizing AI training and deployment and that Ding uploaded many files to personal Google Drive accounts before copying them to his personal computer shortly before resigning on Dec. 26, 2023.
A news account summarized the files and evidence in detail: coverage at The Register.
Role at Google
Ding was hired by Google in May 2019 as a software engineer focused on improving GPU efficiency for machine learning workloads at Google and Google Cloud data centers. His position gave him access to detailed hardware and software designs used to boost compute speed and lower AI training costs — access prosecutors said made the copied material especially valuable.
Alleged ties to China-linked companies
Federal prosecutors alleged Ding secretly worked with entities tied to the People’s Republic of China, including serving as chief technology officer and listing himself as CEO of Shanghai Zhisuan Technology Co. Ltd. by May 30, 2023. He reportedly met investors for a firm called Rongshu and applied to a Chinese government-sponsored talent plan. Prosecutors argue the intent was to use Google’s technology to help build AI supercomputers and custom chips for these firms.
At a technology conference, prosecutors say Ding presented himself as CEO of Zhisuan and claimed the company could replicate aspects of Google’s systems — evidence used to show he sought to benefit PRC-linked firms and build relationships in China’s tech sector.
Detection, investigation and arrest
Google’s security team flagged unusual uploads to a secondary personal Drive account while Ding was traveling in China from Oct. 29, 2022, to March 25, 2023. Surveillance footage and badge data introduced at trial showed a colleague scanning Ding’s badge to create the appearance he was in the United States while abroad. Prosecutors say Ding signed a false affidavit on Dec. 8, 2023, denying retention of Google data as he booked a one-way flight to Beijing.
After a Google employee recognized Ding and the company details he shared at a conference, Google suspended his accounts, locked his laptop and seized devices on Jan. 4, 2024. The FBI executed a search on Jan. 6, 2024. An indictment was filed March 5, 2024, and a superseding indictment followed in February 2025.
Legal arguments and courtroom debate
Defense arguments maintained Ding did not hand material to the Chinese government, had no intent to benefit China as a nation, and that Google had not adequately protected the documents. They emphasized there was no proof the files were ever used or sold to a foreign government. After the verdict, defense attorney Grant Fondo said the team was disappointed and would review options.
Judge Vince Chhabria denied a motion to dismiss during the trial and at one hearing questioned whether actions by an individual engineer constituted economic espionage that benefitted “China as a country” versus private firms — a point the defense raised repeatedly.
Sentencing and next steps
Under federal law, each theft count carries up to 10 years and each economic espionage count up to 15 years. Sentencing could expose Ding to decades behind bars and possible multimillion-dollar fines, though the actual term will depend on the judge’s decisions after a presentence report and hearing. Judge Chhabria released Ding pending sentencing and set a status conference for Feb. 3, 2026.
Reactions from government and Google
U.S. Attorney Craig H. Missakian described the verdict as protecting U.S. innovation and national security. FBI officials framed the decision as a warning against threats to American AI technology. Google Vice President Lee-Anne Mulholland said the company was grateful to the jury and reiterated that stealing trade secrets has serious consequences; Google declined further comment on investigation specifics.
Implications for Los Angeles, California
Economic and industry impact: Los Angeles houses major tech teams, startups and universities that build AI tools and rely on cloud computing. The verdict may prompt tighter internal controls, increased investor caution and higher security costs, particularly for smaller startups.
Workforce and talent flows: Southern California employers that recruit internationally may add deeper background checks and stricter remote access rules, potentially slowing hiring and shifting collaboration practices with overseas partners.
Higher education and research ties: Universities may review data-sharing agreements, tighten export-control training and impose clearer rules for visiting scholars working on sensitive AI, semiconductor and robotics research.
Local politics and daily life: Policymakers may pursue public-private partnerships for cybersecurity and training. For most residents, daily life won’t change, but consumers and companies may see more public statements about data controls and compliance.
