Former Google engineer Linwei Ding has been sentenced to 12 months and one day in federal prison after being convicted of stealing thousands of pages of confidential information related to the company’s artificial intelligence infrastructure.
A US federal judge in California handed Ding the sentence on September 1, 2026, along with two years of supervised release and financial penalties.
Ding, 38, was convicted in January on seven counts of trade-secret theft after prosecutors accused him of copying more than 2,000 pages of confidential Google material between May 2022 and April 2023. The documents covered hardware infrastructure and software platforms used to train and run large AI models.
Prosecutors said Ding uploaded the information to a personal Google Cloud account while secretly developing ties with technology companies in China. He also founded an AI and machine-learning company in China while still working at Google.
The case initially included charges of economic espionage. However, Judge Vince Chhabria vacated Ding’s economic espionage conviction in August after finding that prosecutors had not proved beyond a reasonable doubt that he knew his actions would benefit the Chinese government.
His seven trade-secret convictions remained intact.
How Ding obtained the information
According to the US Department of Justice, Ding copied confidential files covering Google’s AI data-centre infrastructure, including systems used for supercomputing and the training and serving of large AI models.
Prosecutors said he transferred the material to his personal cloud account and used some of the information while pursuing business opportunities in China.
The government also alleged that Ding sought participation in a Chinese government-backed talent programme and presented investors with plans to build an AI supercomputer based partly on technology copied from GoGoogl Google detected the activity in 2024, leading to Ding’s arrest and subsequent prosecution.
Judge calls theft a systematic effort
During sentencing, Judge Chhabria described Ding’s conduct as a deliberate and systematic effort to take Google’s intellectual property.
Ding’s defence had sought a shorter sentence, including home confinement, while pointing to his family circumstances and the damage the case had caused to his career.
The case highlights the growing legal and security risks surrounding AI technology as companies invest heavily in the infrastructure and systems needed to develop increasingly capable models.
For Google and other AI companies, the theft of proprietary infrastructure can expose not only commercial information but also technical knowledge that could help competitors reproduce expensive AI systems.
Read also: Paycode: The fintech taking digital payments offline