By: web admin Posted on

San Francisco: A ex-Google software engineer has been found guilty of stealing thousands of pages of confidential artificial intelligence (AI) data and attempting to use them to build China’s next-generation AI infrastructure.

Linwei Ding, also known as Leon Ding (38), was convicted on 14 federal counts — seven for economic espionage and seven for theft of trade secrets — following an 11-day trial before U.S. District Judge Vince Chhabria.

The case, described by U.S. authorities as one of the most serious technology theft prosecutions involving AI, centers on Ding’s alleged plan to copy Google’s core AI systems to fuel Chinese startups and state-backed research.

Stole 2,000+ Pages of AI Blueprints

Prosecutors told the court that between May 2022 and April 2023, Ding secretly downloaded and uploaded more than 2,000 pages of sensitive AI documents from Google’s internal systems to his personal cloud account.

The stolen files revealed the hardware and software backbone of Google’s AI supercomputers — the same systems used to train large language models and run global data centers.

The data included:

  • Google’s custom Tensor Processing Units (TPUs)

  • Advanced GPU cluster architectures

  • AI workload coordination software

  • Proprietary SmartNIC networking technology

  • Systems that link thousands of chips into one AI supercomputer

China Link and Startup Plan

While still working at Google, Ding secretly affiliated himself with two China-based technology firms. By mid-2022, he was negotiating to become CTO of a Chinese AI company. By early 2023, he had begun setting up his own AI startup in China, where he served as CEO.

In investor pitch decks, Ding allegedly claimed he could replicate Google’s AI supercomputing systems to create a competitive Chinese alternative.

In December 2023, less than two weeks before resigning from Google, Ding transferred the stolen files onto his personal computer.

Ties to Chinese State Programs

Court evidence showed Ding referenced China’s national AI development strategy in business presentations. He also applied for a Shanghai government-backed “talent plan”, stating his goal was to help China achieve global-level computing power.

Prosecutors said Ding aimed to assist two Chinese government-controlled entities by:

  • Developing an AI supercomputer

  • Conducting custom machine-learning chip R&D

U.S. Officials: National Security at Risk

U.S. Attorney Craig H. Missakian said the verdict sends a strong warning:

“Silicon Valley leads the world in AI innovation. The theft of this technology puts our economic security and national security at risk. This will not go unpunished.”

FBI San Francisco chief Sanjay Virmani added:

“Stealing advanced AI technology for the benefit of the People’s Republic of China threatens America’s technological edge. This case shows that we will protect U.S. innovation.”

What’s Next

Ding was first indicted in March 2024, with expanded charges added in February 2025. He is scheduled to appear in court again on February 3, 2026.

He now faces:

  • Up to 10 years per trade secret theft count

  • Up to 15 years per economic espionage count

The final sentence will be determined under U.S. federal guidelines.

The case highlights growing global tensions over AI dominance, national security, and technological espionage — with courts now treating data theft as a threat on par with traditional state secrets.

Leave a Reply

Your email address will not be published. Required fields are marked *