U.S. convicts ex-Google engineer for sending AI tech data to China

A U.S. federal jury has convicted Linwei Ding, a former software engineer at Google, for stealing highly sensitive artificial intelligence (AI) supercomputer data and secretly sharing it with Chinese technology firms.

The verdict follows an 11-day trial in San Francisco and marks one of the most serious AI-related trade secret cases in recent years.

How the Case Started

Ding joined Google in 2019 as a software engineer working on advanced AI infrastructure. According to U.S. prosecutors, between May 2022 and April 2023, he stole more than 2,000 pages of confidential internal documents from Google.

He uploaded these materials to his personal Google Cloud account without authorization. The files contained critical information about how Google designs and runs large-scale AI systems.

Ding was first indicted in March 2024 after investigators found that he had lied and failed to cooperate honestly during Google’s internal investigation. He was later arrested in California.

What Information Was Stolen?

The stolen documents revealed some of Google’s most valuable AI technologies, including:

• AI supercomputing infrastructure design
• Proprietary TPU and GPU system architecture
• Software used to manage large AI workloads
• SmartNIC networking technology for high-performance computing

These systems are essential for training and running powerful AI models used in search, cloud services, and generative AI tools.

Secret Ties to Chinese Tech Firms

Prosecutors told the jury that Ding was secretly affiliated with two China-based technology companies while still employed at Google. He even negotiated a Chief Technology Officer (CTO) role at one of them.

Later, Ding founded his own AI startup in China, Shanghai Zhisuan Technology Co., where he served as CEO. He reportedly told potential investors that he could build AI supercomputing systems similar to Google’s.

Evidence also showed that Ding applied to a Shanghai government-backed talent program. In his application, he stated that his goal was to help China develop computing infrastructure “on par with the international level.”

Attempts to Hide His Actions

Ding never disclosed his foreign affiliations or frequent travel to China to Google. In one striking detail revealed during the trial, he asked a colleague to scan his employee badge at Google offices to make it appear that he was physically present in the U.S. while he was abroad.

According to the U.S. Department of Justice, Ding intended to benefit entities linked to the People’s Republic of China by helping develop AI supercomputers and custom machine-learning chips.

Verdict and Possible Sentence

The jury convicted Ding on:

• Seven counts of economic espionage
• Seven counts of trade secret theft

Each charge carries a potential prison sentence of 10 to 15 years. Sentencing has not yet been announced.

Why This Case Matters

This case highlights growing global tensions around AI, data security, and advanced computing technologies. As AI becomes central to economic growth and national security, companies and governments are paying closer attention to how sensitive technologies are protected.

For countries like Rwanda and others investing in digital transformation, the case is also a reminder of the importance of strong data governance, ethical innovation, and trust in the global tech ecosystem.

Read more: U.S. Department of Justice Official Press Release