A former Google software engineer has been found guilty in a landmark US case that underscores how artificial intelligence has become a frontline issue in economic security, espionage and geopolitical competition.
A federal jury in San Francisco has convicted Linwei Ding, also known as Leon Ding, on seven counts of economic espionage and seven counts of theft of trade secrets after an 11-day trial in the Northern District of California. Prosecutors said Ding stole thousands of pages of confidential Google documents containing advanced AI trade secrets for the benefit of entities aligned with the People’s Republic of China.
US authorities described the case as the first conviction involving AI-related economic espionage, reflecting the rising strategic value of AI infrastructure, chips and large-scale computing platforms. The stolen material related to the hardware and software systems that underpin Google’s AI supercomputing environment, including custom Tensor Processing Unit architectures, GPU systems, orchestration software and high-speed networking components.
According to evidence presented at trial, Ding accessed and exfiltrated more than 2,000 pages of confidential information between May 2022 and April 2023 while employed at Google. Prosecutors said the data was uploaded to Ding’s personal cloud account and later downloaded to his personal computer shortly before he resigned from the company.
At the same time, Ding was allegedly pursuing business opportunities in China without disclosing them to his employer. The court heard that he entered discussions to become chief technology officer of a China-based start-up and later founded his own AI-focused company in the PRC, acting as its chief executive. In presentations to investors, Ding claimed he could build an AI supercomputer by copying and modifying Google’s technology.
The jury also heard evidence that Ding applied for a Chinese government-sponsored talent program in Shanghai, stating his intention to help China develop computing infrastructure on par with international standards. Prosecutors argued that his actions were aligned with PRC national policies designed to accelerate domestic AI capability and reduce reliance on foreign technology.
US national security officials said the conviction sends a strong signal to the technology sector. Assistant Attorney General for National Security John Eisenberg described the case as a calculated breach of trust involving some of the world’s most advanced AI technology at a critical moment in global AI development. The FBI characterised the verdict as a clear warning that theft of AI intellectual property will be treated as both an economic and national security threat.
For technology leaders, the case highlights how AI infrastructure has moved beyond software models into tightly integrated stacks of custom silicon, networking and orchestration platforms. These systems represent enormous investment and competitive advantage, making them high-value targets for espionage.
The conviction also reinforces the growing risk associated with insider threats in advanced technology organisations, particularly in sectors where AI, semiconductors and high-performance computing intersect with national strategy. As companies race to scale AI capability, the need for stronger internal controls, monitoring and governance around sensitive research and infrastructure is becoming increasingly urgent.
Ding faces a potential sentence of up to 10 years’ imprisonment on each count of trade secret theft and up to 15 years on each count of economic espionage, with sentencing to be determined by the court under federal guidelines. He is scheduled to appear at a status conference on 3 February 2026.
The case reflects a broader reality confronting the AI sector: innovation, talent mobility and global collaboration now coexist with intensified scrutiny, enforcement and strategic competition. For organisations operating at the cutting edge of AI, the boundary between commercial success and national security has never been thinner.
