China Steals $600 Billion a Year in U.S. IP, Eroding Key Industries
Updated
Updated · Fox News · May 16
China Steals $600 Billion a Year in U.S. IP, Eroding Key Industries
1 articles · Updated · Fox News · May 16
$600 billion in annual losses from Chinese espionage are hitting U.S. companies and weakening national security, with theft often going undetected or unreported, according to former CIA officer Tom Lyons.
Lyons said Beijing’s state-directed system targets entire industries—not just individual firms—and cited the Linwei Ding case, in which a Google engineer stole AI chip designs later used to market a China-based venture.
Over 25 years, the report says, systematic theft has eroded or wiped out U.S. positions in steel, telecommunications, solar and semiconductors, while U.S. court judgments remain largely unenforceable against China-based actors.
The piece urges a tougher U.S. response ahead of the U.S.-China summit, including subsidized litigation, harsher espionage penalties, intelligence support for businesses and designation of Chinese-linked groups as state-sponsored economic espionage organizations.
As America toughens its laws, can it truly deter a state-sponsored system built to acquire its technology?
Is America's fight against espionage inadvertently sending top scientific talent directly to its biggest competitor?
2026 U.S.-China AI Espionage Crisis: The Scale, Methods, and Policy Response to Industrial AI Model Theft
Overview
In April 2026, the United States government issued the NSTM-4 memorandum, formally accusing China of running coordinated, industrial-scale campaigns to steal American artificial intelligence technology. The memorandum highlighted a sophisticated and pervasive threat to U.S. innovation and economic security. According to Michael Kratsios, foreign entities based mainly in China are systematically distilling U.S. frontier AI models. These Chinese operations use advanced methods like distillation attacks, leveraging tens of thousands of proxy accounts and jailbreaking techniques to evade detection and extract proprietary information from American AI models.