Cohere CEO Challenged US Claims on Chinese AI Development
Aidan Gomez argued that Chinese AI models demonstrate innovation that outperforms original Western technology.
Updated on Sept. 19, 2026 in Artificial Intelligence
Live Poll
Should nations prioritize developing sovereign AI systems that operate independently of foreign technology stacks?
Cohere CEO Aidan Gomez has disputed United States intelligence claims that Chinese firms are building AI models through industrial-scale distillation campaigns. He contended that these models are innovative enough to outperform their American counterparts on key benchmarks.
Why it matters
The dispute centers on whether Chinese technological advancements are based on illicitly copied software or genuine innovation. Gomez argues that if these models were mere distilled copies, they would not be able to consistently outperform the original technology.
Chinese GLM models currently process roughly 10 trillion tokens daily using domestically produced chips. This performance occurs despite U.S. export controls and claims of distillation from powerful Western models.
The players
Aidan Gomez
He is the CEO of Cohere, a company that provides enterprise-focused large language models.
China Ministry of Foreign Affairs
This is the executive department of the Chinese government responsible for foreign relations.
NSA
The National Security Agency is a United States intelligence agency responsible for global monitoring and data collection.
CISA
The Cybersecurity and Infrastructure Security Agency is a federal agency tasked with protecting U.S. critical infrastructure.
FBI
The Federal Bureau of Investigation is the domestic intelligence and security service of the United States.
The details
Distillation involves training a smaller, cheaper AI model using the outputs of a more sophisticated, powerful version. U.S. agencies, including the NSA, CISA, and FBI, have alleged that Chinese firms used this technique to illicitly advance their capabilities.
Timeline
In June 2026, 28.8 million exchanges occurred through 25,000 fraudulent Claude accounts.
On September 8, 2026, the NSA, CISA, and FBI issued a joint advisory naming six Chinese firms.
On September 9, 2026, China's Ministry of Foreign Affairs denied accusations of intellectual property theft.
On September 10, 2026, Aidan Gomez discussed the claims on CNBC.
The Tech Race
The emergence of high-performance domestic chip usage in China marks a critical shift in the global AI hardware arms race. This trend suggests that international sanctions may be accelerating domestic self-reliance rather than halting progress.
The dispute could lead to stricter scrutiny of AI model usage and potential new authentication requirements for cloud-based tools. Users may experience more robust identity verification processes when accessing powerful generative AI platforms.
The takeaway
The disagreement highlights the difficulty in distinguishing between stolen intellectual property and independent technical achievement in rapidly evolving markets. Ongoing intelligence monitoring is likely to remain a central tension in global technology development.
Further reading
For more on global AI development, visit the Artificial Intelligence section.
Source note: This article includes information reported by Crypto Briefing.
Live Poll
Should nations prioritize developing sovereign AI systems that operate independently of foreign technology stacks?
