- Nvidia CEO Jensen Huang urges caution about China’s computing capacity
- China has “enormous” amounts of computers, including lots that sit unused
- Huang says China practicing a Mythos-like model could cause problems
Nvidia CEO Jensen Huang has warned of the potential threat China poses in harnessing vast amounts of computing power to train its next generation of AI models.
Speaking on the Dwarkesh Patel podcast, Huang said China could soon be able to train an AI model similar to Anthropic’s recently announced Claude Mythos — which could have troubling effects on global cyber security.
Huang also raised concerns about the “enormous” amount of computers China currently has that are not being used, but said both his company and the US in general should still be in a strong position.
The article continues below
Chips and more chips
“The amount of capacity and the type of computer (Mythos) was trained on is abundantly available in China, so you just have to first realize that there are chips in China,” Huang said, noting how Mythos was trained on a “pretty banal capacity”.
“They (China) have data centers that are completely empty, fully powered,” Huang added. “You know, they have ghost towns, they also have ghost data centers. They have so much infrastructure capacity. If they wanted to, [could] just collect more chips…Their capacity to build chips is one of the greatest in the world.”
China already manufactures a huge amount of the chip hardware used around the world and boasts many leading universities and AI researchers — offering enormous potential for cooperation rather than rivalry, Huang noted.
“This is an area that is sorely missed because of our current stance on China as an adversary,” he said. “It’s important that our AI researchers and their AI researchers actually talk.”
“Sacrificing them, making them an enemy, is probably not the best answer,” he added. “They are an adversary.”
“We want the United States to win. But I think the safest thing is to have a dialogue and to have a research dialogue.”
“It is important that we both try to agree on what we should not use AI for.”
Mythos is a central part of Project Glasswing, a new cybersecurity initiative that Anthropic is leading with technology leaders to identify and fix vulnerabilities in critical software.
It quickly gained widespread attention, and OpenAI also recently unveiled GPT-5.4-Cyber, its Mythos rival, designed for cybersecurity professionals to spot the next level of attacks.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews and opinions in your feeds.
