Rene Haas, CEO of Arm Holdings
In an interview with Reuters on Tuesday, Haas expressed that it would be a significant challenge to halt the export of CPUs to China. He said restricting these CPUs is particularly challenging because they are widely deployed across a range of applications. Their broad adoption and the difficulty of identifying whether they are being used for AI workloads make targeted controls hard to enforce.
According to Haas, the difficulty of setting specific performance thresholds and memory bandwidth limits, unlike Nvidia's
"They would have to limit everything," Haas said, implying that while the U.S. could try, it would be a tougher area to regulate than AI chips.
CPUs Emerge As AI's Next Battleground
The Arm CEO's comments come as the U.S. recently closed a loophole that allowed AI chip shipments from Nvidia and AMD to Chinese firms overseas. This move came despite ongoing U.S. measures to limit China's access to semiconductors crucial for AI advancements. The U.S. now plans to impose license requirements for advanced chips on entities headquartered in China, regardless of their location.
Although Nvidia's GPUs have been the primary beneficiaries of the AI boom, demand for CPUs has surged in recent months as the industry shifts toward AI inference. CEO Jensen Huang already sees its Vera CPUs sitting on a $200 billion untapped market, and considers China a 'very important' destination for its CPU business despite ongoing U.S.-China technology restrictions.
Coming back to Arm, its AGI CPU is gaining momentum in the AI industry. On Tuesday at the Taipei Computex conference, Hass said Oracle Corp.
