China ‘Will Win’ the AI Race Against the US, Nvidia CEO Says

Vinod Dsouza
China USA Chip AI
Source: Getty Images

Nvidia CEO Jensen Huang said on Wednesday that China will win the AI race against the US. He pointed to lower energy costs in China and looser regulations as leverage, while it’s the opposite in the US. “China is going to win the AI race,” said Huang to the Financial Times.

He also stressed that the US, UK, and the West in general are being held back by “cynicism” in the AI sector, which China is not. “We need more optimism,” Huang said at the Financial Times’ Future of AI Summit. His stark comments come after Trump banned Nvidia from selling its most advanced Blackwell chip to China.

Also Read: China Wants $4 Billion US Dollar Bond, Order Sheet Tops $65 Billion

“We cannot sell them to China. We cannot sell them to other people,” said Trump, citing that the most advanced and powerful chips must not be sold to Beijing. Trump is tightening the grip on China, making it come second to the US in the AI sector. The President also cited military and surveillance operations among the other reasons for the ban.

“These are super powerful chips. We’re not going to let them fall into the wrong hands,” said Trump. However, the White House allowed Nvidia to sell the same powerful chip to South Korea for commercial and research purposes only. Nvidia is yet to disclose the licensing details in the agreement to South Korea. Both the US and China have made AI tech a national priority, aiming to become the next global hub.

Also Read: Everyone’s Buying Tech Stocks, Warren Buffett Is Doing the Opposite

AI Race Swings Towards China, US and South Korea Still Competing Hard

Tesla AI5 Chips
Source: TechBuzz.ai

Even South Korea is in the race but faces lighter restrictions than China. “Just as Korea’s physical factories powered industrial growth, these AI factories will drive digital transformation,” said a government spokesperson. Blackwell chips can be used to build high-performance computing clusters capable of training large language models.