China's military will steer clear of U.S. AI technology, according to Jensen Huang, as they don't necessitate Nvidia's chips or American tech infrastructure to fortify their military.
In the ongoing saga of US-China technology relations, Nvidia CEO Jensen Huang has taken a nuanced approach, balancing support for American policy objectives with facilitating Nvidia's role in global AI growth, including in China's expanding AI infrastructure.
Huang has publicly supported the US government's efforts to maintain leadership in AI, while actively engaging with Chinese officials and industry to promote AI development in the region. He emphasises AI's potential to raise productivity and expand opportunities worldwide, including in China.
Despite initial US export controls that halted sales of Nvidia's H20 AI chips to China, Huang has worked to secure licenses allowing Nvidia to resume these sales. This move is linked to broader US-China trade negotiations, where export controls on AI chips are used as a bargaining chip in exchange for critical resources like rare earth elements needed for US manufacturing.
However, concerns persist in Washington about the Chinese military potentially using advanced US AI technology. The PLA is believed to be avoiding American tech, similar to how the US armed forces avoid Chinese products. Yet, some Chinese businesses are reportedly obtaining Nvidia's banned chips on the black market.
Huang has downplayed these concerns, stating that depriving someone of technology is a tactic, not a goal, and that this tactic was not in service of a goal. He has also criticised the US strategy of limiting China's access to advanced technologies as a failure.
The US is considering geo-tracking tech on high-end hardware as a potential measure to prevent the use of such technology in sensitive areas. If such a technology exists, China can air gap its systems to prevent remote disabling. Chinese operators of smuggled AI chips may have difficulties getting firmware updates and access to Nvidia's advanced cloud tools and enterprise platforms.
Despite these challenges, China is heavily investing in AI data centers and is sourcing Nvidia's most powerful products, despite export controls. China already has a significant number of supercomputers built by Chinese engineers, which do not require Nvidia's chips or American tech stacks.
In conclusion, Huang's stance reflects a balancing act between supporting US policy objectives and maintaining Nvidia's business and influence in China's AI sector. He promotes open-source AI research and a US technology stack as foundational to global AI innovation, while pragmatically navigating export control regimes to ensure Nvidia's continued growth in the global AI market.
Huang's approach fosters Nvidia's involvement in the global AI industry, supporting US policy while engaging with Chinese officials in finance and technology sectors. Amid concerns about the military use of AI, Huang advocates for a non-restrictive approach to technology sharing, viewing it as crucial for economic growth in artificial-intelligence.
