John Le Carré would have discarded as too implausible the Jensen Huang-Chinese ‘spy’ story. One of the world’s most influential businessmen caught in a ‘honey trap’ inadvertently handing over to the Chinese the most potent weapon in the new ‘Cold War’. Really?
I happened to be in China when the story broke. I was giving a talk at a provincial university. One the students asked me what I thought about the Jensen Huang ‘affair’ and Chinese ‘spies’ The professor hastily shut him up supposedly to save me from embarrassment but more likely his own since university authorities would ask why such ‘sensitive’ subjects were being aired in public. I was rather reassured that the news had got past the censors.
It is very unlikely that Alice Huang, the girlfriend, is a Chinese spy. President Trump, admittedly not a totally reliable source, has dismissed initial allegations linking Ms Huang to the Chinese Ministry of State Security. Ms Huang’s company, Megaspeed, had been a customer of Mr Huang’s Nvidia and at worst there may have been some technical infringements of US export licensing rules. US intelligence sources now talk of Alice Huang’s ‘links to China’: a rather inclusive definition of a ‘spy’.
The truth appears to be much more interesting than a spy thriller and not only because Jensen Huang’s sex life has led to his removal, for now, from the leadership of the world’s most valuable company. The story has brought into the open an important debate within the Washington establishment on how to deal with China’s rise.
One faction in the administration and Congress argues that the best way to maintain the US lead in frontier Ai is to keep all advanced ‘chips’- mostly produced by Nvidia -out of the hands of Chinese companies by imposing strict export controls. Act tough and demonstrate that the Chinese cannot function without US technology. Geopolitical sensitivity is heightened by the fact that the ‘chips’ are made in the world’s most advanced, TSMC, foundry in Taiwan and form the GPU’s (General Processing Units) by Mr Huang’s Nvidia which dominates the global supply of memory chips. These are in turn used for training the AI large language models.