Huang powers and bankrolls AI revolution: Nvidia CEO caught between 'a rock and a very hard place'

And for more on this story, I can welcome our guest today, Bernard Benhamou, Secretary General of France's Institute of Digital Sovereignty. Good morning, Bernard, and thank you very much for joining us. This case involving NVIDIA CEO Jensen Huang is interesting because if indeed he was having a relationship with a Chinese spy, even in the age of cyber warfare and hacking and and the ability to get at one's adversaries from a distance. The old methods still have their place. Why do you think this is? Because humans stay humans and at some moment the very classic methods that the Russians are very famous for are also used by Chinese spies for a long time. And at this moment the question is My question is, has it been used by the Chinese authority to help? uh push the agenda of China in terms of developing AI or was he simply unaware that he was helping the Chinese the the the debate and the The board who had to decide or to consider if he has to be ousted or not had a very, very difficult time, I guess. And there was another scandal going on with an employee who has been having a relationship with him. So at this moment, he is in a very, very tough position. Yesterday, we were talking about the possibility of him being ousted. Now it's the probability of him not being able to stay. Those are classic methods, but the stakes are so high that at this moment the Chinese couldn't help but try to influence the giant in the place. He's not only the biggest company in AI chips, he's also the banker of the global AI system. So he has an influence and a role who is central for all the tech ecosystem. We can say, well, so all the stock market, because that was for a long time the first market valuation for by far 5 trillion. Now it's 4.2, I guess. Thank you. It's a very sensitive issue for the global economy, not only for the tech ecosystem. Now, this case... If Alice Wang, the businesswoman implicated in the affair, was a Chinese spy, the Chinese deny it, would this have been necessary for China, given that it actually did access those chips ultimately through a deal with the Trump administration? They had some of the chips, not all of them. And what has been discovered in those recent days is the fact that Alice Wang was able to have In Singapore and maybe Malaysia, data centers running on... top-level chips by Nvidia to train the models of Chinese companies, which was something that shouldn't have happened. So in a way it was circumventing the measures by the Trump administration and that's also why he's in such a difficult position because he had a problem now with his board, with the company, with a class action by the stockholders, but also The security apparatus of the U.S. is really anxious about how and when those circumventing measures have been taken by him. So he is in between a rock and a very hard place. Now, it might be a little bit understandable if an American... Just a regular American were to be caught out by this sort of honeypot situation. But Jensen Huang was born in Taiwan. He speaks Chinese. He's Taiwanese-American. People from Taiwan are generally very sensitive to possible threats emanating from the Chinese government. Was he a little bit careless here? I guess there is. The fact that he had a relationship with a woman so obviously without hiding and after with an employee, the chief of the health division at Nvidia. Obviously he was having a kind of hubris that he thought everything was possible for him. And reality bites back. At some moment, he was not able to hide and to measure his behavior. And that's often the case when it's about personal relationships. Now, there was another story reported by the U.S. news website The Intercept yesterday, which is of less grave than the NVIDIA case. But interesting nonetheless, they reported that two senior executives at OpenAI were duped by Russian pranksters into thinking they were speaking to Ukrainian government officials. And the pranksters put the recording on the Internet. Now, for a sector that's always warning us about the dangers that their products potentially pose to humanity. People working, senior executives working in AI are fairly careless, aren't they? Most of them are very young and they are not trained. to be cautious about what they say and how it can be revealing of strategic information by their companies and indeed the spycraft has evolved with technology and now we have spies using AI and using technologies to fool their opponents and it works it works but I'm not surprised engineers most of those companies are engineers driven They are not political persons, they are not trained to think in terms of geopolitics, in terms of spying, they are not trained for that. Now, the AI sector has taken quite a beating on the stock market in the past week, what with the prospect of greater regulation following the US midterm elections and then also after the NVIDIA scandal. Is it unusually vulnerable to these sort of stock markets? sort of bubbles relative to other areas of tech? They are, because there are obviously some hyper-valuation, some risky valuation for those companies. The ratio, I will not try to be too technical about that, the ratio are in the red zone. So at this moment, you have both events colliding. The Democrats at the Congress were, for some of them, very opposed to developing data centers in the US. And on the other side, this scandal of the most prominent company in the field. So at this moment, AI is obviously at risk, not of a crash or a bubble burst, but a severe correction for the times to come. It's obvious. Now, AI is, it's not exactly a new technology, even though in recent years with large language models it's been thought of as that, and it has many very useful applications in many fields, but at the moment it has a fairly catastrophic image problem, be it from opposition to data centers or resistance to the perceived culturally corrosive effects. Is there a chance that Silicon Valley might eventually take stock of this? Obviously, it's the most recent revolution in tech, it's quicker than any revolution before, than the web, than the smartphones, than any of them, so indeed it goes very fast, and the very new thing about that is that it changes all sectors at the same time, health, culture, news, industry, it changes everything, so indeed The risks, we have been talking about those risks, especially this summer with all these incidents about AI being uncontrollable and attacking companies or governments. So at this moment, I guess there is some kind of hype of AI becoming sentient, AI becoming conscious. No, we are far from that. Super intelligence, like Trump likes to call it, is far from us, but indeed risky. systems in AI are obviously here and we have to regulate not to over regulate not to have a hands-on approach on everything no but indeed and what's new is even China is accepting that there are some negotiations between China and the US on risks on security for banks on security for weapons that have to have that will have to happen so at this moment we are in a In a tough situation, both on the regulatory front and on the business front. So it's a little bit tricky for those industries at that time. Thank you very much for that. Bernard Benhamou, Secretary General of France's Institute of Digital Sovereignty.