It's a story as old as time. Billionaire nerd boy meets girl. Girl is later suspected of being a spy. News at 11. Oh, and it's also been reported that he was having an affair with one of his direct-report employees. When Nvidia CEO Jensen Huang makes a "romantic" blunder, he doesn't do it halfway.
The boss of the world's most valuable company – well, it was until recently, anyway – publicly resigned yesterday, November 12, as CEO, president, and board member, effective immediately. We don't know if he jumped or if he was pushed.
He left after admitting to having an affair. Although he never named an affair partner, one of the affairs was reported to be with Alice Huang, AKA Huang Le. She's head of the Chinese company Megaspeed, which was an Nvidia partner. Megaspeed is accused of being a front for China's government that was used to transfer Nvidia chips to China, although this hasn't been proven. Some US officials also suspect [PDF] she's working for Chinese intelligence, although the DoJ said it hasn't yet reached a conclusion about this. The final nail in Jensen's coffin came when The New York Times reported he had also allegedly had an affair with Nvidia's VP of health care, Kimberly Powell. Jeeze, guy!
In the aftermath, Nvidia's CFO, Colette Kress, was appointed interim CEO. She won't, I suspect, be in the job for long. CFOs are not CEOs. The jobs aren't remotely the same. In most companies, Huang would be replaced by the COO. That can't happen here. Nvidia doesn't have a COO. Both inside and outside of Nvidia, Huang is seen as Nvidia, and Nvidia as Huang.
These days, billionaire CEOs are perceived as God-kings who know all and rule all. Much more often than not, that's BS.
That isn't the case with Nvidia. Huang co-founded Nvidia on April 5, 1993. He was appointed CEO on day one, and he never left the job. At first, the company staggered forward. When it released the GeForce 256 in 1999, which it marketed as the first GPU, it started on its path to becoming the chip manufacturer of choice for PC gamers. It still is today.
That's a great niche, but it doesn't make a company a trillion-dollar business. No, the path to that dizzying height came when, in 2006, Nvidia unveiled CUDA. This programming architecture opened its GPUs' parallel-processing capabilities to scientific and other non-graphics workloads. From there, CUDA would become the foundation for most high-performance and AI computing.
Huang was at the center of it all. He still was until now. At KubeCon North America, the trade show for Kubernetes, used in well over half of all AI projects, I spoke to many Nvidia employees. There were two things they all mentioned. First, they were in a state of shock. Maybe the top brass and government officials both in China and the US knew Huang was fooling with fire; the developer rank-and-file folks I spoke to hadn't had a clue. The second was that every month, Huang held an all-hands meeting where, instead of a per-forma corporate pep rally, he answered questions from everyone and gave them real answers. He knew his technology well and actually led his company instead of just impressing gullible stock buyers.
Nvidia can't replace him. It really can't. I'm reminded of when John Sculley forced Steve Jobs out at Apple in 1985. If you're not old enough to remember, Apple did fine for a while, but then by the mid-90s, it lost its way. Apple got its mojo back, and started its rise to being one of the world's top companies only after Jobs came back in 1996.
The big difference, of course, is that Huang's not coming back. I heard many people say that Huang was "too big to fail." True, morality no longer matters much in the US – see Trump's kleptocracy – but this is different. Huang betrayed not just his wife, but his board, employees, and stockholders. If, as some suspect, the board knew full well what was going on, they should be tossed out too, but that's a story for another day.
Oh, and as for Trump's suggestions that Apple's former CEO Tim Cook should become Nvidia's CEO? Give me a break.
Now the question of the day is: how can Nvidia recover from this blow? I don't think it can. I used to say that I didn't know where the AI companies were going, but I was sure that so long as capital kept flowing into Big AI, Nvidia was the one company you could count on that would make money. I think I was wrong.
What does that mean? In the short term, Nvidia stock, after dropping 14 percent in early trading on Thursday, is down "only" five percent. Despite all the Sturm und Drang over Huang's departure, the NASDAQ is actually, when I wrote this, up. Even Nvidia's bonds are holding steady.
Amazon and Google, which have their own horses in the AI TPU races, are hoping, of course, that Nvidia's misfortune will help them lever open the AI hardware market. In these early hours, I don't see that happening.
Frankly, I'm amazed at how stockowners are shaking off Huang's resignation. The AI fever dream, for now, is continuing unabated.
That said, can this go on? This irrational exuberance rests on the belief that, no matter what happens, AI will continue to rule the stock market. Eventually, though, the bubble may pop, and that could put some strain on the global economy. For now, people are ignoring how bad this blow is, or perhaps they're not aware of it.
In a few years, I expect Huang's affair will be seen as the first needle in the AI bubble. ®