
Global Policy Watch #1: AI Took Away Jensen’s Job
Global issues relevant for India
—RSJ
Last week when we in India were consumed with (and consuming) kaju katlis and patakhas, inhaling toxic air as part of our hoary parampara and busy killing ourselves in the name of custom (because which mai ka laal liberal can stop us from doing that?), a different bombshell dropped on the world of AI, threatening to upend a sector which is also in the race to kill us.
I’m, of course, talking about l’affaire Jensen Huang. The Nvidia founder, the most important figure of the global AI revolution, found himself embroiled in a scandal with an alleged Chinese intelligence operative. As more details emerged, including the timeline of his involvement, the questions about his influence on the Trump administration's China policy, his role in diverting banned advanced chips through intermediaries in Singapore and his character as more affairs with his colleagues came to light, it became clear that Huang was living on borrowed time.
That he could do all of these while also running a company that last quarter generated over $60 billion of free cash makes Mr N.R. Narayana Murthy, the preeminent advocate of the 70-hour work week, look like a slacker. Huang eventually resigned after a week of skeletons tumbling out of his trousers closet with some terrific memes and jokes at his expense. It is quite a fall from grace for someone who was shaping up to be the most consequential business leader of this century.
The Jensen Huang story is a fairy tale. Born in Taiwan and shipped off to the U.S. at the age of nine, he studied hard, funded himself with odd jobs, including working as a dishwasher at Denny’s (repeated ad nauseam) before turning into an entrepreneur. Nvidia started off making chips for gaming companies (that’s how I met them first back in 2005) and soon found the going tough. There was a steady demand but never enough to support Jensen’s ambition.
It came close to going belly up a few times, and for almost a decade its stock price went nowhere. But Jensen persisted, believing the future to be in GPUs, and by 2016 things started falling into place as Nvidia became the chip to go for heavy computing workloads. The real breakthrough came with the launch of ChatGPT. From then on, Nvidia became synonymous with the AI revolution. It went from a market cap of $350 billion to $4 trillion in about three years. It took many things to fall into place for this to happen, but there was no denying the primary driver of this was Huang’s conviction and his ability to build the company around it.
In the process, he became the voice of AI. He argued for wider and more open access to it around the world. In his worldview, AI is more like electricity at the start of last century than nuclear weapons during the Cold War. It is a foundational technology on which the future will be built. That explains why he did not see the restrictions on the sale of advanced Nvidia chips to China through the lens of American geopolitics. Of course, there was a commercial angle here. Nvidia sold chips, and China is its largest market outside of the US.
Megaspeed, the Singapore-based company whose executive Jensen Huang had an affair with, was a beneficiary of this view. With operations and data-centre arrangements in Malaysia, it emerged as a significant buyer of Nvidia hardware out of nowhere in 2022. It bought billions of dollars of advanced chips sitting in Malaysia, built data centres and rented computing capacity to China. This was a legit business model (renting to China wasn’t banned), though basic due diligence, as done by the media outlets, could have shown to Nvidia that the provenance of these companies was suspect. It is almost certain, based on US investigations and media reports, that these were fronts used to circumvent export restrictions.
The larger point is what this tells us about the AI race and the relationship between technology and geopolitics. I have seven questions.
Q1. Why has this story broken now?
The concerns around companies like Megaspeed and the movement of Nvidia chips through Southeast Asia aren’t new. NYT itself broke this story first almost a year back. One hedge fund published a damning report on this and shorted Nvidia on its China play about six months back. I think the Trump administration kept the story under wraps because it knew how central Jensen Huang is to the US AI sector and possibly because the spy angle didn’t surprise them based on their experience with China.
The story really took wings last week when NYT claimed to have leaked documents that showed the government knew Huang was having an affair with a possible Chinese spy. Whoever leaked it was either worried that this investigation was about to be buried or closed, perhaps because there was no evidence of the woman being a spy, or they were China hawks who wanted to bring down Jensen and create a China scare. The ease with which other stories of affairs came about right after the first news suggests a planned set of leaks given to the media. We will have to wait for the real evidence. I’m somewhat convinced the timing of the news was done in a manner to make it look worse than a corporate leader caught having an affair (which is quite frequent and not always a reason to resign).
Q2. Why would China go to such lengths to acquire Nvidia chips?
The answer is that China still needs them. All the talk of China being able to reverse engineer everything and then scale production doesn’t work for GPUs. Chip design is a bit more than that, and it requires accumulated knowledge and engineering nous that can’t be replicated overnight. Huawei is still a generation behind Nvidia, and its rate of innovation is slower. If these domestic champions were half decent, then you’d have the Chinese frontier AI models using them. Almost every single one of them has been built on Nvidia chips. All this Chinese skulduggery of using spies to influence policy on chip exports, stealing IP or smuggling chips through transhipment routes will only strengthen the hand of China hawks in Washington to go for much tighter export controls and penalties.
Q3. What happens to the idea of open access to AI computing?
Huang was the loudest and the most ardent champion for keeping AI open and letting his chips go wherever the demand was. Huang’s argument was simple. If America shut China out, it would only accelerate China’s efforts to build its own chips and models. Plus, the American companies would lose a very large market. There is some logic to this. But as the full extent of what AI could do or disrupt became evident, the national security argument has become harder to dismiss. Even the U.S. proprietary model companies like Anthropic and OpenAI were uncomfortable with Jensen’s stance on open weights. With Jensen gone and the Bluenami that has swept Democrats into the Senate and the Congress, expect more AI regulations and demand for a kill switch once wrong end use is identified.
Q4. Can Nvidia function without Jensen Huang?
Huang was not just any CEO. He was Nvidia for all practical purposes. He operated with around 60 direct reports, which is borderline ridiculous. He saw the big picture, connected the dots and had an instinctive sense of where the industry was going based on what he saw and heard from his customers. The company has enormous institutional capability with long-tenured experts at all levels who have built a juggernaut, but replacing Huang’s market instincts will be a challenge. The other thing that might worry Nvidia is further revelations of other possible honeytrap victims in its leadership team beyond Jensen. I mean, if China could reach up to the very top, I’m sure they would also be trying to entrap the next level …
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit publicpolicy.substack.com
More episodes from "Anticipating the Unintended"



Don't miss an episode of “Anticipating the Unintended” and subscribe to it in the GetPodcast app.








