Hello Betamax,
Jensen Huang was supposed to be in Asia this week. Donald Trump is heading to the Philippines for the APEC Summit, then to China for the APEC Leaders’ Meeting, and Nvidia’s now-former CEO was expected to be by his side to sell chips to Asian companies.
Instead, Jensen now finds himself engulfed in a scandal few could have imagined. Two alleged affairs have come to light in the span of a week. He has admitted to one, involving Alice Huang, an alleged Chinese spy. The other reportedly involves Kimberly Powell, an Nvidia vice president who reports directly to him.
Image credit: Made by Ulla/Tech in Asia with the help of AI
On Thursday, the company announced his resignation, with CFO Colette Kress taking over as interim CEO.
Interestingly, many people in the AI industry aren’t surprised by the allegation that Megaspeed, the Singapore company where Alice was an executive, routed Nvidia chips to Chinese firms barred by the US from buying them. Following the scandal, a fifth person has been hauled to court in Singapore, accused of trading Nvidia chips without authorization.
Chip diversion has apparently been an open secret in the industry.
“It’s actually nothing new,” says Manoj Sukumaran, an independent analyst who covers the AI and chip industry. “People knew it since, I would say, the past one year or eight months.”
Donald Trump’s administration has largely turned a blind eye to this alleged chip diversion, helped in part by his close relationship with Jensen. Nvidia’s status as the world’s most valuable company and its considerable influence in Washington have been unprecedented.
This week, the New York Times reported that Nvidia got a US probe into Megaspeed halted. This episode alone shows how much influence Nvidia has on both Washington and the AI industry.
As the industry’s de facto central bank, Nvidia bankrolls companies struggling to raise funding elsewhere, helping keep the bubble inflated. But as the internet loves to joke, nobody expected Jensen to puncture it with something he should have kept in his pants.
Still, people in Asia’s AI computing industry have told us it’s business as usual. Nvidia’s chips remain too important for customers to abandon over a scandal involving its former CEO.
Bruce Yang, co-founder and CEO of Singapore-based model maker Sapiens AI, says Nvidia will remain the default choice for both training and inference at his company. He points to the “stronger ecosystem” around CUDA – Nvidia’s software platform for running computations on its GPUs – and notes that training and inference are “better optimized” on Nvidia chips.
Switching to a different chip architecture would mean redoing much of the company’s existing work, he adds.
The fallout may take longer to emerge in Southeast Asia, however.
Megaspeed is based in Singapore. The city-state, alongside Malaysia, has already faced investigations into the alleged diversion of Nvidia chips to China.
Meanwhile, for the company, the bigger risk isn’t lost sales but diminished clout in Washington as Democrats control both chambers of Congress.
Southeast Asia is where billions of dollars are pouring into AI infrastructure. The region is heavily dependent on Nvidia’s chips, and US restrictions have already disrupted its data center expansion plans.
“US export controls on advanced GPUs have already forced regional data centers to redesign capacity plans mid-deployment,” says David Sherman, AI and financial inclusion strategist at io.net, a decentralized GPU cloud company.
He adds that when supply tightens, Nvidia prioritizes hyperscalers. Smaller neocloud providers, including those in Southeast Asia and India, face longer waits or must pay a premium on secondary markets.
Reliance on a single chip supplier also raises risks for lenders, driving up financing costs, Sherman further notes.
In one of the latest developments, the US is finalizing rules to prevent Chinese companies from renting Nvidia chips they can’t directly acquire through data centers in countries such as Singapore and Thailand. That puts Southeast Asia’s neocloud providers squarely in the regulatory crosshairs.
US tech companies are increasingly looking to the region to expand data center capacity amid growing opposition to new facilities at home. Yet the same countries welcoming these investments face tighter restrictions on how their AI infrastructure can be used.
For sure, Nvidia’s dominance won’t disappear overnight.
The AI boom made Jensen one of the world’s richest individuals, with a fortune exceeding US$150 billion. But immense wealth and power can also breed impunity. His downfall is a reminder that even the most powerful figures in tech are not immune to scrutiny.
The stakes extend well beyond one executive. Tech leaders have repeatedly warned that tougher AI restrictions could hand China an advantage. Under Trump, that argument has helped the industry push back against tighter controls, particularly on advanced chips.
Southeast Asia has once again found itself caught in the middle. Markets here want to attract investment from American tech giants and become major AI infrastructure hubs. At the same time, they face growing pressure to ensure that the tech entering them doesn’t end up in the wrong hands.
Jensen may be gone, but Southeast Asia still has to navigate an AI race in which it is becoming a major investment destination without having much say over the rules.
Meanwhile, here are more AI stories worth checking out:
Pranav Balakrishnan, AI reporter (with additional reporting by Glenn Kaonang)
Catch up on past issues of the newsletter here.
|