Jensen Huang resigned this week after back-to-back cheating scandals — including an affair with a woman the U.S. government suspects is a Chinese spy — and the internet responded the only way it knows how: by feeding his likeness into the very machines he built a $4 trillion company selling. The AI boom’s patron saint, memed into oblivion by the technology he made cheap. There’s a Greek tragedy in there somewhere, if Greek tragedies came with pixel-art sweaters priced at $178, which Nvidia actually sold this year with the man’s face stitched onto them. Cults of personality are always great value for money right up until the idol stumbles.
Let’s pause on the economics of that sweater for a moment, because it tells you everything about where markets were mentally. A company whose valuation rests on the assumption that AI reshapes civilization itself was merchandising its CEO like a boy band. When a firm starts selling founder-iconography, it has stopped being an equity story and started being a faith story. Faith stories don’t discount risk well. Ask anyone holding Nvidia through this week’s resignation whether the sacred-leather-jacket premium survived contact with reality.
The memes, naturally, are the lede for most outlets. Huang’s face superimposed onto the Coldplay kiss-cam scandal from last year — that immortal moment when a married tech CEO and his company’s head of HR got caught mid-embrace by a stadium camera — generated by a Belgian software engineer scrolling headlines at 1 a.m. Bollywood parodies, sitcom cutaways, music video spoofs. The crowd that worshipped Huang as the high priest of the machine intelligence revolution spent the week using that machine intelligence to roast him. Poetic? Sure. But the more interesting question is why sophisticated capital was ever buying into a personality cult in the first place.
Because make no mistake: this was a personality cult with a market cap. Huang was not merely a CEO; he was the provenance stamp on the entire AI trade. Every hyperscaler capex announcement, every “we can’t get enough H100s” lament, every sovereign AI fund circling San Jose — all of it laundered through the credibility of one man in one leather jacket. And now that man resigned following allegations involving an asset of the Chinese intelligence services. Set aside the tabloid theater. Sit with the geopolitical implications for a second, assuming anyone on the sell-side can spare a second between downgrades they should have issued months ago.
The consensus will tell you this is a governance embarrassment, a headline risk event, a digestion period. The consensus has an impressive track record of calling every crack in the AI trade a buying opportunity. But consider the actual architecture of the situation: the single most important company in the AI supply chain — the de facto monopoly supplier of the compute substrate — just lost, in a scandal, the leader whose personal relationships spanned Washington, Taipei, and Beijing. Huang’s diplomatic function was never on any balance sheet, but it was real. The man who could sip tea with Jensen Huang was not merely buying chips; he was buying access to the chokepoint of the century’s most strategic technology. Whoever replaces him inherits the chair, not the mystique.
History offers some guidance here, none of it comforting for the true believers. When Steve Jobs died, Apple thrived — but Jobs had institutionalized his role for years before departing, and Apple’s moat was consumer lock-in, not critical national infrastructure. When Jack Welch left GE, the vaunted succession machine produced Jeffrey Immelt and a decade of value destruction so comprehensive it should be taught as a warning at every business school. When Carlos Ghosn fell — spectacularly, in a jet escape worthy of the memes — Nissan began its long decay. The pattern: founder-king companies don’t transfer power gracefully, because the intangible asset the market capitalized was never actually the company’s. It was the man’s. And intangible assets get marked to market in a hurry.
There’s also the small matter of the spy allegation, which mainstream financial media has treated with the enthusiasm of a hangover. If the U.S. government believes Huang’s paramour was a Chinese intelligence asset, the questions are not about memes. They’re about what the most valuable chip design IP in the world was exposed to, who knew what and when, and what this does to the increasingly paranoid regulatory posture toward Nvidia’s China business — which, lest we forget, was already being strangled by export controls. Washington’s hawks now have a gift-wrapped pretext to ratchet restrictions further, request deeper audits, and slow-walk anything resembling Nvidia-Beijing rapprochement. Multiply that across the data center buildout timelines and you have a real supply-chain story, not a tabloid one.
And yet the exuberance machine hums along. The same analysts who never flagged valuation stretched past every historical analogue — Nvidia trading at multiples that would have made 1999’s sock-puppet evangelists blush — spent the week reassuring clients that governance turmoil at the keystone company is “contained.” Contained. That was the word, you may recall, of the 2008 crisis, month after month, right up until it wasn’t. Contagion is always contained until the day it’s systemic. The analogy isn’t perfect — Nvidia has real earnings, real cash flows, real dominance, unlike a CDO squared — but the psychology is identical: the story cannot be questioned because too many balance sheets depend on it.
Consider what the AI trade actually is at this stage of the cycle: a capex arms race in which four or five buyers are spending sums comparable to national budgets on infrastructure whose revenue payback remains, charitably, unproven. Every dollar of that spending is justified by projected AI monetization, and every one of those projections assumes compute demand curves that only go up and to the right. The stability of the supply chain at the center of that bet was Huang, personally. Contracts transfer. Vision doesn’t. Relations with Taipei’s political establishment, with hyperscaler CEOs, with the Pentagon — those were the man, not the org chart.
The meme economy, meanwhile, is doing what meme economies do: converting existential uncertainty into entertainment so nobody has to think about it. There’s something almost admirable about it. A Belgian engineer stays up late making kiss-cam parodies while trillions in market value reprices around a scandal involving espionage allegations, and the collective verdict of the discourse is: ha, funny. The Coldplay couple, at least, only destroyed their own careers. Huang’s stumble lands on top of a market structure where retirement accounts, pension funds, and sovereign wealth are now all concentrated — via index funds that most participants don’t even realize they hold — in a trade whose central figure just walked out the door under a cloud.
So where does this leave us? Not, I’d caution, in the camp of the doomers either. Nvidia’s chips remain genuinely indispensable; the substitution problem for competitors is measured in years, not quarters. The demand from hyperscalers is real and contracted. But indispensable assets and indispensable men are different underwriting exercises, and the market conflated them for a decade at enormous multiples. The repricing of that confusion doesn’t have to be a crash. It can simply be the quiet disappearance of the founder premium that everyone swore didn’t exist until it did.
Huang, 63, built the greatest hardware franchise of the era and then tested the thesis that great men are irreplaceable. The memes will fade by next week. The succession question, the espionage questions, and the export-control ratchet will not. Markets spent years capitalizing a leather jacket. Now they get to find out what the company underneath it was actually worth.
