The Bubble That Isn't, Until It Obviously Is
Axios has discovered something remarkable: an AI bubble that exists entirely independent of whether artificial intelligence actually works. This is the logical apotheosis of late-stage venture capitalism—the ability to simultaneously declare a technology revolutionary and inevitable while admitting that the financial structures built around it are fundamentally unsustainable. It's not that the emperor has no clothes; it's that the emperor's wardrobe is on fire, but everyone agrees the emperor himself is real, so obviously we should keep buying designer outfits.
The crisis apparently hinges on a scandal involving Nvidia CEO Jensen Huang, though the Axios reporting treats this as a secondary shock to an already-fragile ecosystem rather than the inevitable outcome of building a $3 trillion industry on the premise that compute-as-theology requires no actual business model. Nvidia, of course, has built a genuinely exceptional semiconductor company—the kind that actually ships products, generates revenue, and doesn't require quarterly press releases about why it still hasn't monetized its core value prop. But when your entire sector is valued on the assumption that every company touching AI will eventually become a trillion-dollar enterprise, the fall of even one pillar doesn't just crack the foundation; it suggests the entire building was held up by investor FOMO and financial media bullying.
History, as Axios notes, shows that world-changing technology and financial bubbles often travel together—a point that somehow still manages to surprise people every time it happens. Dutch tulips. Dotcom stocks trading at 10,000x revenue. Subprime mortgage derivatives rated AAA. Each era produces leaders and investors who insist that *this time* the fundamentals have changed, that *this technology* justifies astronomical valuations, that only the insane would question whether the thing works. What they never seem to anticipate is that the market doesn't actually care whether the thing works; it cares whether you can convince someone else to buy in at a higher price.
The rhetorical sleight-of-hand here is exquisite: "No sane person questions whether AI works" is doing so much work in that sentence. It's not claiming AI is worthless; it's not even claiming current applications are overhyped. It's simply declaring debate heretical, then immediately pivoting to admit that the financial structure built on that unquestionable technology is visibly collapsing. This is what peak hype-cycle doublethink sounds like—the ability to hold two contradictory thoughts without the slightest cognitive friction.
The real risk isn't that investors finally realize AI doesn't work; they won't, because AI clearly does work, in limited, specific, increasingly useful ways. The real risk is that they realize it works at fundamentally different scales and timeframes than the valuation multiples implied. That a $500 billion company can be transformative without becoming a $5 trillion company. That not every application of large language models requires a 500-person Series D and a $10 billion valuation. The bubble pops not when the technology fails, but when it succeeds *too modestly*.
What Axios is really documenting is the moment when the financial industry admits it has no exit strategy—no honest way to reconcile current prices with actual revenue, actual unit economics, or actual competitive dynamics. The scandal involving Jensen Huang is just permission structure for a reckoning that was always mathematically inevitable. In another six months, we'll probably read another story insisting that no sane person questions whether AI works, while casually mentioning that another $200 billion in market cap evaporated overnight. At some point, distinguishing between a technology bubble and a financial bubble stops mattering—they're just different words for the same catastrophe.
"No sane person questions whether X works"