The AI boom is real, but the bill is coming
Nvidia’s stellar results confirm the AI boom is ongoing, but soaring compute costs and weak monetisation show the economics are still far from solved.
Nvidia’s blowout earnings last week revealed something important about the AI boom. The numbers are extraordinary and the demand is real — and yet the market cannot quite decide what story it is living through. Is this the start of a megacycle or the prelude to a reckoning?
Nvidia added almost USD300 billion to its market cap in a single session, with revenue up 62% year-on-year and Blackwell orders sold out quarters ahead. But within hours the rally had faded. Nvidia’s share price reversed more than 2%, the Nasdaq gave back early gains to finish down 1.6% and Wall Street’s volatility index jumped more than 15%. The boom is accelerating, but so are the nerves beneath it.
This tension sits at the centre of the AI market. Nvidia is thriving at a scale that almost defies belief, but the economics of the companies building on its hardware remain deeply strained. The cost of running AI at global scale is rising far faster than the revenue it generates.
Nvidia is the clearest beneficiary of the AI cycle, providing the high end chips needed to train and run frontier models. But if the companies relying on that infrastructure cannot make the economics work, investor sentiment will continue to swing between exuberance and anxiety, no matter how impressive Nvidia’s financials are.
