Nvidia Isn't Just Selling AI Chips — It's Financing the Firms That Buy Them

Nvidia has quietly shifted from chip supplier to financier of the entire AI build-out. Rather than waiting for data-centre operators and AI labs to raise their own capital, the company is putting roughly $1 trillion in deals, cash, and guarantees behind its customers so they can afford to buy its GPUs. That role — funding demand for its own product and effectively setting the terms on which AI infrastructure gets built — is what earns it the ‘central bank of AI’ label.

The maneuver draws obvious skeptics. Critics liken it to the vendor financing that inflated Cisco’s revenues during the 2000–01 dotcom bubble, and cast CEO Jensen Huang as an illusionist propping up demand that might not exist on its own. The worry is circularity: if Nvidia is bankrolling the buyers of its chips, reported growth may reflect financial engineering more than genuine, self-sustaining appetite.

The Economist’s own read is more sanguine. If the bets pay off, cheap access to compute accelerates AI adoption and lifts productivity across the economy. If they misfire, the losses land mainly on Nvidia’s shareholders rather than the broader financial system — which, the argument goes, is how capitalism is supposed to allocate and price risk. The deeper significance is structural: one company now sits at the center of AI’s capital flows, making its balance-sheet decisions a systemic factor for the whole sector.