Jensen Huang's "compute is revenue" case for Nvidia's valuation

On the fiscal Q2 2027 earnings call, Jensen Huang distilled the bull case for Nvidia into four words: "compute is revenue." The argument reframes how investors should think about AI infrastructure spending, treating it as a billable, profit-generating output rather than a cost center on a customer's balance sheet.

The logic rests on what Huang called an inflection point in AI's maturity. Tokens produced by AI systems are, in his framing, productive work that enterprises can charge for. If that claim holds, demand for the compute that generates those tokens does not slow down the way discretionary capital expenditure does.

The framing shapes how anyone models whether Nvidia's customers will keep buying at the current pace. A compute cycle that sits on the cost side of a ledger is vulnerable to budget cuts. One that sits on the revenue side is self-funding, and cutting it means leaving money on the table.

The skeptical read is that Huang's framing is self-serving, and the earnings call is the natural venue for it. But the burden of proof shifts: any analyst underwriting a demand slowdown now has to explain why AI-generated revenue would decelerate fast enough to outrun the capital commitments already made. Until that case is made with hard data, Huang's reframe is the working thesis the market is pricing.