Nvidia's record quarter, explained simply — and why it matters

You might not buy Nvidia chips directly, but its quarterly results are one of the best available ways to see how big the AI boom actually is — and whether it's still accelerating. Here's the plain-English version of a very large number, and why it matters beyond finance.

On 20 May 2026, Nvidia reported $81.6 billion of revenue in a single quarter — the three months ending 26 April 2026. That is up 85% from the same quarter a year before, and up 20% from the quarter before that. To put it in perspective: that is more money than many countries' governments spend in a year, earned by one company from selling chips in three months.

Most of it — about $75.2 billion, or 92 pence in every pound of revenue — came from data-centre chips. These are the machines that fill the giant warehouses (called data centres) that run AI. The rest came from gaming graphics cards, professional design tools, and car chips — but these barely register compared to the AI hardware business.

Here is what the quarter actually looks like in two lines:

Nvidia's data-centre sales have roughly doubled in a year — that 92% jump means almost twice the AI-chip revenue it booked in the same quarter last year — and they're still climbing.

All the AI tools you hear about — ChatGPT, Claude, Gemini, image generators, coding assistants — run on the kind of hardware Nvidia sells. So when Nvidia reports results this strong, it tells you two things: first, AI companies and big cloud providers are still spending enormous amounts building more capacity; and second, new AI tools and models will likely keep arriving at pace, because the infrastructure to run them is expanding.

Nvidia also decided to hand a lot of its cash back to investors: $80 billion added to its share-buyback programme, and a dividend raised to $0.25 per share. That's the financial equivalent of saying 'we expect to keep earning this well for a while' — companies don't hand back $80bn if they think the good times are ending.

Nvidia's current bestseller is a chip family called Blackwell 300. The next generation is called Rubin, and the company confirmed in May 2026 that it's already being produced in the factory — with the first partner systems expected to arrive in the second half of 2026. For AI companies, Rubin is meant to be faster, more power-efficient, and better at connecting large clusters of chips together. If it launches smoothly, expect another leg of growth. If it runs into delays, that's worth watching as a signal for whether the AI buildout pauses to catch its breath.

The simple bottom line: the AI boom is still running at full speed, and Nvidia's results are the proof. The tools you use every day are powered by chips like Nvidia's, the companies building those tools are spending more than ever, and the next generation of hardware is already on its way. For now, the AI industry and Nvidia are telling the same story — and the story is still accelerating.