Nvidia says the next AI chip is already a hit

The AI money machine keeps printing

Nvidia didn’t just report a monster quarter — it used the call to remind everyone that the AI party is still in full swing. Revenue jumped 85% year over year to $81.6 billion, earnings came in above Wall Street’s guess, and the company’s second-quarter guide also landed above estimates. In other words: the bar was high, and Nvidia still found a way to hop over it like it was a speed bump.

Inference is the new gold rush

The spicy part wasn’t just the numbers. CEO Jensen Huang said Nvidia is growing share in AI inference “very, very quickly,” which is the part of AI where models actually do useful stuff — answer your prompt, power a search result, or make your chatbot sound like it knows what it’s talking about. He said the customer base is widening beyond the usual hyperscaler crew, with names like Anthropic, Perplexity, and Cursor leaning harder on Nvidia’s stack.

And the plot thickens: Huang also said Anthropic’s expanding compute footprint across Amazon Web Services, Microsoft Azure, and CoreWeave is a meaningful boost for Nvidia’s inference business. Translation: more AI workloads, more Nvidia chips, more reasons for investors to keep treating the stock like the S&P 500’s most glamorous overachiever.

Vera Rubin is already getting the halo treatment

Then came the future-gazing. Huang said Nvidia’s next-gen Vera Rubin platform should be “even more successful” than Blackwell and that every frontier AI company will want in early. That’s a big claim, but Nvidia has earned the right to sound a little cocky — the company keeps showing up with the hardware, the software, and the demand curve all pointing in the same direction.

Big picture: Nvidia isn’t just selling chips anymore. It’s selling the plumbing for the AI economy, and right now the market is still happy to pay up for that privilege.