Nvidia will now make new AI chips yearly Leave a comment


Nvidia simply made $14 billion price of revenue in a single quarter because of AI chips, and it’s hitting the gasoline from right here on out: Nvidia will now design new chips yearly as an alternative of as soon as each two years, in keeping with Nvidia CEO Jensen Huang.

“I can announce that after Blackwell, there’s one other chip. We’re on a one-year rhythm,” Huang simply mentioned on the corporate’s Q1 2025 earnings name.

Till now, Nvidia’s produced a brand new structure roughly as soon as each two years — revealing Ampere in 2020, Hopper in 2022, and Blackwell in 2024, for instance.

However analyst Ming-Chi Kuo reported earlier this month that the following structure “Rubin” is coming in 2025, giving us an R100 AI GPU as quickly as subsequent 12 months, and Huang’s feedback counsel that report is likely to be on the cash.

Huang says Nvidia will speed up each different type of chip it makes to match that cadence, too. “We’re going to take all of them ahead at a really quick clip.”

“New CPUs, new GPUs, new networking NICs, new switches… a mountain of chips are coming,” he says.

Earlier on the decision, when an analyst requested him to elucidate how the current Blackwell GPUs would ramp whereas Hopper GPUs are nonetheless promoting properly, Huang defined that Nvidia’s new generations of AI GPUs are electrically and mechanically backward-compatible and run the identical software program. Clients will “simply transition from H100 to H200 to B100” of their present knowledge facilities, he says.

Huang additionally shared a few his gross sales pitches on the decision by the use of explaining the unbelievable demand for Nvidia’s AI GPUs:

We count on demand to outstrip provide for a while, as we transition to H200, as we transition to Blackwell. Everybody’s anxious to get their infrastructure on-line. And the explanation for that’s as a result of they’re saving cash and making a living and so they want to do this as quickly as potential.

He additionally has a FOMO argument that made me smile:

The subsequent firm who reaches the following main plateau will get to announce a groundbreaking AI, and the second after that will get to announce one thing that’s 0.3 p.c higher. Do you need to be the corporate delivering groundbreaking AI, or the corporate, , delivering 0.3 p.c higher?

Nvidia’s CFO apparently says that automotive will likely be its “largest enterprise vertical inside knowledge heart this 12 months,” pointing to how Tesla bought 35,000 H100 GPUs to coach its “full-self driving” system, whereas “shopper web firms” like Meta will proceed to be a “sturdy development vertical,” too.

Leave a Reply