After a quarter where Nvidia’s Sales have nearly doubled, with investors and analysts wondering how long the chipmaker can continue that kind of growth given its $140 billion in annual revenue.
Those hopes rest on Blackwell, Nvidia’s name for a family of server products centered around its next-generation AI chip.
CEO Jensen Huang and CFO Colette Kress gave investors several new data points about how Blackwell’s startup is shaping up on a call with analysts on Wednesday. The pair emphasized that the rollout is on track, and indicated that Blackwell’s sales in the next few quarters will be limited by how many chips and systems Nvidia can build, not how much it can sell. is
Blackwell’s production is fully underway, Huang said. “We will deliver more Blackwells this quarter than we previously anticipated.”
The company’s positive comments on Blackwell are one reason the stock is down just 1%, despite the company quickly missing high expectations from investors who had expected Nvidia to significantly exceed its forecasts. will go
Huang and Kreis’ comments also addressed concerns about shipment delays that Reports That said Nvidia is making ongoing engineering changes to its systems to address the issues.
Nvidia’s most important end users have already received some Blackwell chips, the company confirmed Wednesday. Microsoft, Oracle and OpenAI have posted images of Blackwell-based server racks on their social media accounts, and on Wednesday, the company said 13,000 Blackwell chips had already shipped to customers.
“There’s still a lot of engineering that goes into this,” Huang said. “But as you can see from all the systems that are standing up, Blackwell is in pretty good shape.”
Those sample chips aren’t a big part of the shipments the company expects to make. They are early versions intended to allow customers to begin testing and prepare their systems and software for volume shipments, which will begin in Nvidia’s current quarter.
“We will be sending more Blackwells from this next quarter. [quarter]And we’ll ship more Blackwells next quarter than we did in the first quarter, Huang said.
In July, Nvidia said it expected Blackwell’s revenue to be “several billion dollars” in its current quarter, and on Wednesday, the company said it expects Blackwell’s sales for the quarter to be the same as its original Expected to be higher than forecast. Huang also said that Microsoft will soon begin offering its Blackwell-based systems to cloud users.
A limiting factor in building more Blackwell systems is the amount of components Nvidia’s suppliers can provide, Huang said. Additionally, it takes time to ramp up a manufacturing process that has gone from zero shipments to multi-billion dollar shipments in a matter of months.
“It’s a case of demand outstripping our supply, and that’s to be expected because we’re at the beginning of this creative AI revolution,” Huang said.
He also named some of Nvidia’s “great partners.” TSMC, Amphenol, Virtue, SK Hanks And Micron.
“Almost every company in the world seems to be involved in our supply chain,” Huang said.
Nvidia said Blackwell’s gross margin will decline in the coming months from the 73.5% it reported in the third quarter, but the company said margins will increase as the product matures. Huang pointed out that Blackwell comes as a chip only or in configurations that include an entire rack and other components.
Nvidia’s overall message on Wednesday was that its new Blackwell chip is in short supply because companies like OpenAI need GPUs available as soon as possible to develop next-generation AI models. As Blackwell rolls out, Nvidia’s existing AI chips, which they call Hopper, will be deployed to service AI models instead of creating new ones. Nvidia said Blackwell’s sales will eventually outpace Hopper’s.
“Now you’re looking at the tail end of the last generation of foundation models, we’re at about 100,000 Hoppers,” Huang said. “The next generation starts with 100,000 Blackwells.”
Look: Nvidia can still grow even with Amazon and Microsoft entering the space: Susquehanna’s Chris Roland
