Nvidia earned record revenue of $35.1 billion during the last quarter thanks to "exceptional" demand for its current range of GPUs and a "staggering" clamour for next-generation hardware.
Its results for the third quarter of 2024 revealed that revenue was up 17% on the previous quarter and 94% from a year ago. The GPU giant also announced its best-ever data centre earnings, with revenue in this segment soaring by 17% sequentially and 112% year on year to $30.8 billion.
On an earnings call, Jensen Huang, President and Chief Executive Officer, said "the tremendous growth" is fuelled by "two fundamental trends".
"First, the computing stack is undergoing a reinvention, a platform shift from coding to machine learning, from executing code on CPUs to processing neural networks on GPUs," he said. "The $1 trillion installed base of traditional data center infrastructure is being rebuilt for Software 2.0, which applies machine learning to produce AI.
"Second, the age of AI is in full steam. Generative AI is not just a new software capability but a new industry with AI factories manufacturing digital intelligence, a new industrial revolution that can create a multi-trillion-dollar AI industry."
When will Nvidia Blackwell be released?
Blackwell, Nvidia's next-generation GPU, is now in full production, and 13,000 samples have already been shipped. It is being integrated into a wide range of data centre setups as customers prepare their infrastructure.
Oracle, for instance, has unveiled the world’s first Zettascale AI cloud clusters, capable of scaling to over 131,000 Blackwell GPUs, to support the training and deployment of cutting-edge AI models.
During its earnings call, Nvidia sidestepped a question about reports of overheating when Blackwell GPUs are connected together into server racks holding up to 72 units.
Huang said: "There's still a lot of engineering that happens at this point. But as you see from all of the systems that are being stood up, Blackwell is in great shape."
"The integration of all of those systems into the world's data centers is nothing short of a miracle," the Nvidia CEO continued. "And so, the component supply chain necessary to ramp at this scale, you have to go back and take a look at how much Blackwell we shipped last quarter, which was zero.
"And in terms of how much Blackwell total systems will ship this quarter, which is measured in billions, the ramp is incredible. And so almost every company in the world seems to be involved in our supply chain."
Soaring Blackwell demand
Blackwell production is now in "full steam", he added, with Nvidia expected to smash its own targets.
"It is the case that demand exceeds our supply," Huang revealed.
Nvidia built seven custom chips to deliver the Blackwell systems, which can be used in both air and liquid-cooled contexts: NVLink 8, NVLink 72, NVLink 8, NVLink 36, NVLink 72, x86 and Grace.
During recent benchmark testing, Blackwell achieved a 2.2x performance improvement over Hopper - the current generation of GPU.
"Blackwell demand is staggering, and we are racing to scale supply to meet the incredible demand customers are placing on us," said Colette M. Kress, Chief Financial Officer and Executive Vice President. "From liquid-cooled to air-cooled, every customer is racing to be the first to market."
What about Hopper?
However, Hopper is expected to keep selling well. Kress described it as "an amazing product" that's "the fastest growing and ramping that we've seen".
She predicted that it would continue to sell throughout the next quarter and into next year, even though customers are "looking to build out their Blackwell".
Nvidia shouted out xAI's Colossus supercomputer, which draws on 100,000-Hopper GPUs and its Spectrum-X ethernet AI networking platform, achieving "zero application latency degradation" and hitting 95% data throughput compared to 60% for traditional ethernet.
"Hopper demand is exceptional," Kress said.
