Nvidia's data center business hits $2 billion annual revenue run rate

The Volta ramp is just starting and deep learning and artificial intelligence with a dose of high-performance compute is driving some serious growth for Nvidia.
Written by Larry Dignan, Contributor

Nvidia's data center business has hit a $2 billion annual revenue run rate and is showing no signs of slowing down as cloud providers gobble up the company's GPUs.

The company's third quarter earnings results were stellar as all of its businesses showed strong growth. But what's really notable is how Nvidia's data center business has doubled revenue compared to a year ago.


Nvidia's third quarter non-GAAP earnings of $1.33 a share on revenue of $2.64 billion easy topped Wall Street estimates. Analysts were expecting earnings of 94 cents a share on revenue of $2.36 billion.

Nvidia's Tesla V100 GPU gets broad backing as server vendors eye AI workloads | Nvidia expands partnership, courses for Deep Learning Institute | Nvidia morphs from graphics and gaming to AI and deep learning | Computing shifts may rearrange pecking order for Intel, Qualcomm, AMD, Nvidia, chipmakers

With those results it's not all that surprising that Intel and AMD may wind up being frenemies aligned against Nvidia -- at least for a bit.

Read more: Intel aims to be inside your artificial intelligence stack

CEO Jensen Huang said the data center business -- powered by Nvidia's Volta platform -- is just starting to ramp.

As you know, we started ramping very strongly Volta this last quarter. And we started the ramp the quarter before. And since then, every major cloud provider from Amazon, Microsoft, Google to Baidu, Alibaba, and Tencent and even recently, Oracle, has announced support for Volta and we'll be providing Volta for their internal use of deep learning as well as external public cloud services. We also announced that every major server computer maker in the world has now supported Volta and in the process of taking Volta out to market.

HP and Dell and IBM and Cisco and Huawei in China, Insper in China, Lenovo, have all announced that they will be building service -- families of servers around the Volta GPU. So I think we -- this ramp is just the first part of supporting the build out of GPU-accelerated service from our company for data centers all over the world as well as cloud service providers all over the world.

Nvidia's Huang said it's possible that every supercomputer will have acceleration. If that bet plays out Nvidia will represent a lot more than the 15 percent of the world's top 500 supercomputers.

Here's why the wind is at Nvidia's back in the data center:

  • Deep learning training needs more compute power. The market is a natural extension of high-performance computing.
  • Inference will require hyperscale data centers that support billions of billions of queries. GPUs are likely to take inference workloads from CPUs.
  • The cloud will democratize artificial intelligence and those services will be built on GPUs.
  • High-performance computing will be used more in industry via the data center or cloud.

Huang said:

We have the ability now to address high-performance computing and deep learning training as well as inference using one common platform. And so I think the -- we've been steadfast with the excitement of accelerated computing for data centers. And I think this is just the beginning of it all.

Huang went on to say the Moore's Law era is over -- especially since one Volta processor can equate to 100 CPUs. For Nvidia, it's nothing personal with Moore's Law. It's just physics.

Editorial standards