Cloud computing continues to grow in importance. And with it, artificial intelligence — or AI — is an increasingly critical component. It’s also extremely resource-intensive, requiring huge amounts of computing power. Setting up a cloud computing center that can support AI is time consuming, complex and scalability is a problem.
But Nvidia Corporation (NASDAQ:NVDA) is making a move to put itself at the center of cloud computing and AI with the announcement of the HGX-1, a new architecture standard featuring NVDA hardware, developed in partnership with Microsoft Corporation (NASDAQ:MSFT).
If HGX-1 takes off, it could do for cloud computing and AI what the Wintel partnership between Microsoft and Intel Corporation (NASDAQ:INTC) did for the PC era.
The Cloud Computing Challenges
Cloud computing hit new levels in 2016, with cloud revenue growing 25% on the year and no end in sight. Besides the basics of hosting applications, the cloud is increasingly tied up with AI. Personal digital assistants like Apple Inc.’s (NASDAQ:AAPL) Siri may be accessed from a local device like a smartphone, but all the computing needed to analyze speech and spit out answers is done in the cloud.
That’s the only way to get the massive computer firepower needed for AI and make it accessible to the devices, apps and services that need it. Autonomous vehicles are another example of AI, and reliance on the cloud.
There are several challenges with all this cloud-based infrastructure.
The first is that traditional CPUs aren’t up to the task of crunching the data. At least not affordably. That’s where Nvidia comes in. Its graphics cards have proven to be the ideal solution, providing massive processing power at a reasonable cost.
The second issue is that with so many companies pursuing cloud computing solutions and such a range of applications, configuring a cloud computing center is a big job. Scalability of custom solutions is also a problem.
What’s needed to keep costs down and support rapid expansion is standardization.
Nvidia and Microsoft Position HGX-1 as the Solution
Microsoft knows a thing or two about cloud computing. Its Azure solution is a big player in the cloud computing market, second only to Amazon.com, Inc.’s (NASDAQ:AMZN) Amazon Web Services.
Microsoft also knows first-hand the value of being a key partner in setting an standard. Working with Intel as part of the “Wintel” partnership helped make Microsoft the giant it is today,
And Microsoft sees NVDA as being critical to the future of AI and cloud computing. The company is switching to chips based on ARM architecture, from ARM Holdings plc (ADR) (NASDAQ:ARMH) for server CPUs as part of its Project Olympus initiative. One of Microsoft’s partners in Project Olympus is Nvidia.
Together, the companies are promoting NVDA’s new HGX-1 hyperscale GPU accelerator as the new standard for cloud-based AI. The idea is that Microsoft’s cloud servers with those new ARM CPU’s will provide the general purpose computing power. Nvidia provides the muscle. The HGX-1 is a modular box with eight GPUs inside it. NVDA says one HGX-1 provides up to 100 times the deep learning processing performance of legacy, CPU-based servers.
As additional processing power is needed for intensive cloud computing tasks such as AI, plugging additional HGX-1 modules into the system takes care of the additional demand. Nvidia and Microsoft are positioning this as a standard for cloud computing centers going forward, a solution that provides a huge amount of raw processing power, doesn’t require customization and offers easy scalability.
The personal computer era made giants out of Microsoft and Intel. Nvidia, was a small player for most of that time, content to focus on video cards. But now cloud computing and AI are the future and the demand for raw processing power is eclipsing traditional multi-purpose CPUs. Those video cards are the key and NVDA is taking full advantage of its position.
With HGX-1, the company is seizing the opportunity to become Microsoft’s partner for the new era. If successful, Nvidia will not only help to accelerate the development and deployment of cloud computing and AI, it will reap the financial benefits of being in the driver’s seat.
As of this writing, Brad Moon did not hold a position in any of the aforementioned securities.
