NVIDIA (NVDA, Financial) is already the clear leader in the artificial intelligence chip segment, and the company scored yet another win by getting International Business Machines (IBM, Financial) to use its NVIDIA Tesla P100 Pascal GPUs through NVIDIA NVLink.
Last week IBM revealed a series of Linux-based servers which, according to the company, deliver higher levels of performance and greater computing efficiency than any other x86 based server.
Why IBM chose NVIDIA
IBM has been steadily building its cognitive division, and its analytics segment alone is now nearing $5 billion in quarterly revenues. Since the decision to put Watson at the center of things the company has been expanding its AI capabilities, and the decision to use NVIDIA’s products clearly underscores the lead NVIDIA has created in this segment.
“The open and collaborative model of the OpenPOWER Foundation has propelled system innovation forward in a major way with the launch of the IBM Power System S822LC for High Performance Computing,” said Ian Buck, vice president of Accelerated Computing at NVIDIA. “NVIDIA NVLink provides tight integration between the POWER CPU and NVIDIA Pascal GPUs and improved GPU-to-GPU link bandwidth to accelerate time to insight for many of today’s most critical applications like advanced analytics, deep learning and AI.”
Despite NVIDIA’s leadership position in this segment, the competition is not too far away. Intel (INTC, Financial) seems to be at least a year behind, but with accelerated focus on Internet of Things, data center and artificial intelligence being the keys to its future, that gap could narrow considerably.
Significant client acquisitions
But in the meantime NVIDIA’s Tesla GPUs seem to be scoring client after client since the start of 2016. Baidu (BIDU, Financial), the Chinese web services company that operates the country’s largest search engine of the same name, is using NVIDIA’s single-processor computer called the Drive PX 2 to power its self-driving car. Moreover, the Baidu Silicon Valley AI Lab is also using NVIDIA processors to help train its Deep Speech 2 speech recognition system:
“Our software runs on dense compute nodes with 8 NVIDIA Titan X GPUs per node with a theoretical peak throughput of 48 single-precision TFLOP/s.”
But that’s not all. Two of the world’s top 10 supercomputers use NVIDIA’s Tesla K20X GPU accelerators.
So NVIDIA clearly wants to leverage its GPU products in the data center and deep learning space, both of which are growing rapidly on the back of increased cloud adoption and artificial intelligence proliferation.
As such, the company is ideally positioned to exploit the potential in these two segments. In its press release, IBM noted that the new servers can be acquired at lower prices while delivering “80% more performance per dollar than latest x86-based servers.”
For NVIDIA, these might be small wins, but they are coming in at an alarming frequency. In this particular market segment there isn’t one single competitor that can challenge its products, and as such the company is clearly headed toward absolute dominance unless others catch up with it in the next year to two. By that time, however, NVIDIA will likely stretch the gap once again with the next generation of products.
Intel’s focus on IoT and data centers could be aided by the fact that it's now given up on the mobile-processing segment, but the question remains: Will Intel be able to catch up with an ever-accelerating NVIDIA?
Disclosure: I have no positions in any of the stocks mentioned above and no intention to initiate a position in the next 72 hours.
Start a free seven-day trial of Premium Membership to GuruFocus.
