Will Intel Be Able To Gain An Upper Hand Over Nvidia In The Data Center Coprocessor Market?

Though Intel (NASDAQ:INTC) is the leader in data centers microprocessor market, with more than 90% market share, it has more recently lagged in the market for coprocessors, where Nvidia (NYSE:NVDA) has generated momentum. Whereas Intel is dominant in the microprocessor market at large, Nvidia has successfully identified this coprocessor niche where it has garnered significant market share. This is reflected in their recent growth rates. While Intel witnessed a mere 13% year over year growth in its revenue from the Data Center Group in Q3 2017,  Nvidia data center revenues nearly tripled during the same period.  Of course, Intel’s data center revenues are a vast multiple of Nvidia’s. Still, it is in the market for HPC (High-Performance Computing) coprocessors that they directly compete. And both firms are out with new offerings. Intel has recently released its Xeon Phi processor family (formerly code-named Knightsbridge). And Nvidia is out with its Tesla P100, the processor that powers its DGX-1 Deep Learning System (the so-called “Supercomputer in a Box”). Clearly, the battle is on.

Nvidia’s AI And Deep Learning Expertise Provides It An Edge Over Intel

It is notable that in the HPC segment of the market, Nvidia has been gaining traction. The graphics processing leader has developed a series of products for use as a coprocessor in advanced computing applications. Coprocessors and the accompanying software are used to offload compute-intensive operations from microprocessors to increase performance.  For example, in the most recent Top 500 Supercomputer report, 455 (or 91%) use Intel microprocessors, 55 (or 26%) use IBM Power microprocessors, and 13 (or 3%) use AMD microprocessors. And of the 500, 93 (or 19%) are designed to use accelerator/coprocessor devices, of which 67 (or 72%) use Nvidia coprocessors.  Nvidia is also a leader in artificial intelligence and deep learning and has even developed a range of products. These include both coprocessors and the Cuda software environment on which they run.

Nvidia claims that it could foresee the huge deep learning opportunity five years ago. The company also claims that it has been investing heavily to build a platform for deep learning in its GPUs since then, and the performance improvements it has brought is reflected in its current line of GPUs with Pascal architecture. With a strong surge in the data center revenues in the last few quarters, it looks quite evident that Nvidia’s efforts are paying-off well.

However, Intel Has Fortified Its Position With The Acquisition Of Nervana Systems

Earlier this year, Intel announced the acquisition of the deep learning technology startup Nervana Systems. According to Diane Bryant, executive vice president and general manager of Intel’s Data Center Group, Nervana has a fully-optimized software and hardware stack for deep learning and has the advanced expertise in accelerating deep learning algorithms, which can help Intel expand its capabilities in the field of AI (artificial intelligence).

The Nervana acquisition should allow Intel to enhance its existing line of processors for deep learning capabilities. This comes from the fact that Nervana’s Neon framework can exploit GPU capability and has been ranked fastest in various performance measures. Further, as we noted above,  the AI startup plans to deliver a chip based on deep learning technology in the coming year. This should help Intel from losing market share to Nvidia, which has the most advanced GPUs in the market currently.

Sources report that Nervana has gained traction against Nvidia with its Cuda-compatible Neon software offering. The company is also developing a Deep Learning accelerator (i.e., coprocessor ) that is expected next year. Surely, Nervana’s advanced HPC stack will find fertile soil, once it is integrated into Intel’s resource-rich environment. The acquisition of Nervana Systems is going to help Intel fortify its position versus Nvidia in the markets that develop to capture the power of these emerging compute technologies.

View Interactive Institutional Research (Powered by Trefis):

Global Large Cap | U.S. Mid & Small Cap | European Large & Mid Cap More Trefis Research