Partners: Intel Pushes Innovation Envelope With Acquisition Of AI Startup Nervana Systems
Partners are applauding Intel's news Tuesday that it will scoop up Nervana Systems, which specializes in deep learning and and neural networks, as the company doubles down on artificial intelligence.
Kent Tibbils, vice president of marketing at ASI, a Fremont, Calif.-based Intel system builder, cheered the acquisition as another example of "Intel's commitment to new technologies that leverage their core strengths."
"Although AI is still relatively new, Intel’s investment in this segment as well as their devotion to the health of the channel will ultimately help provide another path of opportunity for new products, new services and new solutions," he told CRN. "This isn’t to say that everyone will switch to AI-focused businesses models, but with Intel’s involvement there will be a market for the channel."
Deep learning often requires the powerful and efficient parallel computing capabilities of GPUs – which Nvidia specializes in – to teach machines how to process text, voice and other types of data as part of artificial intelligence.
However, instead of investing in GPUs, Intel seems to be taking a different approach to AI with its acquisition of San Diego-based Nervana Systems, which was named a ’Cool Vendor’ by Gartner this year for its novel semiconductors in neural networks. Intel did not disclose the price nor terms of the deal.
Nervana Systems touts a full-stack hosted platform for deep learning, Nervana Cloud, which helps businesses develop and deploy high-accuracy deep learning systems. The company is also developing Nervana Engine, an application-specific integrated circuit for deep learning.
’[Nervana Systems'] IP and expertise in accelerating deep learning algorithms will expand Intel’s capabilities in the field of AI,’ said Diane Bryant, executive vice president of Santa Clara, Calif.-based Intel's Data Center Group. ’Nervana’s Engine and silicon expertise will advance Intel’s AI portfolio and enhance the deep learning performance and TCO of our Intel Xeon and Intel Xeon Phi processors.’
While Intel owns FPGAs (field programmable gate arrays) that have specific function accelerators, including inference engines for deep learning, it can now flesh out its artificial learning capabilities with the addition of Nervana Cloud.
"Intel’s strength has always been to take innovative ideas and products and commercialize them," said Stephen Monteros, vice president of business development and strategic initiatives at Sigmanet, an Ontario, Calif.-based Intel partner. "Intel for the most part is driven by a vision to develop large-scale commercially viable products, [and] they must feel this will help them get there."
Intel’s deal for Nervana Systems follows Apple’s reported $200 million acquisition of artificial intelligence Seattle startup Turi Friday, as more vendors continue to invest in machine learning capabilities.
Randy Copeland, president and CEO of system builder Velocity Micro, an Intel partner based in Richmond, Va., said the acquisition will put Intel shoulder-to-shoulder with other vendors investing in AI, including competitor Nvidia, whose GPUs shoulder its deep learning capabilities.
’[The acquisition is] a win-win, and should give Nvidia more competition as this market grows,’ said Copeland. ’Nervana Systems has some very promising but untried solutions for deep learning, and Intel has the capability to finish up the project. Also, Nervana carries IP that Intel can leverage as it moves quickly into the deep learning space, and they don't need to start from scratch with a new GPU-style design.’
While consumers may not see anything ’meaningful’ in the market from the acquisition for at least one to two years, in the long term end users will certainly reap the benefits of the increased competition between the two chip makers in developing artificial intelligence systems, said Copeland.
’For consumers, the net result is that artificial intelligence is getting dramatically better in the next decade, and will be more economical to filter down into our everyday lives,’ he said. ’For the data center and cloud, Intel has acquired new methods that will speed up the whole interface between people and computers.’