Nervana takes $600K to build hardware for deep learning

Big technology companies — Baidu, Facebook, Google, IBM, Microsoft, Netflix, Yahoo — have been working hard to make computer servers act more like human brains, deducing information, recognizing patterns, and gradually becoming smarter over time. Now a startup is aiming to build specialized hardware for such complex computation.

The startup, Nervana Systems, announced $600,000 in seed funding today.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":1451571,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"big-data,business,entrepreneur,","session":"B"}']

That’s not much, but it’s a start. And more venture funding for Nervana in the next few years could be critical for it to come out quickly with a viable product.

Surely longstanding hardware makers like Nvidia also smell an opportunity to monetize this hot area of computing, known as deep learning. Some of the companies engaged in it are known for devising designs for their own servers and buying at great scale. Now the trick will be to stand out and convince at least one of the companies interested in deep learning to try Nervana’s wares.

At least San Diego, Calif.-based Nervana’s startup team is well versed in deep learning and hardware design.

Cofounder and chief executive Naveen Rao previously worked on artificial neural computation inside the neuromorphic research group at
Qualcomm. Cofounder and chief technology officer Amir Khosrowshahi was also at Qualcomm, where he worked on neuromorphic sensors. Another cofounder, Arjun Bansal, specialized in algorithms for deep learning at Qualcomm.

The team also includes two hardware veterans, Carey Kloss and Andrew Yang.

It sounds like Nervana won’t be aiming for the kind of hardware that every small and medium-sized business would buy. Rather, the company is thinking about handling vast quantities of data.

“Deep learning has emerged as the leading technique for finding structure in very large data sets, and Nervana Systems’ specialized hardware will do this at a new level of scale to meet growing demands, Rao said in a statement.

Other vendors aspire to handle hard computing work at scale, too. Nevertheless, some smart investors seem convinced of Nervana’s opportunity.

[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":1451571,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"big-data,business,entrepreneur,","session":"B"}']

The startup’s funding comes from Allen & Co., SV Angel, Adithya Agrawal, Y Combinator president Sam Altman, Eric Baker, Scott Banister, Farzad Khosrowshahi, Expedia chief executive Dara Khosrowshahi, Ali and Hadi Partovi, Geoff Ralston, Dropbox vice president of operations Ruchi Sanghvi, and Owen Van Natta.