A stealthy startup known as Cerebras raised round $25 million to construct deep studying


A stealthy startup known as Cerebras raised round $25 million to construct deep studying

Benchmark is doing one other massive funding earlier than the tip of the 12 months, and this time it’s in a  startup known as Cerebras Programs. From what we hear from sources, the spherical (which we heard others had been collaborating in) is someplace round $25 million, although it might be as excessive as $30 million.

Particulars on the corporate are extraordinarily sparse — we will’t even actually discover that a lot details about it on the Web for starters. What we’ve heard from sources is that the corporate is engaged on a specialised next-generation chip (presumably a GPU) for deep-learning purposes. Right here’s the corporate’s description from LinkedIn (actually the one factor that may be discovered from a Google search):

READ NOW !  Makeblock raises $30 million for robot-building kits for teenagers

On LinkedIn, Andrew Feldman — the co-founder and CEO of SeaMicro after which at AMD following the corporate’s acquisition — is listed as the corporate’s CEO. Its CTO, Gary Lauterbach, can be a co-founder of SeaMicro after which at AMD, based on LinkedIn.

The calls for for these sorts of processing is barely going to proceed to extend over time as an increasing number of merchandise shift to a extra AI/machine-learning pushed strategy. As firms like Fb and Google acquire an increasing number of information and wish to use machine studying at an enormous scale, they’re going to want to determine new methods to evolve that or construct higher algorithms.

READ NOW !  Toyota and NTT to collaborate on related automotive tech, together with AI

Amazon just lately unveiled a brand new instrument that enables builders to faucet into as a lot GPU energy as they want — signaling an clearly sturdy demand for processors that may use the sorts of capabilities of a GPU. So startup might entice funding if it’s conceived a brand new approach to construct GPUs or other forms of chips that may enhance that computing energy or effectivity, it might result in extra highly effective purposes.

Nonetheless, we’re attempting to dig up as a lot info as we will on the startup, and we’ll replace this story and proceed to observe it as we study something new. Benchmark declined to remark. Feldman, reached through LinkedIn, stated the corporate was nonetheless in stealth mode and never speaking to the press concerning the firm.