A next-generation artificial intelligence (AI) chip startup called Luminous has raised $9 million in a seed funding round and attracted investors like Microsoft Co-Founder Bill Gates, Uber CEO Dara Khosrowshahi and Uber Co-Founder Travis Kalanick, according to a report by CNBC.
The investment illustrates that tech players still think there’s a need for growth when it comes to AI hardware, which can be used in everything from cars to smart refrigerators.
There have been a few startups popping up lately working on AI hardware, including one called Nervana that was acquired by Intel in 2016.
Nervana investor Ali Partovi, who heads a VC fund called NEO, has also invested in Luminous.
“I always prefer to bet on a talented young team over a big established company,” Partovi said. “While there’s certainly a ton of risk, that’s what makes it worth investing now: if the race was already over, it would be too late to invest.”
The next big jump in AI chip technology hasn’t yet appeared, and many companies struggle with making big gains with chips that are advanced enough to cleanly, quickly and efficiently run AI models. Luminous Co-Founder and CEO Marcus Gomez said his company wants to build a chip that can replace 3,000 boards that have Google’s latest Tensor Processing Unit AI chips (TPUs).
“We know this works because we already have working silicon,” Gomez said.
While Google doesn’t sell TPUs to other companies, Luminous eventually wants to do so with its own chips. Gomez, who previously worked at dating site Tinder, said he would start training an AI model in the morning and wouldn’t come back until evening to see if it was done learning. It should happen in minutes, he said, not days or hours.
The company uses a technology called silicon photonics, a process that involves using light to transfer data more rapidly and inside small places like servers. The process parallels how fiber-optic cables work and move data.
There are other companies also learning to use the technology, like Lightmatter and Lightelligence. Gomez said he doesn’t know when his company’s chips will be available.
“We’re using exact same technologies that everyone else is using, and so as a consequence, our timelines are on similar paths to other companies,” he said.