“The advantage that silicon photonics can offer is a small form factor solution that can ultimately result in a compact in-car size of the device,” said Kiyoul Yang, a Stanford University postdoctoral fellow who focuses on photonic hardware. Many companies today use a lidar system based on rotating mirrors, says Yang, which requires the production of discrete, expensive components. “If everything can be integrated into a chip in a small form factor, then everything can be produced at a low cost,” he says.
Again, Mobileye isn’t the only company banking on FMCW, or lidar chips in a broader sense. But it has a definite advantage: Intel already has a silicon photonics manufacturing facility in New Mexico. “Building an FMCW lidar requires know-how, but even if you don’t have the special factories to make the lidar on a chip, it will be too expensive. It’s getting impractical, ”says Shashua. He expects the cost of each lidar SoC to be in the hundreds of dollars each, orders of magnitude cheaper than what systems cost today.
Even if Mobileye’s production roadmap remains stable, an uncertain regulatory outlook could delay the timeline. Still, it is also making short-term progress, announcing today at CES that it would expand its autonomous vehicle testing to Detroit, Paris, Tokyo and Shanghai in 2020. (The locations are strategic; each is located near an automobile manufacturer that Mobileye provides self-driving technologies for.) And it has used the millions of cars with Mobileye on board to crowdsource a map of nearly 1 billion miles of roads in the world so far, processing 8 million kilometers every day. Despite all the attention Tesla is getting, Mobileye is by far the market leader in autonomous driving.
That reputation and Intel’s deep pockets will help it out against smaller competitors in the lidar SoC race. “I strongly believe that reliability in the automotive industry is a great differentiator,” said Mike Ramsey, automotive analyst at Gartner. “Can I trust this supplier to deliver on time and in quality? And Intel has the very important feature of having a very big throat to choke on if something goes wrong. Don’t underestimate its value. “
Mobileye makes up a small percentage of Intel’s total revenue. But along with the client computing group – that is, the chips used in PCs and adjacent products – this is the only segment that grew in the company’s most recent quarter. It’s just the kind of new terrain that Intel must aggressively plot to avoid another smartphone-style miss.
“If you look at the long term, a company like Intel must look for new growth areas. It’s not easy to find one. You want to find a new market worth hundreds of billions of dollars, ”says Shashua, and a market that leverages Intel’s strengths. “Those domains are rare. We are in that domain. “
XPU marks the spot
Mobileye’s lidar SoC is the sharpest example of what Intel calls its “XPU” strategy – that is, looking beyond the CPU to computers in all its many forms. The company launched its first discrete graphics card last fall, has a dominant position in data center processors, and in 2019 acquired AI chipmaker Habana Labs, which won business from Amazon Web Services a few weeks ago to use its accelerators to train deep learning models. .
“At heart, we are a computer company,” said Gregory Bryant, who heads Intel’s client computer group. “We see this world where more and more things need computer use, more and more things look like a computer, not just the server or the PC, but the car, the house, the factory, the hospital. All of those things need computing and intelligence. “
That broadening comes at a time when Intel is facing more challenges than ever for its traditional businesses. Manufacturing delays have stuck to a 10 nanometer process for manufacturing its chips, while competitors have moved to smaller molds. The company’s chief engineering officer, Murthy Renduchintala, left last summer. And the hedge fund Third Point issued a burning public letter in late December calling on Intel to “retain a reputable investment advisor to evaluate strategic alternatives, including whether Intel should remain an integrated device manufacturer and the possible divestment of certain failed acquisitions.”