At the EcoMotion self-driving conference held (in cyberspace) from Israel this week, Amnon Shashua, founder and CEO of MobilEye, now a unit of Intel INTC , declared their intention to offer robotaxi service, with no safety drivers, in early 2022. They will begin in their headquarters town of Jerusalem, then move to Tel Aviv, then France, Korea and China.
He makes this statement while many other companies, particularly car OEMs, are scaling back their plans and timelines on full robocar service. We’re recently seen layoffs at several companies, including some of the biggest start-ups like Zoox and Cruise. Zoox also confirmed last week it is shopping around for a potential buyer.
This is not the first time this date has been stated by MobilEye, but they are not following the negative swing for now. MobilEye has been a major leader in the driver-assist space, and Shashua is probably the most astute leader of an ADAS company, and the most equipped to make the jump from ADAS to real self-driving. They enjoy showing demos of driving in the difficult and chaotic streets of Jerusalem, and it’s a challenging place to start.
Shashua also predicted there will be significant consolidation among players because of the downturn which has impaired venture funding, stock markets and OEM plans. Even before the virus, robocars were heading into the Gartner IT “trough of disillusionment” which all revolutionary technologies face, and the global downturn has certainly not helped.
After being bought by Intel for $16B, MobilEye has the funding to weather major storms like this. Their driver assist business has taken a hit as car OEMs turned off their production lines, but this will resume, though the lack of driving and economic woes will slow car sales.
MobilEye began doing ADAS with computer vision and radar. The original Tesla Autopilot was their product. For robotaxis though, MobilEye takes a different position from most companies. Tesla wants to continue with the vision/radar approach while most big players have a lidar+vision approach using top quality lidars. MobilEye advances a “vision+lidar” approach which puts vision first, but wants to use lower-end lidars to compete the sensor suite. Depending on the progress in computer vision, this could be a path to lower costs. Most non-automaker companies, however, are strongly focused on getting a working safe system at almost any cost, planning to reduce the cost later, rather than aiming for low cost at first.
While car OEMs still put a lot of focus on making self-driving cars for private ownership, MobilEye, like Waymo, Cruise, Zoox and other non-OEMs believes the fleet-owned robotaxi model is the right path.
Shashua also outlined their acquisition strategy, and made note of the fact that they build their own maps, taking advantage of the many millions of cars driving around with MobilEye chips in them today. While several companies and startups are hoping to become map suppliers to all the self-driving companies, he believes that maps and other core components must be owned in-house. He’s not the first to come to that conclusion: Google GOOGL , which is really the world’s top mapping company, used to import maps from many sources to make Google Maps. In the mid-2000s they undertook the expensive course of building their own maps instead, so they could control the quality. This was just for phone apps, not trusting your life, but the same philosophy applies. “Here” (formerly Navteq) is now owned by several German automakers. TomTom has a long reputation as a “Tier One” supplier to the automakers but faces a harder battle, as do the other mapping startups.
While generally, there is no path from ADAS to robocars (at least in the view of many) MobilEye is taking advantage of all they have learned in ADAS to make their vehicles, and they are a serious contender. It’s generally unwise to name dates with revolutionary technology, and as such, Shashua probably has reasons he’s decided to name those dates. The world will be watching.