A Leader in Autonomous Tech Has Yet to Emerge

Whether it’s a concept car, a pilot, or a press release, almost every US automaker has planted a flag in autonomy.

Meanwhile, driver assistance systems get more and more robust. Who is the leader? Based on announced roll-outs of autonomous vehicles, Waymo is first, followed by General Motors but that’s based on company statements. “It’s incredibly hard to know who is an actual leader in terms of technological capabilities, because there is no good way to measure the capability of these systems,” says Mike Ramsey, research director, automotive and smart mobility, for Gartner.

According to Ramsey, such leader-boards are based on circumstantial evidence: how many vehicles deployed in tests, how many miles driven and reports to the State of California on how many times autonomous systems had to disengage. Even this latter data is flawed as a way of comparing, he says, because not every automaker participates, conditions for disengagement could be very different and there’s not even a standard on what counts as a disengagement.

Another way of evaluating leaders is looking at which companies have been working on autonomy the longest and which have the most cars on the road, Ramsey says. With that lens, he’d identify nuTonomy, now owned by Delphi as the leader. “This was the first company globally to start offering autonomous rides to consumers. I consider them to be one of leaders but they don’t get talked about a lot,” Ramsey says.

Partnerships emerge as key

While some automakers and Tier 1s continue to build proprietary systems bulked up by vendor contributions or start-up acquisitions, other players are forming partnerships to bring autonomous driving to market. “Partnership was not the strategy a couple of years ago,” Ramsey says. “The strategy was, we must own this but the cost to develop this is becoming increasingly huge.”

For example, Waymo’s self-driving car will not be a cute bubble. Instead, Waymo has partnered with Chrysler and Jaguar Land Rover, relying on automakers for their core competencies. While Waymo’s current offerings with Chrysler have the bulky sensors and LiDAR mounted on top, Jaguar and Waymo will build the I-Pace as a fully autonomous vehicle.

This partnership exploits Waymo’s strengths, says Angelos Lakrintis, industry analyst for Strategy Analytics: “Waymo has a whole software stack in their availability with layers spanning from machine learning, specialized and geo-fenced mapping approaches, simulations and in-house software technology which is patented and used in their cars.” He notes that the US Patent & Trademark Office has rejected most of Waymo’s LiDAR patents but, still, it’s a formidable array of software.

Even General Motors, considered a leader in the race to autonomy, has partnered with Honda to develop a purpose-built autonomous EV using Cruise technology. Company statements pointed out the synergies: GM provides its design, engineering and manufacturing expertise. Honda contributes expertise in space efficiency and design and its global reach. According to Ramsey, the GM/Honda partnership shows just how difficult and expensive autonomy initiatives are. “Honda says it’s worth $2.7Bn to us not to do this on our own. That’s a pretty big number.”

Crucial technologies still in play

Lakrintis notes that LiDAR is one arena of competition among developers of autonomous systems. The Valeo ScaLa mechanically-scanning sensor is installed on the 2018 Audi A8 four-door, while Fisker’s EMotion electric coupe uses the Quanergy S3 optical phased array sensor.

BMW is also partnering with Innoviz to develop a MEMS-scanning LiDAR to launch in 2021, while Ford has also invested $75M in Velodyne to develop a future sensor. He adds: “Since major deployments are due later in the mid-2020s timeframe, with the possibility that full-autonomous ride-sharing robo-taxis will start to arrive, there is still some time before newcomers with more robust, cheaper and/or more reliable technologies can win over potential customers.”

Ramsey argues: “The entire way that sensing is done is still not settled upon.” There’s still debate about how many and what kind of sensors a self-driving car needs, with new vendors popping up insisting that autonomous systems need infrared cameras or ground-penetrating radar, for example.

Is autonomy a dead end for automakers?

Most automakers continue to pursue autonomy, either with internal R&D, in collaboration with their vendors, or even in joint ventures with other automakers. However, what happens if one self-driving system turns out to be better than the others? Won’t every automaker have to use it?

Says Ramsey: “The cost is huge and the question of whether there is any sort of differentiation in the end game is not clear.” Ramsey thinks that, ultimately, autonomy will not deliver any kind of brand value. It will simply be an expected function of the car.

That doesn’t mean autonomy initiatives are a waste of money. For one thing, they can affect the stock price. “If you wind the clock back two years when the real hype around autonomous vehicles was at its height and you are a car company that has zero investment in autonomous vehicles, it looks like you have abdicated to everyone else,” Ramsey points out.

This can mean that autonomous R&D is money well spent because such programs attract talent and provide prestige, he says, adding: “The froth generated by every car company pursuing its own thing is driven by keeping up with the Joneses.”