If 2018 was when the industry was shocked into sobriety, 2019 would be a year when pragmatism and the challenges of trying to develop and scale a technology got a lot more real. The upshot: the industry is in the midst of its adolescent years, where change can occur suddenly, causing confusion and awkward encounters. It’s a time when others develop far faster than their peers.
And this coming year promises some of the same.
Cracks in the autonomous vehicle industry — concealed by quixotic zeal and a seemingly bottomless bucket of venture and corporate capital — became too conspicuous to ignore in the opening months of 2018.
A high-profile trade secrets case that pitted Waymo against Uber revealed a cutthroat and reckless side of the burgeoning industry. Just a few weeks later, a self-driving Uber vehicle killed a pedestrian while testing on public roads in Tempe, Arizona, leading Uber to immediately halt all testing. And while other companies only temporarily paused testing of their own, the incident cast a pall over an industry that aspires to make roads safer.
Flashy self-driving car demos slowed to a trickle and some grumbled to TechCrunch about raising capital. Timelines for commercial deployments of robotaxis got fuzzier. Research and advisory firm Gartner released its annual Hype Cycle chart and autonomous vehicles were shown entering into the trough of disillusionment.
And so the chatter and announcements at CES 2019 — the giant tech show held each January in Las Vegas — shouldn’t have surprised anyone. It did anyway. Suddenly, or so it seemed, buzz words weren’t “driverless” and “robotaxis,” they were “safety” and “advanced driver assistance systems,” a less-capable level of automation that is found in new sedans, SUVs and pickup trucks.
This renewed focus on ADAS, along with flurry of partner-swapping, strategic deals and the beginning of consolidation in the industry would be the overarching themes in 2019. This year, with CES 2020 just days away, the ADAS love story will continue along with consolidation and a slow march towards limited robotaxi deployment.
ADAS resets industry expectations
Nvidia CEO Jensen Huang set the stage for one of the themes at CES and the rest of 2019 when he introduced a Level 2+ system called Nvidia Drive AutoPilot. The new product was introduced as a reference platform that automakers and suppliers like Continental and ZF would be able to use to bring more sophisticated automated driving features — not full self-driving — into their production vehicles.
The reveal felt like whiplash to some industry watchers.
Just two years before, Jensen was on the CES stage touting how Nvidia’s tech would lead to Level 4 autonomy by 2020. Level 4 is a designation by the Society of Automobile Engineers (SAE) that means the vehicle can handle all aspects of driving in certain conditions without human intervention. Level 2 systems, in which two primary functions are automated, still have a human driver in the loop at all times.
“The idea of full autonomy went a bit behind the curtain in 2019 and ADAS took more of center stage,” Jeremy Acevedo, senior manager of insights at Edmunds told TechCrunch in a recent interview.
Nvidia was hardly alone. The shift was driven partly by technical and regulatory challenges for self-driving car developers. It was also in response to demand from automakers looking for nearer-term solutions that would improve the technical capabilities of consumer passenger vehicles.
It was a vindication of sorts for Intel, which had spent more than $15 billion back in 2017 to acquire Mobileye, a leading automotive developer and supplier of sensor systems that help prevent collisions. Mobileye, which operates as a subsidiary within Intel, believes that Level 4 or Level 5 (completely driverless in any and all conditions) can’t be reached without going through lower levels of automation first.
“You started to see a lot of Level 2+ kind of talk from players that were looking at platforms that were supposed to be for full autonomy now being repurposed for higher end ADAS,” said Jack Weast, a senior principal engineer at Intel and vice president of autonomous vehicle standards for Mobileye. “I think it was both a reflection of the realization that ‘hey, this is a little harder than we thought, to make something good enough to remove the human,’ and public perception.”
As a result, Weast thinks 2020 will be the year that ADAS-equipped vehicles will become the new normal. Customers will become more familiar with automated driving technology, which will hopefully lessen fear around fully autonomous vehicles, Weast added.
There’s already evidence that this has started. Consumers are opting to pay more for driver assistance and other “content” features in vehicles, according to Edmunds’ Acevedo.
In 2019, consumers spent on average $10,042 more than the base price of a vehicle, a 10.4% increase from the previous year. Look back a few more years and the gap is even wider. Consumers in 2019 spent 36.2% more on “content” add-ons in vehicles like ADAS compared to 2014.
And that’s been a boon for suppliers. Bosch generated 2 billion euros in sales from driver assistance systems in 2019, a 12% increase from the previous year, according to Kay Stepper, vice president at Bosch who heads the company’s driver assistance and automated driving regional business unit.
“That doesn’t even include what we’ve seen a renewed interest in, which is the Level 2 and Level 3 systems that will be rolled out in the next several years,” Stepper said.
Capital and consolidation
Autonomous vehicle companies have continued to press ahead, despite this new attention on ADAS. But it hasn’t been easy.
Developing autonomous vehicles that can shuttle people between two points — without a human safety operator behind the wheel — has proven to be an expensive endeavor.
A handful of AV startups such as Argo AI, Aurora and Cruise captured considerable capital in 2019.
Aurora raised $530 million in a Series B funding round led by Sequoia Capital and “significant investment” from Amazon and T. Rowe Price Associates; and VW Group invested $2.6 billion in capital and assets into Pittsburgh-based and Ford-backed Argo AI. In a separate deal, VW Group and Ford expanded an alliance to include electric vehicles and agreed to collaborate with Argo AI to introduce autonomous vehicle technology in the U.S. and Europe.
And then there was Cruise, which raised another $1.15 billion in new equity from a group of investors that included T. Rowe Price Associates, Honda, SoftBank Vision Fund and its parent company GM. The investment increased Cruise’s post-money valuation to $19 billion, inclusive of SoftBank’s previously announced investment commitment. The round followed two big deals from 2018: a $2.25 billion investment by SoftBank’s vision fund and a $2.75 billion commitment from Honda as part of an exclusive agreement with GM and Cruise to develop and produce a new kind of autonomous vehicle.
But this seemingly rich capital pie wasn’t cut up evenly. While companies like Argo AI, Aurora and Cruise landed big deals, other companies struggled to raise funds. Some facing down rounds sought out, or continue to seek out, buyers or strategic partners.
This inequity gave the better capitalized companies a chance to beef up their technical coffers, particularly around sensor and simulation.
For instance, Alphabet business unit Waymo recently acquired simulation company Latent Logic for an undisclosed sum, and Aurora bought lidar company Blackmore and 7D Labs, a simulation startup founded by former Pixar software engineer Magnus Wrenninge that makes photorealistic synthetic data sets for street scenes. Drive.ai, the buzzy AV startup once valued at $200 million, was acquired by Apple as it prepared to shut down.
Even Tesla got into the acquisition mix and bought DeepScale, a Silicon Valley startup that uses low-wattage processors to power more accurate computer vision, in a bid to improve its Autopilot driver assistance system and deliver on CEO Elon Musk’s vision to turn its electric vehicles into robotaxis.
The dream of driverless
Just a few years ago, if you were to believe the rhetoric from the autonomous vehicle industry’s biggest players, robotaxi services were imminent. Executives spent the latter part of 2018 and the opening months of 2019, reeling back boastful predictions and adjusting deployment timelines.
Kelly Funkhouser, who heads up connected and automated vehicle technology testing at Consumer Reports, said in 2019 the divide between the levels of automation grew wider.
“Some of the larger companies attempting Level 5 publicly admitted that it’s harder than they thought,” Funkhouser said.
Waymo CEO John Krafcik set the sobering tone for the coming year during an on-stage interview in late 2018 at the Wall Street Journal Tech conference.
“Driverless cars are here, they’re truly here and people are using them — at the same time they’re not ubiquitous,” Krafcik said. “It’s going to be a really long time, I think decades before you see this kind of technology everywhere in the world.”
He later added that he’s not certain if the industry will ever achieve Level 5 of automation, a designation that means the vehicle can drive in any environment under any weather conditions.
“Autonomy always will have some constraints,” he said at the time.
It was the kind of talk that would be repeated by executives at other companies in the months that followed. In April, Ford CEO Jim Hackett said during a talk at the Detroit Economic Club that the industry had “overestimated the arrival of autonomous vehicles.”
As 2019 reached its midpoint, Cruise joined the fray and announced it would scrap its ambitious target to launch a commercial robotaxi service before the end of the year in San Francisco.
“They’ve pivoted on when is a realistic deployment date for higher levels of automation,” said Stepper. “You don’t hear the aggressive statements anymore about when this will be in consumers’ hands. 2019 was really a year of more realistic expectations and now the conversation has changed to when can this be truly deployed for series production in larger volume.”
Musk, and by extension Tesla, is perhaps the one outlier to this larger trend. In April, the CEO announced plans to launch the first robotaxis as part of broader vision for an autonomous ride-sharing network in 2020.
“From our standpoint, if you fast forward a year, maybe a year and three months, but next year for sure, we’ll have over a million robotaxis on the road,” Musk said at the time. “The fleet wakes up with an over the air update; that’s all it takes.”
Even as others have punted on timelines, Musk has dug in. The company’s approach has been to roll out over-the-air software updates on its Autopilot software. Over time, Musk contends that owners who have paid for FSD (or full self-driving) will enjoy that level of driving automation their production vehicles.
Signs ahead for 2020
While Acevedo expects 2020 will in many ways mirror 2019 with full autonomy taking a backseat to ADAS, he was quick to note that there had been some notable milestones in the pursuit of fully driverless vehicles, notably with Waymo.
Waymo has operated an early rider program, which connects vetted members of the public with its self-driving minivans, in the Phoenix suburbs since April 2017. Last December it took the next step and launched Waymo One, a commercial self-driving car service and accompanying app in the same area. Members of the early rider program were transferred to Waymo One, which allowed them to bring guests and even talk publicly about their rides. But all of the self-driving vans that carried members of the public still had a human safety driver behind the wheel.
That changed in October of this year when Waymo began to let members of its early rider program hail rides in driverless minivans — no safety driver in sight. This month, the company began to charge for rides.
The number of driverless rides and the area they operate in are still small. But it was enough to keep Waymo in a leading position, in terms of deployment.
Now, others are preparing for driverless deployments of their own — a shift that will spur further investments in sensors and simulation startups.
To be sure, these deployments will be limited in terms of volume and geography; 2020 will not be the year when driverless becomes ubiquitous. In the near term companies must start justifying the huge investments made into their tech. In short: business models matter.
“Only those who have a defined and needy customer will see success with product-market fit,” Voyage CEO and co-founder Oliver Cameron told TechCrunch.
Comment