Amid economic uncertainty, tight capital markets, wary investors and consolidation in the nascent autonomous vehicle industry, Aurora Innovation is still motoring on towards its target of commercializing self-driving trucks by 2024, ticking off waypoints and milestones along the way.
It’s even managed to rack up a few wins in the past year, including launching or expanding pilot programs with FedEx, Ryder, Schneider and Uber Freight. More recently, Aurora announced its self-driving system is “feature complete” — industry jargon that means the technology has all the capabilities required for hauling freight autonomously on public roads. And it deployed the first of its “commercial ready” autonomous shipping terminals in Palmer, Texas, infrastructure needed if the company’s trucks are going to be on the road 24 hours of every day of the year.
Three years ago, it wasn’t so clear if Aurora would survive its ambitious and risky moves.
In late 2020, Uber sold its self-driving unit Uber ATG to Aurora in a complex deal that would bolster the startup’s valuation to $10 billion along with its autonomy efforts, while giving Uber a 26% stake in the company.
The deal liberated Uber from what had been a troubling development path, including a 2017 lawsuit from Waymo for trade secret theft and patent infringement (which Uber settled in 2018), as well as the 2018 fatal crash involving an Uber ATG autonomous test vehicle. But while Aurora gained talented engineers it also faced the tricky process of integrating about 1,200 former Uber ATG employees into its operations.
Six months later, Aurora would make another leap: this time into the public markets through a merger with a special purpose acquisition company, an increasingly fraught path that has stymied many other automotive startups. Aurora’s stock has suffered along with other mobility SPACs; its stock price debuted November 4, 2021 at $11.25 a share. Aurora stock has fallen 85% since then and was trading Monday at $1.46 a share.
“I think that there is a kind of best house in a bad neighborhood kind of sentiment. That there’s a lot of stuff that’s maybe not going as well in this space, right? Some very public failures and backstabbing,” co-founder and CEO Chris Urmson said in a recent interview. “But, you know, I think that’s short term. A short term issue because as we continue to demonstrate progress and execution and the fact that we made strategic bets years ago that are paying off, right? That we expect that that will be recognized and, you know, things will move forward.”
Aurora is now encountering another batch of challenges as it drives towards commercialization, from assuaging shareholders and its growing list of partners to navigating regulatory hurdles in California and raising more money.
And despite that “feature complete” achievement, Aurora still has plenty of testing to do, including ramping up to haul 100 loads per week between Dallas and Houston by the end of this year.
The company has more than 30 trucks on the road in Texas today hauling goods with more than 50 trips a week. To date, the company’s trucks have traveled more than 400,000 miles and hauled 20 million pounds of freight for FedEx, Uber Freight, Werner, and Schneider.
If all goes according to plan, Aurora will deploy its self-driving system called Driver commercially in Texas, without onboard human supervision, by the end of 2024.
According to Urmson, his company’s autonomy tech will fix what he calls a “systemic problem” in the U.S. supply chain: “There’s not enough people willing to drive trucks. We’re short about 80,000. We expect to be short about 150,000 by the end of the decade.”
The trucking industry also suffers from a whopping 90% turnover rate, and its workers must abide by federal 11 hour daily limits behind the wheel.
Urmson says that the Aurora driver should be able to keep a truck moving for about 20 hours on an average day and that, plus associated fuel savings and reduced insurance costs due to safer driving, means theoretically doubling the revenue per truck for a fleet operator.
“Getting from Houston to LA, for example, takes about three days by truck today, because of that 11 hours of service limitation. The Aurora Driver should be able to make that trip in 24 hours,” Urmson said.
Regulatory speed bumps
Those increased efficiencies and subsequent reduced costs are crucial to the appeal of Aurora’s technology to the trucking industry. But, there’s one ugly fly threateningly buzzing around the soup: California’s AB-316. If passed, this bill would require human operators in trucks within the state of California, and could kick off similar legislation elsewhere.
While Urmson hopes that California will “see the economic and safety benefits” of autonomous trucking and allow driverless operation, Aurora is prepared for an eventuality where it becomes law: “It’ll mean helping our customers haul freight between Arizona and the East Coast,” Urmson said.
In other words: no business in California. If policies like this gain momentum, that trend could pose a threat to Aurora’s business plan.
“It would certainly be if we ended up with kind of a checkerboard across the United States,” Urmson said. “But, I think in practice that’s unlikely.”
The Uber ATG boost
For now, Aurora’s primary goals are validating the functionality of its Driver within the state of Texas and doing everything it can to cut costs. Urmson told TechCrunch that the ATG acquisition has helped on both fronts.
With ATG, Aurora went from 600 to 1,800 employees “overnight,” and Urmson says they’re still operating with about 1,700. While some Uber ATG employees chose to leave after the deal was closed, Urmson said none were made redundant: “The ones who want to be here are, and that’s really all we can hope for, right?”
At this point, the teams are well integrated, according to Urmson, but it was something of a painful process.
“We were very deferential around the culture on both sides. And I think that just led to confusion,” Urmson said, with conflicting lines of reporting and duplicated efforts. “Like any organization, there’s pockets of challenges, but I think that, again, those folks that were united in the mission of let’s go get the vehicles on the road, let’s do something that’s going to change the world, they fit in really well.”
The Uber ATG deal also helped Aurora become more efficient and cut costs with the goal of extending the company’s runway in the lead-up to commercialization. For instance, Urmson said that ATG’s batch processing of cloud computing was far more efficient than how Aurora was previously handling distributed tasks. “As you can imagine, for all of our ML [machine learning] stuff, all of our testing and simulation, there’s a massive amount of cloud orchestration that happens,” Urmson said.
Using ATG’s batching process, Aurora’s distributed systems team was able to take disparate cloud requests and lump them together into fewer, larger requests, saving time and reducing costs. “That’s been really powerful and it’s the heart of what we use today,” Urmson said, estimating the cost savings there alone to be in the “tens of millions of dollars.”
On the functional side, ATG’s offline Simultaneous Localization and Mapping (SLAM) procedure has Aurora generating high-res maps more efficiently. ATG’s sampling-based motion planning has been integrated to help the Aurora Driver in emergency, near-collision situations. The Aurora Driver’s behavior in construction zones also derives from ATG algorithms.
All of this is included in the Aurora Driver Beta 6.0 release, which “introduced the final driving capabilities needed to commercially haul freight without vehicle operators” according to Aurora. That commercial launch, Urmson says, is still on target for next year, and those cost-cutting measures have ensured Aurora will have enough cash to get the company through the middle of 2024.
More cash, more runway
Aurora is now focused on converting all of those pilot programs into long-term customers. The company will host its first customer summit at the end of April in Dallas — an event that will bring together its pilot customers and partners to discuss the next steps towards commercial launch, according to the company.
However, another raise is still in the plans. On April 6, Aurora filed for a proposed mixed shelf offering to raise $350 million.
“We have not been shy about the fact we would need to raise additional capital in the future,” an Aurora spokesperson told TechCrunch when asked about the filing. “This is a standard filing that provides Aurora the flexibility to raise at a future date but is not a signal we intend to imminently fundraise.”
In the meantime, Urmson said he is focused on de-risking the business.
“We’re going to wait for the right time and the right partners. What we see is, as a company, we look at these milestones, and our execution against them, and we see them as major moments of de-risking our business,” Urmson said.
Reducing risk and inspiring confidence now is key, particularly in the wake of widespread consolidation throughout the industry that has left some investors wary.
The abrupt termination of Argo AI last year has left many questioning the future of autonomy as a business. Meanwhile, Alphabet’s self-driving unit Waymo reportedly ended its own trucking effort, Waymo Via, in January. (Urmson previously led the former Google self-driving project.)
“I think one of the really important early decisions we made as a company was to be independent,” Urmson said, referencing Argo AI’s funding by Ford and Volkswagen. “As a company, we have a mission to deliver the benefits of self-driving technology safely, quickly and broadly,” he said. “We work with amazing partners, but for none of them is that their mission.”
For Urmson, that misalignment between Argo’s quest for autonomy and its corporate parents’ need for profitability that ultimately sealed Argo’s fate during tough economic times.
Until things improve, Urmson continues to drive home the importance of frugality on his team: “At each of our all-hands, we have a section we call ‘Every Little Bit Matters’ and we highlight places where teams across the company have saved between thousands and millions of dollars by either not doing something or finding a less expensive way to do it. I think that is a muscle that will serve us well, even when economic times become better.”
How Aurora is navigating the bumpy road to commercial self-driving trucks by Tim Stevens originally published on TechCrunch