Earlier this week, Aurora announced its acquisition of Uber’s Advanced Technology Group (ATG) in a complex deal that is an all-stock transaction. This makes Aurora the largest AV developer—at least in number of employees.
Questions we should be asking are: what’s in it for each party, what would be the net gain for the combined company, and what impact it might have on the overall AV market.
Let me break it down.
In the following chart, I summarized the key parameters for Aurora, Uber ATG — before and after the acquisition. I also outlined the VC investments and key AV suppliers when available. The blue blocks summarize Aurora before the deal. Black boxes show Uber ATG and red boxes are the combined company after the acquisition. The green boxes show Aurora’s new potential customers from the ATG connections.
Aurora Before Acquisition
Aurora Innovation was founded in 2017 and has received two VC funding rounds of $90 million in February 2016 and $600 million in June 2019. Amazon and Hyundai’s VC groups were among the 12 investors. Aurora’s latest valuation was in the $2.5 billion range.
Aurora’s has grown rapidly and had nearly 700 employees in four cities with Bay Area and Pittsburgh as the main centers. Aurora has its lidar group in Bozeman, Montana from its acquisition of Blackmore in May 2019. Aurora just located an autonomous truck group in Dallas. Blackmore’s frequency modulated continuous wave (FMCW) lidar is well suited for the long-range sensing needed for a truck at highway speeds.
The name of its AV software platform is Aurora Driver, which is focused on three AV use-cases: AV goods delivery, autonomous trucks and robotaxis. Initially robotaxi was the leading Aurora AV use-case, but this has shifted towards goods AVs and autonomous trucking. The need and growth in truck services from the pandemic is now putting autonomous trucking as the first priority for Aurora.
At CES 2018, Aurora announced it would be using Nvidia Drive Xavier processor. This indicates that Aurora’s current and future AV hardware platform will use Nvidia-based GPU platforms. Similarly, Aurora’s L4 ECUs will be based on Nvidia processors from Tier 1 suppliers. Aurora has used various Ford cars for L4 testing vehicles. Aurora has started using Chrysler Pacifica SUVs in preparation for future goods AV customers.
Aurora has been very deliberate in planning and executing its AV strategy. It is focused on AV safety and AV software simulation before it does road testing. Virtual miles are a major focus and Aurora has developed a Virtual Testing Suite, which allows Aurora to run millions of virtual miles per day. Aurora estimate that just a single virtual mile can be just as insightful as 1,000 miles collected on the open road.
Its founders have taken the opportunity to start with a clean sheet based on what it learned and experienced—good and bad—from Waymo, Tesla, Uber and others. This strategy seems to be paying off in terms of VC interest and industry reputation. And the ATG acquisition is a validation of their strategy.
Uber started its AV research in early 2015 with headquarters in Pittsburgh and its core group of people coming Carnegie Mellon University (CMU). ATG launched its first AV trials in in the fall of 2016 in Pittsburgh. A second pilot followed in Phoenix in early 2017. In January 2018, Uber said it had completed 2 million+ AV test miles with safety drivers. ATG also did intermittent AV testing in California.
Uber ATG lost a lot of momentum in March 2018 when one of its AV test vehicles killed a pedestrian in Phoenix. This event also took a lot of luster from the whole AV industry. This Uber crash was extensively investigated and showed that ATG’s safety strategy and operation was below par. Uber ATG has worked diligently on mitigating and correcting its weaknesses. However, Uber ATG is not perceived as being a leader in the robotaxi software platform sweepstakes.
Uber also entered the autonomous truck segment with the acquisition of Otto in August 2016. Google sued Uber for IP theft in February 2017 from the Otto acquisition. The court activities showed that Google had a valid case for lawsuit. Uber settled the Google lawsuit in February 2018 and in August 2018, Uber closed the Otto project to focus on robotaxi AV use-case. It is not clear if any of the Uber ATG employees have autonomous trucking experience from Otto. If there is still autonomous truck experience at ATG, it would be useful to Aurora.
Uber made ATG a subsidiary in April 2019 when it received a VC investment of $1 billion from Toyota ($400 million), Denso ($267 million) and SoftBank ($333 million). This gave the three companies a 13.8% ownership. This means Uber ATG was valued at $7.25 billion at that time. In retrospect it looks like Toyota-Denso-SoftBank should have received a higher ATG ownership share.
Uber signed up with Nvidia at CES 2018 to use Nvidia Drive platform for its AV hardware. Uber began using Nvidia GPU hardware platform in its first test fleet of Volvo XC90 SUVs, which has continued throughout its AV testing. Volvo used Veoneer as its Tier 1 supplier for its Nvidia-based AV hardware. ATG also used Nvidia AV platforms in its Otto autonomous truck testing while it lasted.
Uber ATG has been a turbulent ride for most of its existence. It still has impressive resources with around 1,200 employees and VC investment of $1 billion. It looks like ATG has found a new home with the expertise to leverage these resources and be part of much better results than the previous five years.
Aurora with ATG
Aurora will be a much larger company with ATG and will have much larger resources—people, capital, investors, testing facilities and expertise. It will have nearly 1,900 employees and even if some people move on, it is well over twice bigger than Aurora’s current personnel.
Total VC funding will reach $2,090 million including the extra $400 million Uber is investing. Uber is also handing over its equity in ATG, which will give it a 26% stake in the combined company. Uber ATG investors including ATG employees with equity will own another 14%. The valuation of Aurora with ATG is $10 billion or about 4X of Aurora’s valuation in June 2019.
The Aurora Driver software platform will have more opportunities and will continue to focus on its three main use-cases: autonomous trucks, goods AVs and robotaxis. It looks like autonomous trucks and goods AVs have higher urgency than robotaxis for Aurora.
Information on Aurora AV suppliers are limited except for its use of the Nvidia Drive platform that it will use in current and future systems. It looks like Aurora is designing its own L4 ECU hardware. This is probably to get a better performance with its Driver software platform. Aurora will probably use a Tier 1 supplier to build L4 ECU that it has designed. Aurora is likely to get multiple suppliers/customers for autonomous trucks, goods AVs and robotaxis.
What Is Next for Aurora?
I believe this is a win-win acquisition for both Aurora and Uber. ATG was a drain on Uber’s finances as it was an expensive R&D effort with no or minimal near-term revenue potential. Uber is likely to get a better robotaxi software platform from Aurora than from ATG in a few years. Uber will also get a goods AV software platform from Aurora for its Uber Eats business.
Since Toyota is now an Aurora investor, it provides some interesting potential. Toyota has its own extensive AV investments and R&D efforts for multiple AV use-cases. Hence there is potential that Toyota could become an Aurora Driver customer for three use-cases—goods AVs, robotaxis and fixed route AVs.
SoftBank is the largest investor in ride-hailing companies and is likely to have significant impact on which robotaxi software platforms they will use. That is the reason SoftBank is listed as a potential Aurora “Customer” in the above figure. However, SoftBank is also a leading investor in Cruise.
The biggest issue Aurora will have with the ATG acquisition is the integration of a much larger company than itself. There is some good news, but also potentially tough issues. The good news is that both companies use the Nvidia hardware platforms and they have major facilities in Pittsburgh. Many of the employees also know each other due to CMU ties and have worked together in previous jobs.
The toughest issue is probably the different corporate cultures. Aurora has prominently featured its so-called “no jerks” policy for hiring people. Uber and ATG have had significant cultural issues, including a willingness to cut corners for early project completion, which will not work at Aurora.
It is likely that Aurora’s success and positive culture will work for the vast majority of ATG employees, but it is likely that some (100 to 200?) ATG employees will not be with Aurora in a few months.
Aurora’s software development tools, procedure ad strategy will probably be good news for the ATG employees, and the merged company will do well.