Aurora – Shirtless in Mountain View pt. I

A look at the Aurora SPAC in two parts

• The autonomous trucking company is going public at $11bn with mediocre technology and a financial plan that will not see the company generate any profit before 2027.
• This profile reveals a proposition that is on the extreme end of risk meaning that the valuation should be orders of magnitude lower in order to compensate investors for the outsized risks that they are assuming.
• This analysis is based on RFM research and the investor deck released in conjunction with the announcement of the deal which can be accessed here.

Part I – The shaky fundamentals

• Aurora is a self-driving company that intends to license its technology to fleet operators for both trucking and consumer ride-hailing in return for a per-mile royalty.
• This is an excellent business model in the long run but it results in lower revenues in the short term and increases financial risks should self-driving take longer than expected to come to market.
• This is a very real possibility as the technology is having great difficulty in driving as safely as humans and it is not until it is materially better, that it will be allowed into the market at scale.
• In 2017, RFM set a target of 2028 for self-driving technology to be a commercial reality, and to be fair to Aurora, its business plan more or less reflects this target.
• Furthermore, Aurora is absolutely right to go for trucking first.
• This is because trucking is a big market in the USA, is short of drivers, and is a technologically easier problem to solve if one just sticks to the highways.
• Hence, trucks on the highway are very likely to drive more safely than humans sooner than passenger vehicles in urban and metro areas.
• On this basis, there is nothing wrong with Aurora, but the problems arise when one looks at its technological performance compared to its peers and the risks that the company is asking its investors to assume.
  • First, technology: For the last 4 years, RFM has used California disengagement data as an indication of how good competing offerings are compared to one another.
  • There are large flaws in this rationale (as it is not very difficult to game one’s data) but to date, it has offered a reasonable representation of reality, which is why RFM has continued to use it in its analysis.
  • In both 2019 and 2020, Aurora has performed in the middle of the pack ranking 15^th out of 36^th in 2019 and 12^th out of 29^th in 2020.
  • In 2019 it made mistakes 189x more frequently than the leader (Baidu) which improved to 91x more frequently in 2020.
  • This is not good performance and reflects badly on the company overall.
  • Furthermore, Aurora recently acquired Uber’s autonomous driving effort which was demonstrably the worst in the industry casting further doubt in my mind with regards to its technological prowess.
  • A central part of the investment case is that Aurora is “architecting best-in-class technology” (investor deck page 11) which will allow it to get to market early and compete effectively when the others arrive.
  • However, the evidence I have seen indicates that the technology is an also-ran, meaning that it will be late to market and will struggle to make high margins.
  • Second, risk: Aurora takes risk to a whole new dimension with no revenues until 2023, $622m in 2026, and $2.0bn in 2027.
  • One thing that has become very clear in the last 18 months is that uncertainty is on the rise, and I think that the reality is that Aurora has no more certainty on its 2027 forecast than if it drew the number out of a hat.
  • This is especially the case for Aurora because unlike Lucid Motors, Ouster, Innoviz, or even Faraday Future, its technology is not even close to being market-ready.
  • This means that in addition to the huge execution risk that all of these companies bear, Aurora also bears substantial technology risk on top of that.
  • The data points I have seen (see above) point to this technology risk being much greater than the investor presentations would have us believe.
• The net result is that even before looking at the $11bn valuation being asked (see part II tomorrow), there are some serious issues with this transaction that lead me to question whether I would want to be involved.
• Stay tuned for part II tomorrow when the deal and the valuation will be examined.