Three reasons Uber should never have invested in self-driving technology.

R&D and innovation are not the same thing.

There’s a classic Venn diagram generally attributed to Ideo’s Tim Brown that points to the reality that for an idea to be considered an innovation, it needs to satisfy three criteria: desirability (people would want it), feasibility (it is something that can realistically be created) and viability (it can be made and offered in a way that makes financial sense for the business). Academics or inspired home tinkerers may be satisfied with any combination of one or two of these qualities, but a business, especially a publicly held business, needs to satisfy all three.

Viewing Uber’s sustained autonomous vehicle R&D efforts against this framework, I think we can safely question the reasoning that led to this endeavor. That’s not to say that I disagree with the underlying premise that self-driving vehicles will ultimately take over the role of your cousin Tommy who earns a little extra cash driving for Uber or Lyft, now. I just question whether Uber did the smartest thing for its business or its investors by taking on the Herculean task of developing the technology themselves. 

The liabilities are potentially massive.

The National Transportation Safety Board (NTSB) just released a report outlining the results of their investigation of a fatal accident involving a pedestrian pushing a bicycle and a self-driving Uber vehicle. The NTSB chose not to find Uber criminally liable for the accident. But according to new documents released as part of the investigation, the software inside the Uber car in question was not designed to detect pedestrians outside of a crosswalk, and the self-driving car, in general, failed to consider how humans actually operate.

I have to think this was likely Uber’s one and only free pass. As we have seen in corporate liability cases lately, juries are quite willing and capable of doling out judgments in the multibillion-dollar range. One might argue that even that risk would be worth it for Uber to reap the rewards of working autonomous driving technology long term. But, as we’ll see in point two, one might be overestimating Uber’s ability to reap those rewards in any significantly advantageous fashion.

Long term, there is no competitive advantage.

In 2018, Uber spent nearly $500 million on transportation research. Of course, all of that was not on driverless vehicle research. Some was on autonomous delivery drones. The point here is that massive R&D budgets should be at least theoretically correlated with massive payouts—as in risk and reward should be expected at equal levels. But the potential rewards seem unlikely to be Uber’s alone, if theirs at all. It appears that if Uber does, somehow, develop working autonomous driving technologies, it’s unlikely they will be alone in doing so. And as it stands today, Google is more likely to win the race to market, and perhaps more germane, Uber will most likely have to license much of their self-driving tech stack from Google, anyway.

So, let’s assume Uber eventually figures out self-driving. They already know they will be multiple billions in the hole, day one, on the software front. But now they’ll need a fleet of vehicles. I have doubts Uber is poised in any way to become an auto manufacturer as well, so they’ll need to purchase a fleet built to run their software. Meaning, the costs they’ll need to recoup on capital expenditures will be disproportionately higher than a competitor who simply buys vehicles that will most certainly be available through any number of auto manufacturers who will license Google’s tech directly.

It seems like a scenario that has Uber “winning” this competition would need to be based on the assumption that customers have such goodwill toward Uber that they would be willing to either pay a premium for their services or that Uber would be able to once again operate at a loss for an extended period of time while they roll out the service. Both seem unlikely, especially the latter, which brings me to point three.

Short term, they simply cannot afford it.

Scale is a funny thing. It’s either on your side or it’s not. And in Uber’s case, it is not. At least not yet. In their earnings report for the third quarter of 2019, bookings, or total rider receipts before expenses (like paying drivers), grew to $3.7 billion, a 29% increase over the same period in 2018. Total net loss, however, grew to $1.1 billion—18% more than the same period of 2018. So, losses are positively correlated with revenues. Not good. Since that report, the stock has fallen by nearly 10%. Now, we cannot attribute all of their losses to the investment in self-driving technology, but it is a significant portion of that. It will be seen in the next few quarters whether the market believes those losses are an investment in a more lucrative future or another outsized Silicon Valley wager. Obviously, from my previous point, I fall into the latter, more pessimistic camp. Further, in the past few weeks, we’ve seen states coming after Uber with massive employment tax bills. If the courts side with any of the local and state governments, you can be sure every other state and municipality will get in line to file a suit of their own.

A lesson in sunk costs?

While one would be hard-pressed to find fault in the reasoning that prompted Uber to begin researching autonomous vehicles (it is undoubtedly the future of on-demand transportation), in my humble opinion, it’s time they reconsider their internal calculus on the build-or-buy question. Would Uber potentially lose face by abandoning the project? Potentially. But I think if the company accelerates its transition to profitability, they will be in a more competitive position, with no shortage of “buy” options when the technology matures.