Who Is Really #1 In Self-Driving Cars?

Navigant’s controversial report doesn’t use enough criteria.

byAlex Roy| UPDATED Jul 16, 2019 1:24 AM
Who Is Really #1 In Self-Driving Cars?

Who is really #1 in self-driving cars? You wouldn’t know it from this week’s unintentionally entertaining Navigant Research Leaderboard Report on Automated Driving, which placed Ford first, GM second and Renault-Nissan third. Waymo? Seventh. Tesla? Twelfth. The media—most of whom appear not to have paid $3,800 to read the raw report—lapped it up. Wired’s summary ran with the mother-of-all-clickbait heds, “Detroit Is Stomping Silicon Valley In The Self-Driving Car Race.” 

The Navigant report is well researched—it’s Navigant, after all—but it has one major flaw: It doesn’t really make sense.

No less than Elon Musk biographer

Ashlee Vance

launched a Twitter war against Navigant's Senior Analyst Sam Abuelsamid, suggesting the report was skewed by the company's client list, which includes Ford and other companies that ranked higher than conventional wisdom would suggest. I disagree with Vance. Navigant has a long history of transparency and authoritative research. The authors’ credibility—especially that of the widely respected Abuelsamid—is unimpeachable.

The problem isn’t with Navigant’s research, it’s with the report’s scope and methodology.

The overall thesis—that self-driving technology is nothing without the might of a traditional manufacturer behind it—is as myopic as Silicon Valley’s belief that technology investments alone can “disrupt” the car industry.

This type of disruption mythology makes me sick. Disruption isn’t magic. Disruption isn’t the art of executing an idea competitors can’t or won’t. Disruption is the science of executing an idea better than competitors can or will. Disruption mythology harms both sides of an industry under attack, because it masks the nature of realities everyone must face if they want to survive and prosper.

Navigant’s report is a perfect example of counter-disruption mythology, a document that satisfies a calcified industry who want to believe buying is as good as building, money can solve for time, and being a Foxconn in a new transportation paradigm is for losers. 

What Does The Navigant Report Really Say?

The report’s actual title is “Assessment of Strategy & Execution for 18 Companies Developing Automated Driving Systems.” The report is very impressive, but you don’t need to read past the abstract to recognize its inherent flaw. Here are the ten criteria by which Navigant ranked the players:

  1. Vision
  2. Go-To-Market Strategy
  3. Partners
  4. Production Strategy
  5. Technology
  6. Sales, Marketing & Distribution
  7. Product Capability
  8. Product Quality & Reliability
  9. Product Portfolio
  10. Staying Power

That sounds like a pretty comprehensive set of criteria, but by giving them all equal weighting, Navigant presupposes the car industry will stay pretty much as it is, that self-driving cars will be built and marketed through existing channels, and that new channels/platforms will be controlled or owned by existing manufacturers.

By making technology only 1/10th of any given player’s score, numerous car manufacturers outrank even the most advanced Silicon Valley players, and would do so even if the formers’ R&D budgets dropped to zero.

Two missing criteria? Government relations and data.

Self-driving cars won’t become ubiquitous overnight. Different types of vehicles will propagate—on different software platforms, in different regions—based on local laws, all of which depend on wildly expensive, long-term lobbying efforts. Whichever platform/car conforms to local laws first—especially if they helped define those laws during R&D—wins.

Once hardware is commoditized, software is everything. Actually, it isn’t everything. Data is, because data is the fuel that feeds self-driving technology, or, as Intel CEO Brian Krzanich stated, "data is the new oil." The greatest AI in the world needs data—mountains and mountains of data it can turn into information to act upon.

Navigant Research

Here’s Navigant’s list by rank:

  1. Ford
  2. GM
  3. Renault-Nissan Alliance
  4. Daimler
  5. VW Group
  6. BMW
  7. Waymo
  8. Volvo/Autoliv/Zenuity
  9. Delphi
  10. Hyundai Motor Group
  11. PSA
  12. Tesla
  13. ZF
  14. Toyota
  15. Honda
  16. Uber
  17. nuTonomy
  18. Baidu

If technology and data gathering were the only criteria, the list should look like this:

  1. Waymo
  2. Tesla
  3. The other Americans
  4. The Germans
  5. Volvo
  6. Hyundai
  7. The Japanese
  8. FCA

Of course, it's difficult to find anyone who disputes Waymo’s advantage. One could debate swapping Waymo and Tesla’s rankings, but I’ll leave that to the experts on Twitter. Waymo and Tesla have gathered orders of magnitude more data than anyone else, but the precise nature of the data, the rate at which they’re gathering it, and the value of that data is difficult to ascertain.

What about George Hotz's Comma.ai, with nearly 1M miles of data, including video? Hard to say what that’s worth without knowing how much and what kind of data automakers have gathered. Maybe there’s a reason they’ve been so quiet on the data front.

What about government relations? Fairly opaque, but stealth startup Zoox’s example highlights the folly of trusting Navigant’s methodology. Former NHTSA Administrator Mark Rosekind just joined Zoox’s management team. I’d call that a pretty big advantage. Zoox has raised $290 million and is rumored to have drafted self-driving car legislation in multiple states.

Zoox isn’t even mentioned in the report.

Tier 1 suppliers Delphi and ZF are on it, but Bosch isn’t. Does NVIDIA deserve a spot? How much credit is Volvo getting from the Autoliv/Zenuity alliance? How much of BMW’s position is derived from the Intel/Mobileye partnership?

Imagine the same list for the computer industry. Would you place Foxconn on the same list as Apple and Dell, and declare Foxconn the loser? Of course not, and yet Navigant has ranked key suppliers in the self-driving “race” alongside manufacturers with whom they share few (if any) characteristics.

So who is really #1 in Self-Driving Cars?

This one’s easy. The company closest to delivering a Level 4 self-driving car—even if it’s geofenced. Why? Because multiple overlapping industries will realign overnight, from car rental to sharing platforms to public transportation. It will be exactly like bringing a gun to a knife fight (but without the guns or knives). Mobility is the watchword, and self-driving technology is a foundation element. Everyone will want in and nothing will be sacred. Whomever can deliver will be able to pick and choose their partners, as no one can afford to be left out. If Morgan Stanley is right when they say that only six car companies are likely to survive, then the only question is who will be happy to play Foxconn in exchange for the technology they lack.

By that standard, Waymo and Tesla are tied for first in the only race that matters: the mobility race. If it’s Waymo, they’ll be able to sign up everyone low on Navigant’s list (and unlisted players like FCA), forming an alliance eradicating any lead held by those higher up on the report. If it’s Tesla, their allegedly inflated stock price will justify itself, spike overnight, and crack open capital markets guaranteeing not only the survival the shorts love to question, but long-term dominance of a sector whose future shape remains unclear.

I'd bet on Waymo, who already have a partnership with FCA.

As Brad Templeton pointed out, Navigant’s rankings aren’t completely wrong, but a true leaderboard would have to include alliances whose true nature remains unclear, define all the players across the mobility continuum, and attempt to define the nature of “mobility” itself. Silicon Valley and traditional automakers need each other, but they don’t need all of each other. Beware reports with conclusions born in a vacuum, for that’s where billions of dollars are most likely to disappear.

Alex Roy is Editor-at-Large for The Drive, host of Autonocast.com and /DRIVE on NBC Sports, author of The Driver, and set the 2007 Transcontinental “Cannonball Run” Record in 31 hours & 4 minutes. Follow him on Facebook, Twitter and Instagram.