Self-driving cars are mostly Fake News

This article is the first in a series to help GTK’s network understand and invest successfully in “self-driving” and adjacent markets.  Please forward to friends to subscribe via the link at the bottom.

The transportation technology market is striking in two ways:

  • The $2 trillion annual market size means small & broad impacts, as well as nichey & deep changes, are big news in themselves.
  • There is a significant disconnect between corporate leadership (or at least what leadership thinks investors want to hear) and corporate technologists regarding technical feasibility: rumors of the widespread death of steering wheels are greatly exaggerated.

Every year the world purchases approximately 75MM passenger and 25MM commercial vehicles, with enormous benefits and downsides:

  • Trillions of dollars tied up in infrequently used vehicles.
  • Driver fatigue and 1.3MM accident deaths per year.
  • Enough fuel burned to fill the TransAmerica Pyramid twice every day.​

To make sense and make money in the transportation technology market, focus on the following:

  • Specific features applicable to broad populations.  These include features consumers seek (blind spot assist, performance, and entertainment systems) or ones that auto manufactures standardize (Toyota with collision avoidance) or almost force to be bought (Mercedes’ version of “AutoPilot”).  Consumers don’t pay for expensive, optional safety features they only use episodically.​
  • “Driver-less” deployment in “closed field applications” (rather than general purpose consumer vehicles): shuttles, buses and taxis are already deployed in limited areas (the Bay Area, Singapore and Las Vegas, respectively).  Waymo is making great progress in this segment. Trucking is interesting too.

Furthermore, for cars to someday drive themselves, hundreds of thousands if not millions of suitably equipped cars are needed to collect information on how humans drive.  This is a big reason why even if you don’t purchase the AutoPilot option Tesla includes the 22 sensors and keeps the data rights, and this is why “LIDAR” vision startups have attracted $100MMs of capital and many buyouts.

So you shouldn’t make decisions assuming the prevalence of self-driving car purchases in the next many years, but lots of the components (cameras, radar, connectivity) are already table-stakes in developed-world new vehicles and increasingly elsewhere.  Affordable and aesthetic autonomy isn’t there yet, which is why the big deals are about talent, not product: acquirers and investors will pay a $20MM+ per Ph.D. for teams of technologists. GM and Ford paid a $1BN each for Cruise and Argo, respectively.

So forget about the Fake News of getting rid of the steering wheel, but come and ride shotgun with us: there’s lots of value to create and capture. Upcoming topics:

  • “What’s with the weird-mobiles?” or What it takes for the car to drive the car.
  • “If computers can play chess and Jeopardy, why can’t they drive?”, or How to get smart on AI and Autonomy in five minutes.
  • “I’ll take the premium stereo instead,” or how autonomy features sell.
  • “Tesla: deservedly hated but under-appreciated” or why the Silicon Valley trope that “legacy” companies are in trouble could become true.
  • “Where is everybody?” Or why Apple, Samsung, and others shouldn’t be forgotten.
  • “I’m from the government, and I’m here to help,” or why claims of regulatory bottlenecks are Fake News.