By Colin Barnden
What’s at stake?
Autonomous driving was supposed to save lives, restore the environment, and make transportation accessible to all. Suddenly the future of the industry looks shaky.
Ford’s decision to call time on Argo AI marked the end of the beginning of autonomous driving. Right now, no one knows what happens next, but short-term developments aren’t going to be dictated by technology demonstrations or regulatory decisions. The future of autonomous driving rests entirely in the hands of investors.
Let us begin with what went wrong at Argo. The answer is akin to what is wrong at the other autonomous developers: Roads are complex, not complicated. AI and machine learning is too brittle to navigate phase-transitions inherent in complex systems. Simulation and brute-force training cannot “solve” the reality of emergent properties.
Translation: Weird stuff happens on public roads. It always has, it always will.
We frequently hear the argument that humans walked on the moon, so why shouldn’t tech companies succeed in developing cars that drive themselves on public roads? The flaw in the argument is that space travel is complicated, not complex. However, it is futile to make such fundamental observations in the face of an autonomous driving industry that runs on magical thinking and gaslighting.
So let us state the problem in the simplest terms possible: Machines may be safer than human drivers at some time in the future. But precisely when, and how much it costs to get there, no one knows.
Recommended: Argo AI’s Demise Automakers’ AV Miscue
We’ve all seen and heard the forecasts. In 2012, Google’s Sergey Brin promised “autonomous cars for all” within five years. In 2017, Intel told us the ‘Passenger Economy’ will be worth $7 trillion.
What followed was a race to be first to “solve” autonomy. Autonomous taxis, toasters (fixed-route shuttles) and trucks. You name it, venture capitalists, sovereign wealth funds, and “vision” funds battled each other to invest. They all wanted to claim their slice of the trillions, and no one wanted to listen to the skeptics.
Then Ford dropped a stun grenade, and everything changed.
For years we were told the future of transportation was C.A.S.E. (Connected, Autonomous, Shared, Electric). For robotaxis, the future suddenly looks more like J.U.N.K.
CEO Jim Farley has cauterized Ford’s autonomous losses by severing the Argo limb. That was bold leadership. But let’s consider what the decision means for GM and for its autonomous subsidiary Cruise, developer of robotaxis and the Cruise Origin toaster-on-wheels.
Farley just forced GM’s CEO Mary Barra to go all-in with Cruise against Waymo. Barra can easily make the call to continue funding Cruise, but events thereafter would be out of her control. The cost to GM to commercialize the Cruise technology is still unknown; the California Public Utilities Commission could pull Cruise’s driverless permit for repeatedly blocking traffic. If Cruise kills someone, it is game over.
The other option now open to Barra is to fold and close Cruise. But then GM gets positioned as a follower of Ford, which made the big call first. Ford forcing GM to go double-or-quits on Cruise is a masterstroke.
In comparison, Waymo has no automaker parent to turn to, and Alphabet stopped funding it several years ago. Alphabet is experiencing its own short-term challenges with the performance of Google and looks highly unlikely to start writing billion-dollar checks to fund the Waymo moonshot.
If willing investors are still around to fund autonomous driving technology, Argo didn’t find them. But in the current investment environment, can Waymo still obtain funding? Will GM’s investors permit Barra to continue pouring money into Cruise?
Share prices for market leaders such as Amazon and Meta (Facebook) have fallen sharply in the last week, indicating that investor sentiment toward “big tech” has suddenly soured. This does not bode well for more expensive bets on autonomous driving. Ironically, the share price of Mobileye gives a precise indicator of investor sentiment on a mark-to-market basis.
It seemed the first company to “solve” autonomy would be crowned the winner. Instead, the winning move looks like being the first investor to cut its losses. Farley is the decisive leader who got the big call right. Attention now shifts elsewhere, precisely at the moment investors have become unforgiving.
Post-Argo, Waymo is left battling with Cruise in a fight to the death. My judgment is that both Barra and Cruise will be gone within 12 months. Waymo has at best one funding round left to “solve” autonomy. But since roads are complex, not complicated, its likelihood of success looks extremely doubtful.
If no one is left to fund development, more company failures are inevitable. Focus will inevitably turn to autonomous trucking and logistics suppliers, such as Aurora, Embark, Gatik, TuSimple, and so on to the lidar suppliers. The future of these companies rests in the hands of investors.
If the shuttering of Argo AI marked the end of the beginning, the beginning of the end is the realization and acceptance that technological development in the automotive industry is always an iterative process.
The near-term future of driving is a human-machine collaboration. We will see very sophisticated supervised automation systems on divided highways, with advanced driver monitoring using human factors research to support drivers in supervisory roles. Argo’s achievements will live on, but in a future version of Ford’s BlueCruise, not in robotaxis–as was so widely expected.
What about the future of autonomous driving? Right now, no one knows. But every investment professional understands that you do not contunue funding a basket case indefinitely.
Companies don’t go bust simply because they lose money, they go under because they run out of cash. This year has seen investor sentiment sour significantly against big tech. The next 12 months will be a rough ride for the autonomous driving industry. Argo AI won’t be the only casualty.
Colin Barnden is principal analyst at Semicast Research. He can be reached at firstname.lastname@example.org.