GM Didn't Stay in the Robotaxi Race. It Changed Races.

General Motors spent $12.1 billion building a robotaxi company, then killed it, then immediately started talking about building something that looks a lot like the same thing. That is weird. Or at least it should be, if you're trying to figure out what GM actually is.

The basic point is not that GM killed Cruise and might come back. The basic point is that GM killed one kind of autonomous business and started building a different one. They look similar from far away. Up close they have completely different economics, different customers, and different risk profiles. Confusing them is like treating a bank and a credit card company as the same business because they both deal in money.

GM announced in December 2024 that it would stop funding Cruise's robotaxi operations. It owned about 90% of Cruise, having acquired the startup for over $1 billion in 2016 and poured another $11 billion-plus in over the following decade. GM laid off roughly 1,000 Cruise employees - nearly half the workforce - in February 2025 and expected the shutdown to save $1 billion annually. The robotaxi fleet was parked. The ride-hailing service was done.

The official story was that GM was retreating from robotaxis to focus on what it called "advanced driver assistance" - a technical term for features that help human drivers but don't replace them. Cruise's remaining engineering assets were absorbed into GM's corporate autonomous driving team. In August 2025, GM was already rehiring former Cruise employees.

But here's the part the headline gets wrong. This isn't GM "still in the race." This is GM admitting that the robotaxi race was a terrible business for an automaker to be in, and switching to one that actually fits.

A robotaxi is a rideshare business. You own the fleet, you pay for the infrastructure, you deal with regulation, liability, and the kind of unit economics that have so far convinced no one - not Waymo, not Tesla, not Cruise - that they work at scale. You're trying to be Uber but without drivers, which means you're Uber's problem AND Uber's margin pool AND your own balance sheet at the same time. The capital requirement is enormous, the path to profit is theoretical, and the regulatory risk in every city is unique. GM was trying to build a logistics business on top of its car business, which is fine until you realize the logistics business might cost more than the cars.

Super Cruise is a different machine. It's a software subscription feature on cars GM already sells. The customer is the car owner, not the rider. GM doesn't need to own the fleet or dispatch trips or deal with municipal ride-hailing permits. The customer has already paid $80,000 to $130,000 for a Cadillac or Chevy, and Super Cruise adds a monthly fee on top. Super Cruise generated $234 million in revenue in 2025, and GM expects that to climb toward nearly $400 million. That's not a robotaxi business. That's a SaaS business riding on top of a $170 billion vehicle sales base.

The gap between those two models is the gap between the story and the mechanism. GM spent $12.1 billion trying to build the former and is now building the latter with a fraction of the cost and a business model that actually connects to what the company already does.

The technology pipeline is still real, just reoriented. GM killed Ultra Cruise - the ambitious system meant to cover 95% of U.S. and Canadian roads - but in October 2025 announced a successor: an "eyes-off" version of Super Cruise coming to the 2028 Cadillac Escalade IQ, starting with highways. GM began supervised highway road testing in California and Michigan in March 2026. It's also integrating Google's Gemini AI for in-car conversational features. This is the same sensor stack, the same mapping work, the same learning - just sold to the person behind the wheel instead of to nobody at all.

That distinction matters because the economics are not the same even if the neural networks are. Super Cruise revenue is recurring, it scales with GM's existing vehicle deliveries, and it doesn't require building dispatch infrastructure or maintaining a driverless fleet. The customer base is people who already bought expensive GM vehicles - a captive audience with demonstrated willingness to pay for premium features. The robotaxi customer base is everyone in a given city who might take a ride, competing against Uber, Lyft, public transit, and their own car keys.

GM's stock trades around $81 with a $72 billion market cap, which suggests investors are pricing GM as an automaker with a decent software sideline rather than a robotaxi platform waiting to happen. That's probably the right read. The Super Cruise business, even at the projected $400 million, is a rounding error compared to GM's core operations. But it's a rounding error with gross margins that look more like software than auto hardware, and that changes how you think about the company's trajectory.

The simplest model is this: GM treated Cruise as a R&D lab that cost $12.1 billion to run. The lab produced sensor stacks, mapping data, and a hands-free driving system that now ships as a feature on Cadillacs. The lab closed. The product stayed. That's not a graceful pivot story. It's just what happens when a company realizes the business model it built around the technology was wrong.

So no, GM isn't "still in the robotaxi race." GM exited the robotaxi race and is now in the slightly more profitable business of charging people who already own Cadillacs an extra monthly fee to take their hands off the wheel. It's a smaller dream. It's a better business. The question for investors is whether $234 million in annual Super Cruise revenue, growing toward $400 million, is enough to make the $12.1 billion write-off feel like tuition instead of a mistake.