One AI Driver, Three Badges: Stellantis and Wayve in Turin

A Fiat 500e is about as small as a Stellantis car gets. A Maserati Grecale is a luxury SUV with a price tag to match. Last week in Turin, both drove hands-free, door to door, on the same piece of AI software, and by the companies' account each car went from integration to road testing in under four weeks.

The demo rides took place at Wave by Vento in Turin, October 7 to 9, with Wayve saying the cars ran at the Lingotto track. They used "STLA AutoDrive powered by Wayve," which pairs Stellantis' in-house driver-assistance platform with the Wayve AI Driver. The two companies call it Level 2++ supervised automated driving: the car handles much of the work on highways and city streets, but the driver stays responsible and must be ready to take over at any moment. It is not a robotaxi. It is not "self-driving" in the legal sense either.

Same brain, different bodies

The headline isn't that a prototype drove itself around a Turin test track. Prototypes do that every week somewhere. The headline is that two vehicles at opposite ends of Stellantis' range ran one core system, after a third, a Jeep Grand Cherokee, did it first.

The Jeep was shown in connection with Stellantis' Investor Day in May, when the partnership was announced. On that day Wayve CEO Alex Kendall told Reuters the startup would integrate its AI Driver into a Stellantis prototype "in less than two months." Wayve now says the Jeep took under two months and that the Fiat and Maserati each took under four weeks. All of those timelines are company claims, not independently verified, but the direction matters: each new vehicle got cheaper and faster to bring up.

Kendall framed it as the point of the whole exercise. "We're building a general-purpose AI Driver that can power any vehicle, anywhere," he said, according to Automotive World, adding that Wayve had "brought the same core intelligence to the Fiat 500e and Maserati Grecale." Stellantis' own line, per Carscoops, is that the technology can be adapted across its lineup without developing separate systems for each model.

That is the software-defined OEM in one sentence: many badges, one driving stack.

Why Wayve's architecture makes this possible

Wayve's pitch rests on two technical choices. Its driver is mapless: it does not need the detailed HD maps that traditional systems rely on, which Wayve says simplifies global deployment and cuts costs. And it is camera-first with radar, with lidar optional for automakers that want it. Wayve has said it designs for hardware costing under $1,000, according to The Next Web's coverage summarized by The Outpost.

Those choices are what let one model sit in a city car and a premium SUV. A rules-based system tuned to specific sensor placements and mapped roads has to be re-engineered for every vehicle. A learned driver that generalizes needs calibration and validation, not a rebuild. That is the licensing thesis Flux covered last month, when Wayve's London robotaxi pilot looked tiny next to Waymo: Wayve sells the driver, not the fleet. Turin is what that looks like when a carmaker with a sprawling portfolio actually buys in.

It also explains the stakes for Stellantis' current system. Its Hands-Free Active Driving Assist works on roughly 200,000 miles of mapped North American roads, according to Carscoops. The Wayve-based system is meant to cover far more road types, including urban streets.

Filosa's math and Kendall's seatbelt

Stellantis CEO Antonio Filosa shared the stage with Kendall on October 9. Reuters reported that Filosa said the tie-up could significantly cut self-driving costs and development time, and that Stellantis currently works on a 24-month development timeline for these systems, which the partnership aims to shorten. Set that against four-week integrations and the appeal is obvious, even allowing for the gap between a demo-ready prototype and a validated production car.

Kendall went further on where this ends up. In coverage of the event, he argued that every car will eventually need an AI driving system the way it needs seatbelts and emergency braking. "It's really morally wrong to not have some system, some intelligence in the vehicle to save those lives," he said, as reported by The Next Web. That is a safety argument, and also a market-sizing one: if AI driving becomes standard equipment, the supplier of the driver gets a slice of every car.

The Momenta asterisk

Stellantis is not betting on one supplier worldwide. In late September, Chinese autonomous-driving company Momenta signed a global strategic partnership with Dongfeng Stellantis, the joint venture Stellantis and Dongfeng formed in August, to co-develop driver assistance for new mass-produced Peugeot and Jeep models across China, Europe and other markets, Reuters reported. Momenta's software already runs in more than 1.2 million vehicles.

So "one AI driver, many brands" comes with a geographic footnote. Jeep, the brand that debuted the Wayve integration, will also carry Momenta-developed assistance through the joint venture. One reading is that Stellantis is dividing driving software by region, with Wayve leading North America, where Chinese connected-vehicle software faces US restrictions. Stellantis has not framed it that way publicly, and which brand gets Wayve first in 2028 is still unannounced. Carscoops notes it will not be the redesigned Jeep Cherokee, which is due in the second half of 2029.

What to watch

The 2028 North America launch is the only date on the board, and it is two model years away. Between now and then, three things decide whether Turin was a turning point or a well-produced demo.

First, validation. Bringing a car up in four weeks is engineering; signing off a Level 2++ system for public roads across thousands of edge cases is a different, slower discipline, and nobody has said how Stellantis will shorten that part.

Second, the supplier map. If Stellantis keeps Wayve and Momenta on separate continents, it is not running one AI driver. It is running two, with brand-agnostic software on each side of a regulatory wall. That is still a software-defined company, just a more complicated one.

Third, the driver handoff. Level 2++ asks a human to supervise a system that is good enough to lull them. How Stellantis monitors attention, and how it markets the feature, will matter as much as the model.

The bet itself is clear. Stellantis is treating driving as a software layer that sits above the badge, and Wayve is treating carmakers as distribution. If both hold their end, the Fiat and the Maserati will differ in leather and horsepower, not in how they see the road.