Pony AI Inc. is chasing robotaxi scale—and just found a way into Europe

Pony AI’s big week: profit, permission, and a passport stamp

Robotaxis have spent the last decade living in the “cool demo, see you in 2030” zone. Then Thursday happened.

On March 26, 2026, Pony AI Inc. said it expects to more than double the number of robotaxis running its tech to over 3,000 vehicles across more than 20 cities globally in 2026—and, more importantly, it framed that growth as something that can coexist with a first-ever quarterly profit. That’s a meaningful vibe shift for a category that usually treats losses as the cost of doing futuristic business.

The timing isn’t accidental. The company also dropped news that it’s landing in Zagreb, Croatia—yes, Croatia—as part of a partnership that includes Uber Technologies and a local autonomous startup called Verne, positioning it as a contender in what’s being pitched as Europe’s first commercial robotaxi service.

Why Zagreb matters more than it sounds

If you’re wondering why an autonomous driving company with deep China roots is making headlines in a European capital that isn’t exactly a global tech hub, that’s the point.

Robotaxis don’t scale like apps. Every new city means negotiations with regulators, mapping work, fleet ops, safety processes, and someone willing to attach their brand to the whole thing. Europe is famously cautious around transportation and safety rules, so getting a real, commercial-facing launch pathway is less about “one more city” and more about proving you can navigate the slow, bureaucratic parts of globalization.

And Uber’s involvement is a signal. In 2026, Uber doesn’t need another flashy pilot; it needs a long-term answer to the question investors always ask in the background: what does Uber look like when human drivers aren’t the default? Partnering is the pragmatic move—Uber brings demand and distribution, while partners bring the vehicles and autonomy stack.

Profit is the headline, but the subtext is unit economics

Pony AI calling out a first-ever quarterly profit on March 26, 2026 isn’t a declaration that the robotaxi business is suddenly easy. It’s a declaration that the company believes the math is starting to work.

The robotaxi industry has a simple, brutal rule: you can’t “story” your way to scale forever. Eventually, the cost per vehicle, the cost to keep it running, and the ability to generate repeat paid rides have to stop being theoretical.

Pony AI has been signaling for a while that it’s trying to reduce what each vehicle costs to deploy and operate, and it has leaned into partnerships that let it avoid owning every part of the fleet itself. The “over 3,000 vehicles” target for 2026 is basically Pony saying: we’re ready to stop acting like a lab and start acting like a transportation company.

Where the stock fits into the narrative

Pony AI Inc. American Depositary Shares (PONY) trades like the market can’t decide whether it’s looking at an industrial-scale AI platform or a high-beta experiment. That uncertainty is understandable: autonomy is both a real product and a regulatory process, and those two timelines rarely match.

Still, a few things are now clearer than they were even a year ago:

• The company is publicly tying 2026 to fleet expansion on a specific, measurable scale (3,000+ vehicles).
• It’s attaching that expansion to profitability language, not just “growth.”
• It’s pushing outside its home turf with a high-visibility consumer use case in Europe.

If you’ve been treating robotaxis as a distant sci‑fi checkbox in your mental model of “AI investing,” Pony’s latest updates are a reminder: some of this is turning into regular business execution—just with weirder vehicles.

The broader takeaway: autonomy’s winners will be the ones who can ship

In 2026, the autonomy conversation is less about who has the cleanest demo video and more about who can stack wins that look boring: permits, fleet ops, partnerships, repeat rides, and improving economics.

Pony AI is trying to prove it can do exactly that—at the same time.