Xpeng’s New Robotaxi Is Already in Production While Tesla’s Still Testing

While Tesla continues its methodical testing phase for a robotaxi that may or may not ever actually exist as a consumer product, Xpeng has already moved into production with its GX-based robotaxi. This is the kind of decisive action that separates actual autonomous vehicle programs from vaporware—and it represents a significant shift in how Chinese automakers are approaching self-driving technology.

The Shortcut That Makes Sense

Rather than designing a bespoke robotaxi platform from scratch like Tesla’s rumored Cybercab project, Xpeng took the smarter route: adapt an existing production vehicle. The company’s robotaxi is essentially a heavily modified version of the GX SUV, which already sells to regular consumers. This decision dramatically cuts both development time and manufacturing costs—two things that matter infinitely more in the real world than engineering purity.

Under the hood (or rather, in the computer), Xpeng equipped the robotaxi with four in-house Turing AI chips delivering 3,000 TOPS of processing power, plus steer-by-wire steering. The company developed the autonomous system entirely in-house, giving it complete control over the tech stack. That’s a major advantage when you’re racing to deploy a service before the regulatory environment shifts again.

Vision-Only, Just Like Tesla—But Already Working

Here’s where Xpeng gets interesting: it copied Tesla’s most controversial decision. The robotaxi uses a vision-only autonomous system with no LiDAR or high-definition maps, relying purely on cameras and an advanced AI model. This is the same bet Tesla made with Autopilot and Full Self-Driving, and it’s been wildly debated by engineers for years.

The difference is that Xpeng isn’t theorizing about this approach—it already has permits to test Level 4 autonomous systems on Chinese public roads and has been running tests since January across multiple cities. By early 2027, the company plans to remove human safety operators entirely, moving from supervised autonomy to true driverless operation. Tesla’s robotaxi timeline remains vague at best.

Three Seating Options, Real Interior Upgrades

Xpeng will offer the robotaxi in five-, six-, and seven-seat configurations, maximizing flexibility for fleet operators. The interior gets passenger-focused upgrades: privacy glass (passengers probably don’t want to be gawked at), rear entertainment screens, and new plush seating. It’s not revolutionary, but it shows the company is thinking about the actual user experience, not just the engineering challenge.

The one unknown: Xpeng hasn’t disclosed whether the robotaxi uses the consumer GX’s range-extender hybrid powertrain or goes full electric. For a vehicle that will likely spend most of its life in constant operation, that efficiency question matters. EPA fuel economy data suggests hybrids still have an advantage in per-mile operating costs, but extended autonomy testing data would tell the real story.

The Broader Picture: China’s Robotaxi Reality Check

Xpeng’s move to production reflects a harder truth that the Western robotaxi narrative often glosses over. China’s autonomous vehicle industry isn’t waiting for perfect sensors or settling abstract philosophical debates about safety standards. Companies like Xpeng are building, testing, and deploying. Meanwhile, Tesla has been promising a robotaxi for years without shipping a consumer vehicle.

That said, China’s progress hasn’t been friction-free. Earlier this year, Chinese regulators hit pause on new robotaxi permits after 200 autonomous vehicles got stuck in traffic, a humbling reminder that scaling this technology isn’t a straight line to glory. Xpeng’s plan to launch pilot operations in the second half of 2026 will test whether the company can keep its momentum in an increasingly cautious regulatory environment.

Why This Matters More Than You Think

The competitive dynamic here is important. Tesla built its empire partly on being first with bold ideas, but Xpeng is demonstrating that execution—not innovation theater—is what counts in autonomous vehicles. Building a robotaxi based on an existing platform isn’t sexy, but it’s exactly how you get a working product into the real world while your rivals are still perfecting specs sheets.

The vision-only approach is the real wildcard. If Xpeng’s AI-only system works reliably at scale, it validates Tesla‘s bet and opens a path to cheaper autonomous vehicles globally. If it stumbles, expect a rush back to LiDAR and HD maps. Either way, we’re about to get actual data instead of promises, and that changes everything.

Sources: Carscoops