Wayve's AI Self-Driving Tech to Power Stellantis Vehicles by

Wayve's AI Self-Driving Tech to Power Stellantis Vehicles by 2028

Stellantis partners with Wayve to bring hands-free driving to its North American vehicles by 2028. Wayve's end-to-end neural network adapts to any sensor, chip, or vehicle without HD maps. The deal follows Wayve's $1.2B Series D and a prior Nissan partnership.

4 min readMay 21, 2026

Wayve's AI Self-Driving Tech to Power Stellantis Vehicles by 2028

Stellantis Picks Wayve for Hands-Free Driving

Stellantis, the automaker behind Jeep, Ram, Chrysler, and Dodge, has selected UK-based self-driving startup Wayve to deliver hands-free driving technology for its North American vehicles by 2028. The deal was announced Thursday at Stellantis’ investor day in Michigan.

This is Wayve’s second automaker contract, following a partnership with Nissan. It comes after Wayve closed a $1.2 billion Series D funding round with strategic investors including Nissan, Stellantis, Microsoft, Nvidia, and Uber.

What Wayve Brings

Wayve’s self-driving system is fundamentally different from traditional approaches. Instead of relying on high-definition maps, lidar, or specific hardware, Wayve uses an end-to-end neural network trained purely on driving data. The software adapts to whatever sensors and chips are already in the vehicle.

Wayve CEO Alex Kendall emphasized the system’s adaptability: "Our AI is so adaptable; we can generalize to the variety of products that they offer, and means that because of the diversity of sizes, shapes of vehicles, different driving styles, different geographies they run in our AI is built to scale across them all."

Technical Details

Wayve offers two products: a hands-off assisted driving system (comparable to Tesla Full Self-Driving Supervised) and a driverless system for robotaxis. Stellantis will use the hands-off, eyes-on version.

Kendall noted that a prototype for Stellantis was developed in just two months. Within weeks, engineers had a vehicle driving using the AI system. The software runs on existing chips from Stellantis’ OEM partners, reducing hardware costs.

Stellantis’ Turnaround Plan

Stellantis plans to launch 11 new vehicles in North America by 2030 as part of a $70 billion turnaround plan. Seven will be priced under $40,000, and two under $30,000. It is unclear if Wayve’s tech will appear in these lower-cost models, but Wayve’s efficiency pitch suggests it’s plausible.

Why This Matters for Developers

Wayve’s approach represents a shift from rule-based autonomy to data-driven neural networks. Developers building AI systems should note the emphasis on generalization: one model trained on diverse data can handle multiple vehicle types, sensor configurations, and driving conditions. This reduces the need for per-vehicle calibration.

For those working with embedded systems, Wayve’s ability to run on commodity chips is a key advantage. It avoids vendor lock-in and allows automakers to use their existing compute platforms.

The partnership also highlights the growing interest in end-to-end learning for complex real-world tasks. Wayve’s system is not modular — perception, planning, and control are all handled by a single neural network. This is a departure from traditional autonomous stacks that separate these functions.

What’s Next

Wayve and Stellantis are targeting the North American market first. With a 2028 timeline, developers should watch for integration details, especially regarding the hands-off system’s performance in diverse environments like highways, suburbs, and cities.

For engineers interested in self-driving technology, Wayve’s approach offers a glimpse into a future where AI models replace hand-coded rules. The ability to adapt to different vehicles and hardware could lower the barrier for automakers to deploy autonomous features.

If you’re building AI for robotics or autonomous systems, consider how end-to-end training might simplify your stack. Wayve’s success suggests that data-driven generalization can outperform traditional engineering in some domains.

Summary

Deal: Stellantis to use Wayve’s hands-free driving by 2028 in North America.
Tech: End-to-end neural network, no HD maps, works with any sensors/chips.
Scale: Stellantis plans 11 new vehicles by 2030, some under $30K.
Context: Second automaker deal after Nissan; $1.2B Series D.

Developers should track Wayve’s progress as a case study in applying deep learning to safety-critical systems. The company’s ability to deliver a working prototype in two months is a notable achievement.

Editor's Take

I've been skeptical of end-to-end systems for safety-critical tasks, but Wayve's ability to prototype in two months is impressive. The lack of HD map dependency is a genuine advantage — maps are expensive to maintain. Still, I wonder about edge cases: without explicit redundancy, a single model failure could be catastrophic. I'd like to see how they handle rare scenarios. For now, this is a bet on data scaling over hand-crafted rules.

— DevDigest Editorial

Key Takeaways

•End-to-end neural networks can reduce deployment costs by eliminating HD maps and sensor-specific tuning.
•Wayve's system runs on existing OEM chips, avoiding expensive hardware upgrades.
•Prototyping in two months suggests rapid iteration cycles for AI-based driving features.

Why It Matters

Wayve's end-to-end neural network approach could disrupt traditional autonomy stacks. If successful, it may accelerate hands-free driving adoption across multiple automakers, reducing hardware costs and complexity. Developers in AI and embedded systems should watch how this scales.

#ai#autonomous-vehicles#self-driving#Stellantis#Wayve

Get the weekly digest

Every Sunday - top tech stories, industry breakthroughs, and developer tools delivered to your inbox.

No spam, unsubscribe anytime.

LLM Judges Show 47-Point Score Swing in Agent Benchmark

A benchmark of six AI agents across eleven skills, graded by three different LLM judges, reveals score swings up to 47 points and shifting rankings. The study shows LLM judges favor their own model family, with Opus giving itself a 4.6-point boost. The author recommends multiple judges and binary rubric criteria for reliable evaluations.

4 min read·about 2 hours ago