Reflections on Ride AI 2026

Plus, Uber bets big on Lucid and Tesla expands in Texas.

This is the Ride AI newsletter: The most comprehensive weekly digest of news and intelligence at the intersection of technology and transportation.

To see all of the sessions from Ride AI 2026 in full, please subscribe to our Youtube channel. All videos will be posted by tomorrow morning.

Reflections on Ride AI 2026

What an incredible event.

From James’s keynote highlighting how controversial technologies become ones we can’t live without through the power of marketing, to hearing from the companies that are actually on the ground doing the work to scale AVs, it was an information-dense and inspirational day.

In the morning fireside chats, Alex Kendall, CEO of Wayve, sat down with Ed Niedermeyer for an inside look at their progress towards generally available, zero-shot autonomy. Chris Urmson reflected on how the 15-year journey of industrializing autonomous trucks had been harder than originally anticipated.

Lior Ron of Waabi discussed the company’s aim to eliminate the need for traditional “hub-to-hub” crutches by mastering surface streets for door-to-door delivery, and how Waabi expanding into robotaxis was because their “single brain” architecture could already drive multiple vehicle forms.

In panels, Woven Capital, Morgan Stanley, and Wells Fargo talked about how the standards for autonomous vehicle companies entering public markets now demand proof of scaled commercialization and mature unit economics rather than just technological promises, alongside a prediction that the market for personally owned AVs is expected to remain significantly larger than that of the robotaxi segment.

Experts from Foxglove, GM, and Toyota Research Institute (TRI) explored how autonomous vehicle lessons are spilling over into other robotics sectors like logistics and manufacturing. Alan Ohnsman of Forbes dove into what’s required to operate under extreme circumstances with Tensor, Mobileye, and Motional. Uber, AVIA, and Aon addressed how in the absence of perfect federal regulations, AV companies must use commercial agreements and safety standards to build consumer confidence. Moove, Terawatt, Voltera, and MOIA got into the weeds discussing “picks and shovels” assets like charging hubs and maintenance depots, and how automated “next-gen” depots that can handle tasks like robotic cleaning and wireless charging are needed to maximize vehicle throughput.

In the late afternoon slate, Rocket Drew from The Information grilled Encord, Weights & Biases, Aeva, and Zenseact in a technical session focused on the role of vision language models and “world models” in generating synthetic training data. Xiaodi Hou of Bot Auto and Michael Wiesinger of Kodiak AI shared what really mattered in trucking, autonomous or not (cost, cost, cost.) and Paul Konasewich of PACCAR shared some secrets of how autonomy hardware integration works between the OEM and the developer in a panel moderated by Thomas Wasson from FreightWaves. And finally, I ask, “What in the world is really going on in China?” to my panelists Tu Le from Sino Auto Insights and Dr. Leo Wang of Pony.ai, one of China’s leading AV companies.

And finally, in pecha kuchas and keynotes, McKinsey’s Ani Kelkar and Darius Scurtu presented data showing how the AV market is shifting from focusing on tech to a focusing on scaling, with unit economics as king. David Moss detailed his experience driving nearly 13,000 intervention-free miles across the U.S. using Tesla’s Full Self-Driving (FSD) version 12. Nigel Drego of Quadric introduced “Chimera,” a programmable and efficient licensable silicon IP core designed to handle evolving model architectures without requiring expensive hardware respins. It turns out that while AI models evolve every one to two years, automotive silicon must be built to last over a decade to support updates for models that do not exist yet.

Ethan McKenna shared his journey in building Robotaxi Tracker, a platform to provide empirical data on fleet counts and wait times by scraping traffic cameras and reverse-engineering ride-hailing apps. Omar Qazi presented data from his FSD Database project, which uses real-time vehicle telemetry to evaluate the performance of Tesla’s Full Self-Driving software independently of the manufacturer. Ashu Rege, head of DoorDash Labs, introduced “Dot,” an autonomous bike-lane robot uniquely suited for the “Goldilocks” zone of delivery where drones have too low a payload and robotaxis are too large.

Outside SFJAZZ, the action didn’t stop. Toyota Research Institute presented its research testbed complete with a VR data collection simulator, letting attendees experience what it’s like to collect snowy road data. Wayve gave demos of its zero-shot autonomy system that will go live in London later this year. Tesla gave attendees a taste of personally owned autonomy with FSD demos. Tensor, Uber, Lucid, Nuro, and MOIA brought production-intent vehicles, providing an exclusive first look at real autonomous vehicles that will be on the road later this year. DoorDash’s Dot attracted waves of onlookers with its cute looks and sounds and forward-thinking design, and Robot.com’s sidewalk robots doled out free swag to attendees.

Sitting in the audience listening to the speakers on stage share solutions to the challenges they’ve overcome, it became apparent to me that indeed, this is not just another hype cycle. Wayve’s fully embodied AI is doing real zero-shot drives. Waymo has already scaled to 10+ cities. Pony.ai has had robotaxi service across China for the past four years, and is already expanding into Europe. These are real products and services actually hitting the road. We are already living in a world where AV technology exists and more importantly, where it actually works.

The challenge now is marketing. Everybody has that first Waymo ride with that “wow” moment, which, make no mistake, is an irreplaceable piece of marketing. As the industry scales, I’m hoping those “wow” moments grow, and vehicular autonomy goes from novelty to commonplace.

From all of us at Ride AI, thank you so much for attending. I hope you’re looking forward to next year!

Top Story

Uber is deepening its bet on Lucid in a big way. Uber is increasing its total investment in Lucid to $500 million (an additional $200M on top of its prior commitment), PIF affiliate Ayar Third Investment Company is putting in another $550 million, and the fleet commitment between Lucid and Uber has expanded to at least 35,000 vehicles spanning Lucid Gravity and a forthcoming Lucid Midsize platform. The week prior, Nuro kicked off employee testing of the premium robotaxi service in San Francisco using Lucid Gravity SUVs with Nuro’s autonomous stack and Nvidia Drive AGX Thor compute, bookable through the Uber app. Commercial launch in the SF Bay Area is targeted for later this year.

Between Uber, PIF, and the fleet commitment, Lucid is now being treated less like an EV startup and more like critical infrastructure for the next wave of autonomous mobility. As part of this push, Uber has also invested in Nuro, although the exact sum is yet undisclosed.

Domestic News

Waymo is now fully open to the public in Miami and Orlando, with over 150,000 riders already served from the initial interest list before the public opening. Miami is also getting highway driving capability as an opt-in through the app.

Tesla’s robotaxi service is now rolling out in Dallas and Houston, making them the second and third cities after Austin, all three in Texas. Per crowdsourced tracking data from Robotaxi Tracker, there’s currently one active vehicle in each new city compared to 46 in Austin, so “rolling out” is doing some work in that sentence… but it’s moving for sure.

The CPUC paid a site visit to Zoox’s first serial production facility in the Bay Area this week, with Commissioner Baker and staff touring the purpose-built vehicles up close. Zoox hasn’t announced anything about paid service officially, but regulators don’t typically make factory visits just for fun.

Tesla has taped out its AI5 chip, targeting roughly 5x the useful compute of the current dual-chip AI4 setup and approximately 2,000-2,500 TOPS total. Volume production is targeted for mid-to-late 2027 at TSMC’s Arizona and Samsung’s Texas facilities.

International News

Waymo has started autonomous testing on public roads in London with a fleet of about 100 Jaguar I-PACE vehicles across a 100-square-mile area of the city, safety operators in seat. The company first announced its London ambitions about six months ago and has since been mapping streets and hiring locally. Commercial service is planned for 2026, pending UK regulatory approval. This will be Waymo’s first right-hand-drive country for general deployment.

Hesai just released the Picasso, what it’s calling the world’s first full-color lidar chip. To understand why this is significant, you need to first understand that standard lidar works like a high-precision black-and-white camera. It captures the shape and distance of everything around a vehicle but has no color information, which is why AVs still need cameras to read traffic lights, lane markings, and signs. The Picasso integrates color perception directly at the sensor level, generating colored 3D point clouds natively without camera stitching. It supports up to 4,320 laser channels at 4K resolution, and Hesai’s next-gen ETX series incorporating the chip is targeting mass production in the second half of 2026.

In short, this could eliminate the need for cameras on AVs entirely.

Alright, that’s it from me… until next week. James and Mike will be back in the next couple of days with additional coverage of the event. If you enjoy this newsletter, share it with your friend, colleague, or boss. Thank you for reading; Sophia out!