A Waymo stopped dead in the middle of a San Francisco street during Super Bowl week, and nobody knew what to do.
Chris was on his way to the Warriors game with his wife — friend had hooked them up with good seats, the whole thing — and the car just quit. No driver to wave at. No number to call that would mean anything in the next three minutes. Just a vehicle sitting in traffic while the city moved around it the way water moves around a rock.
He told me that story, and then — almost in the same breath — told me about a Sky Sports journalist who had one of the most seamless, camera-ready experiences of their career in the back of a different Waymo that same week. Same company. Same city. Same seven-day window. One story is a product failure. One story is a product advertisement. The interesting question is: what does it mean that both are true at the same time?
That's what this episode is actually about. Not Waymo specifically — though we spent a good chunk of time on it. It's about what happens when technology that's genuinely transformative shows up in an environment that wasn't designed for it, and the city has to absorb the chaos before the technology can prove what it's capable of.
The thing that breaks at scale isn't the car
Here's the number Chris gave me: more than 1,000 Waymo vehicles were operating in the Bay Area during Super Bowl week. That's not a test deployment. That's a fleet. And from what I saw on the ground out there — I was in San Francisco for the full week — the vehicles mostly worked. They did the thing. People got in, looked around for five minutes like their brain was rebooting, and then went right back to their phones the way you do when you're in any other Uber. That's the tell. When something stops feeling miraculous because it works, that's when you know the adoption is real.
But here's what the 1,000-car number doesn't capture: the event-day surge problem is not a car problem. It's a coordination problem. Tens of thousands of people leaving the same geographic area at approximately the same moment, all of them trying to get to hotels, airports, after-parties, dinners — Waymo handles the last-mile movement. It does not handle the bottleneck before the last mile, which is getting out of the stadium zone in the first place.
Chris made the point sharper than I could have: Waymo doesn't surge price the same way human rideshares do. They don't call out sick on game day. They don't get road rage, they don't get gridlock anxiety. What they can't do is unsnarl a city that was built before autonomous vehicles existed as a concept and has not retrofitted itself to use them efficiently.
This is where the conversation got interesting to me. Because I grew up thinking the bottleneck was always the technology. The car, the software, the sensor, the LIDAR system that Waymo uses versus the camera-only approach Tesla takes — whoever solved the technical problem would win. And the technical problem is basically being solved. 127 million miles of fully autonomous driving. 90% reduction in serious injury crashes compared to human drivers. Those numbers are not marketing. Those numbers are a product category arriving.
The bottleneck is the city.
Vegas, and why the city is always the real infrastructure
We talked about host cities for a while, and Chris landed in the same place I had independently been thinking about since New Orleans last year: the events that feel effortless are the ones where the city itself is purpose-built for the volume.
Vegas is the obvious answer, and we both said it. Not because Vegas is particularly interesting or culturally rich — it's designed to be neither. Vegas is interesting because its entire physical and logistical infrastructure was built under one assumption: a lot of people are going to be here, they're going to want a lot of things at once, and they need to be able to get from point A to point B without the whole system locking up. Hotels per square mile. Venues per capita. Road layout. Every decision Las Vegas made in its construction was implicitly a decision about throughput.
Compare that to Indianapolis hosting All-Star Weekend, which Chris mentioned — I'll let him tell the story, but the punchline is the Courtyard Marriott was a top hotel option. When the best accommodation available to someone who flew in from across the country is a business-trip hotel off an interstate, the city is communicating something about what it was built to absorb.
New Orleans last year was great — I said it on the show and I'll say it here. Walkable. Tight geography. One main corridor that everything fed into. But New Orleans works because the scale matches the infrastructure. The Superdome to Bourbon Street is a manageable distance. The city was never designed for 100,000 people; it was designed for the density it already has, and major events fit inside that density reasonably well.
The World Cup this summer in New York and New Jersey is going to be the real test. Because that's a city — and a metro area — where the regulatory pathways for autonomous vehicles don't fully exist yet. The infrastructure question isn't just physical. It's legislative. New York will need special permitting for commercial driverless services at scale. That permitting process is a city making a decision about whether it's ready to be redesigned — not by engineers, but by bureaucrats moving at bureaucratic speed.
The private jet version of the same problem
We went somewhere unexpected after that. I asked Chris about private aviation — specifically, the Super Bowl being the single biggest private aviation weekend of the year, with over 500 planes departing the Bay Area right after the game. And he told me he was on United flight 1249 from SFO to Houston, which is the most honest answer a guest has given me in a while.
The private jet conversation sounds like a different topic. It's the same one.
The reason people fly private to major events isn't primarily status — or it shouldn't be, anyway. It's exit infrastructure. When 500 jets depart in a two-hour window, the people on them are solving the same problem Waymo is trying to solve: they need to leave a high-density geographic area at a high-demand moment without sitting in a system that wasn't built for simultaneous departure at scale.
Chris's dad had a line for this that I've been repeating since we recorded: if you've got more time than money, go that way. If you've got more money than time, go that way. Clean framework. The problem is that framework assumes the infrastructure exists to serve both ends of that tradeoff. Private aviation scales up cleanly for the people at the top. Mass transit scales down cleanly for the people at the bottom. The middle — which is most people — is stuck in the Uber surge and the stadium parking lot and the street where the Waymo just stopped.
I've made decisions at that middle level more times than I'd like to admit. I was at the Toronto All-Star game — the one Chris remembered as the cold one, and he's not wrong, I think that's the year I borrowed a mentor's plane for the hop from Detroit — and even a small aircraft on a short route changes the calculation entirely. Not because of the luxury. Because of the time. The event ends, and you're in the next city before the highway has cleared.
But here's the thing: that solution is not scalable. And the question worth asking is what does scale look like, and who's building it?
Waymo is an answer to that question. 126 billion dollar valuation. $16 billion raised. By end of 2026, they're projecting over a million rides a week, expanding from the current 400,000. Tokyo and London are the first international markets. That's not a startup story anymore — that's infrastructure deployment.
PULL QUOTE: "They don't call out sick on game day." — Chris, on the real advantage autonomous vehicles have over human rideshare drivers
The regulatory layer nobody talks about
The competitors are real. Zoox — Amazon-owned, no steering wheel, bi-directional design, free rides in parts of San Francisco to prove the model, starting to charge this year. Apollo Go in China, quarter-million driverless rides per week already, now partnered with Lyft for UK and Germany deployment. Tesla with the Cybercab, camera-only, no LIDAR, Elon's admitted the timeline is long. Uber building their own robotaxi operation with Lucid Motors and Nuro, targeting 100,000 vehicles by 2027.
Every one of those companies is working on the car. The car problem is being solved — by multiple teams, simultaneously, with billions behind each effort.
Nobody is talking about the regulatory layer the way it deserves to be discussed.
Chris made the point almost in passing: New York doesn't have clear pathways for fully driverless commercial services. So the World Cup — one of the highest-density, highest-demand sporting events on the planet, arriving in the New York metro area this summer — is going to happen in a city that hasn't decided yet how to let the technology operate inside it.
That's not a technology failure. That's a governance failure. And governance failures move at a fundamentally different speed than engineering failures. You can fix a Waymo that stops in the street with a software update. You cannot fix a regulatory framework that wasn't designed for autonomous vehicles with a software update.
Three things I'm watching after this conversation
- Which cities are building regulatory infrastructure ahead of the events, not in response to them. Phoenix has a Waymo manufacturing facility coming. They've been permissive on autonomous vehicle testing for years — not by accident. That's a city that made a decision about what industry it wanted to attract and built the governance structure to attract it. The cities that figure out how to permit and integrate autonomous vehicles before the next Super Bowl or World Cup bid are going to have a real advantage in the competition to host these events. I'm watching which mayors are actually paying attention to this.
- What happens when hardware diversifies for different environments. Right now the conversation is city-by-city, good-weather markets first. Phoenix, Miami, LA — places where the variables are manageable. The hardware question for extreme weather, dense urban environments like New York, elevated terrain — that's where the next phase of competition happens. The company that cracks New York in January has cracked the hardest environment in American cities. That's not 2026. But it's not 2040 either.
- Whether the private aviation model gets disrupted from below or above. JSX — JetSuiteX, which is what Chris was actually on — is the interesting middle layer. Not full charter, not commercial. Semi-private, faster boarding, decent experience, fraction of the charter cost. That model exists because the gap between commercial and private is wide enough to build a business in. Autonomous vehicles are trying to do the same thing for ground transportation. The question is whether those middle layers scale, or whether we end up back at a barbell: Waymo for most people, NetJets for the rest, and nothing clean in between.
The stalled Waymo in San Francisco was not a story about a bad product. It was a story about a transformative product meeting infrastructure — physical, regulatory, logistical — that hasn't caught up yet.
The car figured out how to drive itself. The city hasn't figured out how to let it.