Remount RoundupIssue 9
The Unglamorous Half
A robotaxi fleet got towed out of a traffic jam with dead batteries, and our partner's CEO named the missing layer on the record. The interesting part of autonomy isn't the driving anymore. It's the ground underneath it.
For most of this year, when someone asked what I was working on, I could point at the cars. This week I mostly pointed at parking lots.
Fifteen spaces, decent lighting, a gate code, and chargers within a couple of miles. That is what an autonomous delivery pilot actually runs on. My son and I spent the week mapping lots and Level 2 chargers in North Austin for our fall pilot, and somewhere between the second charging map and the first insurance question, the shape of this industry got very clear to me.
The intelligence is arriving on schedule. Tesla just pushed a distilled version of its v14 self-driving stack to the older cars in its fleet, millions of vehicles everyone had written off as frozen in 2023. People are livestreaming coast-to-coast runs on it right now. Still supervised, still imperfect, but the direction only points one way.
Now look at what happened on the ground this same stretch. On the Fourth of July, San Francisco gridlocked after the fireworks, and a number of Waymos sat in that traffic until their batteries ran out. Tow trucks hauled robotaxis out of a traffic jam. The driving software was not the failure. Nobody’s job was making sure the machines could reach a charger.
Same week, Uber and Waymo, partners in two cities, started lobbying against each other over a robotaxi bill in Washington DC. Allies fight like that when a market stops being theoretical.
And Ben Seidl, CEO of our delivery partner Autolane, said the quiet part on the record in this week’s Ride AI interview: “Vehicle financing, insurance, fleet management, and charging: those are distinct problems that need distinct partners.”
Read that list again. Financing. Insurance. Fleet management. Charging. None of it is glamorous. All of it decides whether a machine earns money or dies in a parking lot. The companies building the intelligence do not want that work. The companies building the vehicles do not want it either. That partner-shaped hole in the middle of the industry is exactly where Remount builds, the layer that finances, insures, charges, and runs the machines for the people who own them.
The humanoids tell the same story from the other end. Figure’s first production robots came off an eleven-month run on a BMW assembly line late last year, thirty thousand cars built with humanoid hands in the loop. Real shifts, not demos. Every machine like that needs the same boring layer underneath it before it earns a dollar for anyone.
The intelligence gets the headlines. The ground gets the margins. I know which half we want.
More below on what moved this week.
- Jonas
The week’s evidence
Our partner’s CEO named the missing layer. Ben Seidl founded Autolane in 2024 on the idea that autonomous commerce needs an orchestration layer at the curb. In this week’s Ride AI interview he drew the map plainly: Autolane builds the network connecting demand to supply, while “vehicle financing, insurance, fleet management, and charging” need dedicated partners. He also sketched the endgame, one vehicle running deliveries and rides at the same time, pushing the marginal fare toward “a dollar for ten miles.”
Ground Truth vol.3 | Building the “Air Traffic Control” for AVs with Autolane (Ride AI)
Waymo had a week of growing pains. Days after announcing four new cities and a million-rides-a-week target, the other side of scale showed up: Waymos towed out of San Francisco’s July 4 gridlock after running out of charge, a city supervisor preparing an inquiry, and partner Uber lobbying against a DC robotaxi bill it says would hand Waymo a monopoly (their Phoenix partnership just ended; Atlanta and Austin continue). None of that is a driving-software story. All of it is operations and politics, the layer nobody prices in.
Tesla pushed v14 autonomy backward into the old fleet. FSD v14 Lite, distilled from the AI4 stack, is rolling out to Hardware-3 cars: parking, reversing, and smoother behavior on hardware everyone had written off. Early reports call it a big step past v12 (and more timid than the AI4 version). Still supervised. But the fleet capable of near-autonomous driving just got millions of vehicles bigger, and the community is already running coast-to-coast attempts on it.
https://electrek.co/2026/06/29/tesla-fsd-v14-lite-hw3-rollout/
If the driving is solving itself, the money moves to whoever runs the ground underneath: the depots, the chargers, the insurance, the title. That’s the half we’re building. If you want to see what one earning machine looks like on paper, the modeler at remount.co runs the numbers live. Reply anytime. I read every one.
You’re getting this because you signed up at remount.co. This is a thought experiment, not financial advice and not an offer of securities. Reply anytime or write hello@remount.co.