Lobbyists don’t need lidar.
That’s the odd thing about the self-driving car race right now. We spend a lot of time talking about sensors, neural networks, and how many miles a vehicle has driven without a human touching the wheel. Meanwhile, according to the Financial Times, Waymo has doubled what it spends on lobbying as its tensions with Uber deepen. The technology is real. But some of the most consequential moves are happening in meeting rooms, not on roads.
I write about AI agents for people who don’t build them, and a robotaxi is one of the easiest AI agents to understand. It perceives the world, decides what to do, and acts, all without a person approving each step. No chat window, no prompt. Just a car that shows up and drives you somewhere. If you want to understand where autonomous agents are headed, watch this one, because it’s the version of the idea that has to survive contact with real streets, real regulators, and real competitors.
What the money is telling us
Here’s the picture the recent headlines paint, and it’s worth reading them together rather than one at a time.
Waymo has finalised a $16bn funding round at a $110bn valuation. Uber has pledged $10bn to win the robotaxi race. Waymo is exploring a split with Uber. Volkswagen is pushing into robotaxis to challenge Waymo and Tesla. And Waymo has doubled its lobbying spend.
Read as a set, that’s not a technology story. It’s a positioning story. Enormous sums are moving before anyone has clearly won, which usually means the participants believe the outcome is still up for grabs and that being early to the table matters more than being perfect.
Why lobbying is the tell
When a company doubles its lobbying budget, it’s making a bet about what stands between it and its customers. For most software, the answer is nothing much. You ship an app, people download it, done. AI agents that act in the physical world don’t work that way.
A robotaxi needs permission to exist in a given city. It needs rules about who is liable when something goes wrong, what data it has to share, whether it can operate at night or in rain, whether it can drop passengers at an airport curb. Those rules get written by people, and companies that show up to those conversations get a say in the shape of them.
So doubling that spend is a signal that Waymo sees regulation as a live variable, not a settled background condition. It’s also a reminder of something non-technical readers often miss: an AI agent’s real constraint is frequently permission, not capability.
The awkward Uber situation
The Waymo and Uber relationship is the part I find most interesting to explain, because it captures a tension that runs through the whole agent economy.
Uber owns demand. It has the app, the habit, the millions of people who already open it without thinking. Waymo owns the driving. Those two things fit together neatly, which is why they’ve worked together at all. But they also compete for the same prize, because whoever controls the customer relationship captures most of the value.
Uber pledging $10bn to win the robotaxi race, and Waymo exploring a split, suggests both sides have done that math and come to the same uncomfortable conclusion. Partnerships that work well in the short term can become strategically impossible once the technology matures.
You’ll see this pattern again and again with AI agents. The company with the smart system and the company with the users start out as allies and end up as rivals.
Volkswagen changes the shape of the field
Volkswagen entering the space to challenge Waymo and Tesla matters because it adds a different kind of player. Carmakers think in factories and production volumes. Software companies think in fleets and software updates. When both are chasing the same market, the competition stops being a straight comparison and becomes a question of which approach scales first.
What I’d watch if I were you
You don’t need to track every funding round. A few simple questions will tell you most of what you need to know:
- Which cities allow robotaxis to operate without a safety driver, and how quickly that list grows
- Whether you book a ride through a robotaxi company’s own app or through an aggregator like Uber
- Whether new entrants like Volkswagen actually put vehicles on public streets
- How much these companies keep spending on shaping rules rather than shipping features
That last one is the quiet indicator. Technology maturity is often measured in demos. Market maturity gets measured in lobbying budgets, legal filings, and partnership breakups.
The robotaxi race is the clearest look we have at what happens when AI agents leave the screen and start operating in the world. The engineering questions are getting answered. The harder questions are about permission, ownership, and who gets to stand between you and the machine doing the work. Right now, those are the ones the money is chasing.
🕒 Published: