Waymo’s cars can drive themselves. Waymo’s business, apparently, cannot. The company that has spent years telling us its software no longer needs a human at the wheel just doubled its lobbying budget, because the thing standing between it and a national robotaxi rollout isn’t perception stacks or edge cases in the rain. It’s people in Washington who haven’t decided what the rules are yet.
According to federal lobbying disclosures, Alphabet-owned Waymo sharply increased what it spends persuading US regulators to clear a path for fully autonomous taxi service. It’s pushing for a federal framework that would smooth the rollout. And it’s doing this while trading elbows with Uber, a company that knows a thing or two about arriving in a city before the city has figured out whether it’s allowed to.
What this actually tells you about the technology
I review AI tools for a living, and the pattern here is one I see constantly at a smaller scale. A product hits the point where the demo works, the metrics look good, and the remaining blockers are all organizational. Permissions. Compliance. Someone in legal who wants to know what happens when it’s wrong. The engineering stops being the hard part and the paperwork becomes the roadmap.
Waymo doubling down on lobbying is a signal that it believes its own driving problem is solved enough to bet money on regulatory timelines instead. That’s a genuinely interesting read. You don’t spend heavily on federal advocacy for permission to deploy something you’re not confident deploying. You spend it when the queue is the constraint.
The less flattering read is also available: a patchwork of state and city rules is expensive and slow to navigate one jurisdiction at a time, and a federal framework is a shortcut past fifty separate negotiations. Both readings can be true. Companies don’t lobby out of principle, they lobby out of unit economics.
The Uber dynamic is the part worth watching
This isn’t Waymo against regulators. It’s Waymo against Uber, with regulators as the terrain. Streetsblog reported that Uber and Waymo combined spent more than $15 million lobbying New York politicians this year. That’s two companies fighting over the rules of a market that barely exists yet at scale, in a single state.
The two have fundamentally different shapes. Waymo builds the driver. Uber owns the demand and the app habit of hundreds of millions of people. Whoever gets the rules written closest to their own architecture wins a structural advantage that no amount of later engineering fixes. If the framework assumes a fleet operator with its own vehicles and its own autonomy stack, that favors one model. If it assumes a marketplace routing passengers to whatever autonomous supply is available, that favors the other.
Rule-writing is product design by other means. That’s why the money moved.
Why AI people should care about this specific fight
Robotaxis are the most physical, most legible test case for a question every serious AI deployment eventually hits: who is accountable when an autonomous system acts and something goes wrong? Chatbots get to be vague about it. A two-ton vehicle at an intersection does not.
Whatever framework comes out of this lobbying push will set precedent for how the US handles autonomous systems that take consequential actions without a human in the loop. Liability. Incident reporting. What counts as adequate testing. Who gets to see the data. Those questions will echo into agent deployments that have nothing to do with cars, because regulators reuse the mental models they already have.
So the interesting number isn’t how much Waymo spent. It’s what the resulting framework demands. A federal standard that requires real safety reporting and independent scrutiny would be good for everyone, including Waymo, whose main competitive asset is a safety record. A federal standard that mostly preempts local objections and calls it a day is a different thing wearing the same label.
My honest take
I’m not going to pretend increased lobbying is sinister. Every industry that touches public infrastructure does this, and a national patchwork genuinely is a bad way to regulate vehicles that cross state lines. Asking for one set of rules is a reasonable ask.
What I’d push back on is the framing that this is a story about progress. It’s a story about positioning. The doubled budget doesn’t make the cars better, safer, or more available. It makes Waymo faster than Uber at getting permission, or slower, depending on who spends more effectively. That’s a competitive contest, not a technical one, and the outcome will shape which robotaxi you can actually hail in three years more than any model update will.
Keep an eye on the disclosures rather than the demos. In this phase of autonomy, the filings tell you more about the roadmap than the keynotes do.
🕒 Published: