Remember when the robotaxi race was supposed to be won in the lab? A few years back, every autonomous vehicle company pitched the same story: whoever cracked the perception stack first, whoever logged the most miles, whoever had the best simulation pipeline would take the market. The winning tool would be the best tool. That was the deal.
Turns out the deal has changed. According to the Financial Times, Waymo has doubled its spending on lobbying in its robotaxi battle with Uber. Not doubled its compute budget. Not doubled its safety testing. Doubled its lobbying. Meanwhile, Uber has pledged $10 billion to win the robotaxi race, and separate FT reporting frames the company as fighting to stay in that race at all.
I review toolkits for a living, and I keep coming back to one uncomfortable question: what does it mean when the most important tool in an AI company’s stack isn’t software?
The Stack Nobody Benchmarks
When I evaluate an AI product, I look at the usual layers. The models. The APIs. The developer experience. The pricing. But the Waymo story is a reminder that there’s a layer underneath all of that, one that never shows up in a benchmark chart: regulatory access.
A robotaxi that can’t get permits doesn’t have a technical problem. It has a political one. And political problems get solved with lobbyists, not pull requests. Waymo doubling its lobbying spend tells you the company has done the math and concluded that the marginal dollar buys more progress in a statehouse than in a data center.
From a reviewer’s perspective, that’s a fascinating and slightly depressing signal. It means the product competition has matured to the point where the technology is table stakes and the real differentiation happens in rooms most of us will never see.
Uber’s $10 Billion Bet
Uber’s side of this is just as telling. A $10 billion pledge to win the robotaxi race is an enormous number, and the FT’s framing that Uber is fighting to stay in the race suggests the company knows it’s playing catch-up. Uber famously sold off its own self-driving unit years ago, and now it’s spending heavily to buy back relevance in a market it once tried to build in-house.
If I were reviewing this as a toolkit decision, I’d call it what it is: a build-versus-buy reversal at massive scale. Uber built, then quit, and is now buying its way back in. That’s not a criticism, exactly. Sometimes the honest answer in a product review is “this team should have used someone else’s tool.” But it does mean Uber’s position in this race depends on partnerships and capital rather than a proprietary technical edge, which makes Waymo’s lobbying push even more pointed. If you can’t outspend a rival on the road, you can try to outmaneuver them in the regulatory arena.
What This Means for the Rest of Us
Most of my readers aren’t running robotaxi fleets. But there’s a lesson here for anyone building or buying AI tools.
- Distribution beats capability more often than we admit. The best model doesn’t win. The model that ships, gets approved, and reaches users wins. Lobbying is just distribution strategy wearing a suit.
- Watch where the money moves. When a company shifts spending from engineering to influence, it’s telling you the technical race is closer than the marketing suggests. Nobody doubles lobbying spend when they’re comfortably ahead on merit alone.
- Regulatory moats are real moats. For AI companies in healthcare, finance, transportation, and other regulated spaces, the permit is the product. Evaluate vendors on their compliance posture, not just their demos.
There’s a broader backdrop here too. The FT reports that America’s biggest companies are posting “rock solid” profits even as consumers face higher costs. Big players have the cash to fight these battles on every front at once: technology, capital, and politics. Smaller AI companies don’t get to double their lobbying budgets, because they don’t have lobbying budgets. That asymmetry shapes which tools survive long enough for people like me to review them.
My Honest Take
I’d love to tell you the robotaxi race will be decided by whichever system drives best. The evidence says otherwise. Waymo is arming up with lobbyists, Uber is arming up with billions, and the actual riding experience is almost a side conversation.
As a reviewer, I can benchmark
🕒 Published: