Article — "Waymo Hits 14 Cities as Zoox Maps Two More—But Uber Split and NHTSA Signal Turbulence": Article — "Waymo Hits 14 Cities as Zoox Maps Two More—But Uber Split and NHTSA Signal Turbulence": On September 1, two robotaxi companies announced they were arriving in San Diego on the same day, and CNBC filed both under one headline: "Waymo and Zoox expand into more U.S. markets as robotaxi race heats up."
Read the two announcements again and the word "race" starts to fall apart.
Waymo's tweet: "We're pulling up in Denver, San Diego & Tampa today! Download the Waymo app to be one of the first to take a ride."
Zoox's tweet: "Our test fleet is heading to both cities later this month to map the streets ahead of autonomous rides."
The verbs are the whole story. Waymo is pulling up with riders. Zoox is heading out to map. One company is collecting fares; the other is collecting HD point clouds. These are not two competitors at the same starting line. They are two different phases of the same industry, flattened into a single narrative because both press releases dropped in the same news cycle.
One company is collecting fares; the other is collecting HD point clouds. These are not two competitors at the same starting line—they are two different phases of the same industry.
Last week we mapped a three-tier deployment reality: Waymo running paid revenue at scale, Zoox running a single paid market, and the rest running demos. This week each tier moved—Waymo added three cities, Zoox entered mapping in two new markets, and the Uber-Waymo alliance reportedly came apart in Phoenix, per an opinion blog. The gap didn't close. It widened, and the reasons it widened tell you who actually has a business.
Start with the numbers CNBC's own copy contains, buried below the "race" framing.
Waymo now provides over 500,000 robotaxi rides in the U.S. every week, across 4,000-plus vehicles in 14 cities, and it's targeting one million rides a week by the end of 2026. In each of the three new cities it launches with "dozens" of robotaxis and plans for "hundreds over time," all hailed through the Waymo app.
Zoox's only paid ride-hailing service, by contrast, is in Las Vegas. Its San Francisco service is free and limited to select riders, with no announced date for paid operation. And when Houston and San Diego "open," they open with human drivers on board, mapping streets—not carrying fares.
So the correct sentence is not "Waymo and Zoox both expanded." It's: Waymo added three revenue markets while Zoox added two pre-revenue markets it won't monetize for months, if the Las Vegas timeline is any guide.
Here is the question Zoox has ducked since Las Vegas and continues to duck in Houston and San Diego: what does a Zoox ride cost to provide? The company has attached no fare and no cost-per-mile to any market. It's now expanding its geographic footprint to a claimed 12 markets without ever answering the one number that separates a transportation service from a demonstration program.
City count is the vanity metric of this industry. It's cheap to enter a city with a mapping fleet and a safety driver. It is extraordinarily expensive to remove the driver, satisfy two separate state agencies, price the ride, and still have unit economics that don't bleed. Waymo's Ojai—the Zeekr-built minivan running its sixth-generation system, explicitly "designed to be cheaper to build, operate, and maintain"—exists because Waymo is playing the second game. Zoox is still announcing the first.
Which brings us to the development the "rivalry footnote" crowd keeps mis-filing: the reported collapse of the Uber-Waymo partnership in Phoenix, per an opinion blog.
The two reportedly still collaborate in Atlanta and Austin, and Uber executives are reportedly already "taking subtle jabs at Waymo," per an opinion blog. The instinct is to read this as a spat. That misses the direction of the leverage.
For years, the strategic logic of an AV company partnering with Uber was distribution. Uber owns the demand-side app, the rider habits, the surge pricing, the marketplace. An AV maker with a great car and no riders needs that pipe.
Waymo reportedly dissolving the Phoenix partnership, per an opinion blog, would be Waymo declaring it no longer needs the pipe—at least not in its most mature market. When you're already dispatching 500,000 rides a week through your own app, with tens of thousands signed up in each new city before you launch, the ride-hail distributor stops being an accelerant and becomes a tax and a dependency.
Walking away from Uber in Phoenix is a confidence move only the market leader can afford. Zoox cannot make that move because Zoox has no comparable demand engine of its own; it has one paid city and a free pilot. The Uber split is not a sign of turbulence for Waymo. It's a sign Waymo believes it has achieved distribution independence, the thing every other player still lacks.
Then there's the regulator, and this is where the asymmetry gets almost cruel.
NHTSA Administrator Jonathan Morrison sent a letter to all AV developers emphasizing that autonomous vehicles must not interfere with first responders and law enforcement—framing it as a "functional insufficiency" the industry must fix. In interviews, Morrison has pushed a broader accountability posture: developers must ensure vehicles are safe and mature, obey road rules, and face enforcement when they don't.
On paper this applies to everyone equally. In practice it lands on Waymo hardest, precisely because Waymo is the only one operating at real scale.
You cannot obstruct a fire truck with a mapping fleet that has a human behind the wheel. First-responder interference is a scale problem—it happens when you have thousands of driverless vehicles moving through live emergencies, like the San Francisco fireworks-show incident that reportedly highlighted challenges managing robotaxis during emergencies. The directive targets exactly the failure mode that only appears once you've done the hard thing and pulled the driver out.
So the leader's reward for scaling first is becoming the primary regulatory target. Waymo's 4,000 vehicles are simultaneously its business moat and its largest liability surface. Zoox, still sending humans out to map Houston's heat and San Diego's coastal fog, is structurally insulated from Morrison's core concern—not because it solved the problem, but because it hasn't earned the problem yet.
That is the chasm the "race heats up" headline conceals.
One company is metabolizing the consequences of scale: distribution independence on the upside, regulatory bullseye on the downside. The other is still in the phase where you get to announce cities without disclosing costs, deploy fleets without removing drivers, and expand your footprint without a regulator singling you out—because none of it is load-bearing yet.
Watch the verbs, and watch the fares. When Zoox tweets that it's "heading to map," and still won't say what a ride costs, it's telling you exactly which phase it's in.
The turbulence in the title—the Uber split, the NHTSA signal—is real. But it's the turbulence a business hits when it has passengers to lose. The company mapping streets ahead of autonomous rides isn't flying through it. It hasn't taken off.