Tesla's Robotaxi Just Went Unsupervised in Three Texas Cities. The Price War With Waymo Is Already Over.
Tesla launched fully driverless rides in Dallas and Houston, undercutting Waymo by 56% on price, with Cybercab mass production starting this month and seven cities targeted by June.
Tesla's Robotaxi Just Went Unsupervised in Three Texas Cities. The Price War With Waymo Is Already Over.
Tesla launched fully driverless rides in Dallas and Houston this weekend, undercutting Waymo by 56% on price, and the Cybercab hasn't even entered the fleet yet.
The driver's seat is empty. The ride costs $6.15. And the company doing it just added two cities in a single weekend.
What Happened
On April 18, Tesla activated fully unsupervised robotaxi service in Dallas and Houston. No safety driver. No human monitor. Riders hail a Model Y Juniper through the Tesla app, get in, and ride through a 25-square-mile geofence in each city.
This is not a pilot. It is a commercial service, open to the public.
Tesla already had a supervised program running in Austin since June 2025 with roughly 35 vehicles. That fleet grew to about 500 Model Y units across Austin and the Bay Area by January 2026, logging over 550,000 autonomous miles. The Dallas-Houston launch skipped the supervised phase entirely. It went straight to no-driver.
The timing is not accidental. Tesla's Q1 2026 earnings call is April 22. Walking into that call with three active robotaxi cities and a price advantage over Waymo is the kind of operational leverage Wall Street pays attention to.
Why It Matters
This is the first time Tesla has put fully unsupervised vehicles in head-to-head competition with Waymo in the same cities. Waymo launched in Dallas and Houston in February 2026 as part of its 10-city expansion. Tesla is now operating in three of those markets.
The price difference is the headline. A 2.25-mile, 7-minute ride in Dallas: Tesla charged $6.15. Waymo asked $13.93. That is 56% cheaper for the same trip, based on user-reported pricing shared on X by Sawyer Merritt.
Tesla's per-mile cost is running $2 to $3 across the new cities, based on early rider data. If those economics hold at scale, the unit economics of camera-only autonomy vs. LiDAR-loaded vehicles become very hard for Waymo to match.
And the Cybercab, Tesla's purpose-built two-seat robotaxi with no steering wheel and no pedals, is expected to enter mass production at Giga Texas later this month. Drone footage from April 13 showed over 50 Cybercab units on the Giga Texas campus near the crash testing facility. When those hit the fleet, the cost curve drops again.
The Consensus Take
Wall Street is watching but hedging. The bull case, articulated by Wedbush's Dan Ives, frames robotaxi commercialization as the unlock for a "$3 trillion AI chapter." TSLA surged 12% over four sessions leading into the launch, closing at $400.62 on April 17 with a market cap of roughly $1.28 trillion.
The bear case focuses on the geofence. Twenty-five square miles is small. Only 12% of the Austin fleet operates unsupervised. The service saw disruptions in Dallas-Houston within 24 hours, suggesting capacity constraints. And Waymo has nearly a decade head start in operational complexity across 10 cities.
Both sides have real points. But they are arguing about different things.
The Mr. X Take
The bears are measuring the geofence. They should be measuring the rate of change.
Tesla went from 35 supervised cars in one city to 500+ vehicles, three cities, and unsupervised operation in under a year. The seven-city target for June 2026 (adding Phoenix, Miami, Orlando, Tampa, and Las Vegas) is aggressive by anyone's standard. Waymo took nearly a decade to reach 10 cities. Tesla is attempting to match two-thirds of that footprint in months.
The real comparison is not "who has more cities today." It is "whose architecture scales cheaper." Tesla runs cameras and silicon. Waymo runs LiDAR, cameras, radar, and significantly more expensive per-vehicle hardware. At $6.15 versus $13.93 for the same trip, the pricing tells you whose cost structure has more room to compress.
The Engineering Layer
Tesla's camera-only approach was controversial for years. No LiDAR. No HD maps. Just eight cameras, a neural network trained on billions of miles of fleet data, and the AI4 inference chip running FSD.
The bet was always that compute and data would eventually beat expensive sensor arrays. That bet is starting to pay visible dividends. The Model Y Juniper running today's FSD stack handles unsupervised operation in urban Texas environments without the hardware overhead that makes Waymo vehicles cost a reported $150,000+ each.
The Cybercab takes this further. Built from scratch for autonomy, it strips out the steering column, pedals, and driver-side controls. Less hardware. Lower manufacturing cost. Tesla has said the consumer price will be under $30,000. When deployed as fleet vehicles at scale, the per-unit economics get very interesting very fast.
The Market Layer
TSLA at $400.62 prices in optimism. The 12% four-session run on 113 million shares (April 15 alone) signals institutional positioning ahead of earnings. At a $1.28 trillion market cap, the stock already reflects a company worth more than every other automaker combined.
But here is the math that matters: if Tesla reaches 2,500 active robotaxis nationwide (the reported current fleet size) running at $2-3 per mile with high utilization, the revenue per vehicle starts to compound. And the fleet can grow with every Model Y and Cybercab that rolls off the line.
Ives is not wrong to call this a potential $3 trillion chapter. The question is execution speed. The earnings call on April 22 will tell us how aggressive the ramp plan really is.
The Meme Layer
The internet has already made the joke. A Dallas rider posts a $6.15 receipt next to a $13.93 Waymo quote and captions it: "The driverless car price war lasted about 48 hours."
It is funny because the engineering is real. Tesla spent years getting mocked for "vision only." Now the vision-only car is doing the same job for half the price in the same city. The meme writes itself.
Risks
The geofence is small. Twenty-five square miles per city is not a citywide service. Scaling the operational domain involves edge cases that multiply with geography, weather, and traffic density.
Safety data is still early. Tesla's unsupervised track record is measured in months, not years. Any serious incident will draw intense regulatory and media scrutiny, especially in a state with lighter oversight.
Capacity constraints appeared within 24 hours of the Dallas-Houston launch. Scaling a fleet is a logistics problem, not just a software problem. Vehicles need charging, cleaning, maintenance, and repositioning.
Waymo is not standing still. It operates in more cities with a longer safety record. Its hardware approach may be more expensive, but it has a deeper operational playbook.
Valuation risk is real. At $1.28 trillion, any stumble in the robotaxi narrative could trigger a sharp repricing. The stock is priced for execution, not patience.
Bottom Line
Tesla put unsupervised robotaxis in three Texas cities, undercut the only real competitor by 56% on price, and did it with a camera-only system running on consumer-grade hardware. The Cybercab starts mass production this month. Seven cities are targeted by June.
The bears keep asking when the geofence expands. The factory already answered. It is building the vehicle that makes the geofence irrelevant.
This is commentary, not financial advice. Verify primary sources and official filings before making investment decisions.
Mr. X is an ultra-concentrated futurist investor obsessed with Tesla and the broader mission-driven technology ecosystem. He believes the market consistently underestimates the pace of disruptive innovation and is willing to withstand extreme volatility for asymmetric upside. He tracks product roadmaps, manufacturing ramps, and regulatory catalysts with obsessive detail.
View all articles →