If Tesla Robotaxi miles continue to increase at 20% per week, the math quickly becomes interesting

Support CleanTechnica’s work through a Substack subscription, on Patreon, or on Stripe. Help us produce all the original, high-quality content we publish week after week despite the challenges of AI deleting content, anti-social media, inflation and other obstacles.

Tesla just surpassed Waymo in Austin in weekly driverless miles. If its recent 20% weekly growth holds, it will surpass Waymo nationally in about 11 weeks and will be reaching Uber + Lyft volume in three Texas metro areas by mid-February.

The miles accumulated without supervision are already large enough to begin to answer the question of safety from humans and from Waymo. This has implications for whether LiDAR is actually required.

I noted in my Q3 earnings summary that Tesla’s unattended robotaxi miles increased from 250,000 last quarter to about 1.7 million, and then the same day the company said they were already at 3.3 million. That implies approximately 20% weekly growth in the last period. I want to take that number seriously for a limited period and see where it goes relative to Waymo and to the actual ride market in the Texas cities where Tesla is already operating. Then I will draw some conclusions.

It looks like Tesla has already surpassed Waymo in Austin. Next, let’s look at what it would take for them to overtake Waymo nationally and start eating into Uber and Lyft’s market share.

Mathematics through intersections

Tesla Unattended Weekly Mileage Projection (20% weekly growth compounded from current unmonitored domestic miles/week):

At 558,000 per week, the rate crosses Waymo’s recent national weekly total of ~4 million around week 11. By week 16, the weekly rate would be approximately 10 million miles. That’s the point at which Tesla’s weekly volume would exceed half the current Uber + Lyft volume in Austin, Dallas and Houston combined.

How fast has Waymo really been growing?

Waymo’s recent numbers make the crossover more eye-catching. As of the end of June 2026, the company had logged 271.3 million accumulated passenger-only miles. In the second quarter of 2026, it added only about 50.7 million miles, down from 49.9 million in the first quarter, sequential growth of about 1.6%. Previous quarters of 2025 had seen quarterly growth of 15% or more. City-level data shows the same cooling: Phoenix has recorded two consecutive quarterly declines, Austin’s quarter-over-quarter growth fell from over 100% in early 2025 to around 4% in the second quarter of 2026, Los Angeles was close to 9%, and San Francisco was close to 13%. Atlanta grew faster from a small base, but the big picture is a large installed base whose percentage growth has slowed dramatically. Even if Waymo regains its 15% quarterly growth, that’s dramatically less than the 970% quarterly growth implied by my chart above.

Austin, Dallas and Houston: the realistic short-term market

The Austin metropolitan area has approximately 2.4 to 2.6 million people and about 45 million vehicle miles traveled per day (approximately 315 million per week). Waymo’s recent weekly rate is about 400,000 miles; Tesla’s most recent week in Austin was approximately 438,000 miles. Both are still small fractions of total driving. Dallas-Fort Worth (~5.7 million people, ~157 million daily VMT) and Houston (~5.9 million people, ~144 million daily VMT) are substantially larger. Tesla’s current Dallas footprint is already larger than its initial footprint in Austin (approximately 81 square miles) and has logged more than 600,000 miles accumulated unattended. Houston remains smaller: about 25 square miles and about 226,000 cumulative miles. Waymo is present in both cities with public service. Estimated weekly Uber + Lyft miles (combined paid rider miles only):

• • Austin: 1.5 to 2.5 million

• • Dallas-Fort Worth: 4 to 6 million

• • Total of three cities: approximately 10 to 14 million

Prices change the addressable market

If Tesla’s prices are about half those of Uber, the relevant market is no longer limited to transportation substitution for hail. Personal driving on the three meters is equivalent to hundreds of millions of kilometers per week. Even a 1 to 2% change in trips currently taken privately turns into several million additional weekly miles once the cost per mile falls well below the total cost of owning and operating a personal automobile for many trips. Some off-peak or commute-heavy transit trips (CapMetro, DART, Metro) also become competitive.

Therefore, the practical short- and medium-term opportunity in these three cities alone is greater than the 5-7 million weekly trips portion, potentially 10-20 million once personal travel substitution appears. The 16-week 20% compounding route already reaches ~10 million miles per week. That erases half of the current Uber + Lyft total in the three cities and leaves room for the start of personal ride substitution, provided the Dallas and Houston geofences expand enough and the fleet grows to absorb demand. Maintaining 20% ​​weekly for 16 consecutive weeks while expanding service areas is an operational issue, not a demand issue. Price becomes the lever of demand once vehicles and geofences are large enough.

What could go wrong?

The arithmetic is clean. The path from here is not.

First, the 20% weekly rate is a launch rate that is observed while the absolute numbers are still small and concentrated in one city. Maintaining it for 16 weeks means that the weekly addition has to continue to grow. That requires rapid deployment of vehicles or a sharp increase in utilization per vehicle. Both become more difficult as the fleet expands: Charging infrastructure must keep pace, maintenance capacity must expand, and local regulators must approve larger operating zones. A single high-profile incident causing a pause in any of the three cities would flatten the curve immediately.

Second, geofence expansion is not automatic. Dallas and Houston currently operate in much smaller areas than Austin. Bringing them to Austin-scale coverage involves mapping, local permitting, and public acceptance. Any delay will push the volume absorption problem until 2027.

Third, utilization may decline even as the fleet grows. If vehicles spend more time empty, charging, or waiting for demand that has not yet shifted from personal cars, miles per vehicle decrease and the size of the required fleet increases.

Fourth, pricing it at half the price of Uber assumes Tesla can still cover its costs at that level. If insurance, remote assistance, or cleaning costs are higher than expected, the price cut is reduced or disappears, and the personal travel substitution story weakens.

None of these are fatal. They are the ordinary constraints that turn an early exponential into an S-curve. The 16-week path clarifies the calculations for a material position relative to current transportation in three Texas cities. The open question is whether the vehicles, zones and operations will arrive on that schedule.

my reading

Tesla isn’t primarily trying to overtake Waymo in percentage terms. It’s trying to demonstrate that the unit economics of an unattended Cybercab (or Model Y) can work at scale, with revenue expected to be notable in 2027. The recent 20% weekly growth is the clearest sign yet that the software has surpassed an internal quality bar in the cities where it operates, and the unsupervised miles accumulated already provide a significant basis for comparing safety results with those of human drivers and Waymo. If Tesla’s incident rate continues to hold up as the fleet grows, it undermines the long-standing claim that LiDAR is required to operate safer than humans.

The superior results relative to Waymo would further suggest that more capable software is more important than simply adding sensors to less sophisticated systems. The next two questions are whether the same growth rate can be sustained long enough for the weekly total to surpass Waymo’s, and whether Dallas and Houston can reach Austin-like coverage in the next four months. History suggests that at least one constraint appears before the curve remains perfectly exponential, but arithmetic no longer requires an absurd number of cities or an impossible fleet to become material relative to existing transportation. Any progress they make in Florida (my home state) in Miami, Orlando, Tampa and Jacksonville is a short-term advantage, but they need to continue their growth in the future.

If you want to take advantage of my Tesla referral link to get three months of fully autonomous driving, here’s the link: But if another owner helped you more, use their link instead of mine. Disclosure: I am a shareholder of Tesla and XPeng. I do not offer any investment advice of any kind.