Approval cap of 5,000 vehicles, but only about 20 are actually on the road.
On August 20, the Nevada Department of Transportation unanimously approved Tesla to operate up to 5,000 autonomous vehicles in Clark County, lifting the temporary cap of 10 vehicles set just weeks prior. Tesla stated that paid operations could commence in about 30 days after completing vehicle inspections and fare approvals. The stock price rose 5.14% on August 21 following the news.
However, approval does not equate to actual operational capacity. Third-party tracking data indicates that by mid-2026, Tesla was operating approximately 20 unsupervised Robotaxis across various US cities, down from a peak of around 25 in late April. Dallas and Houston each still only have a small number of vehicles. Even if Las Vegas immediately adds 10 vehicles, it would increase the nationally tracked fleet size by about 50%, highlighting that the business remains in its very early stages.
The real bottleneck is not licensing, but reliability and Cybercab production.
Musk directly linked the expansion of operational scale to reliability, describing the improvement process as a "march of nines," with an ideal target of reaching 99.999999% reliability. He stated that the real limitation on Robotaxi growth is only this point. Therefore, before the software reliability meets the requirements, the 5,000-vehicle license cap will not automatically translate into a fleet of 5,000 operating vehicles.
The Cybercab introduces a second constraint. As it uses a new chassis, Tesla must first accumulate real-world driving data for this model, and early vehicles must be equipped with steering wheels and pedals for calibration. Musk said that once sufficient data is gathered, the fleet size will "increase dramatically." However, the Cybercab only began production in the second quarter, and there is still a ramp-up process before large-scale deployment.
With a forward P/E ratio of 181, only autonomous driving can explain it.
In Q2, Tesla's revenue reached a record $28.24 billion, up 25.5% year-over-year, beating market expectations by 6.6%. Energy storage deployments grew 53% quarter-over-quarter to 13.5 GWh, and global paid FSD users approached 1.5 million. However, the operating margin fell from 4.1% a year ago to 1.4%. EBIT of $398 million was more than 70% below expectations. Adjusted earnings per share were $0.33, missing estimates by about 39%. Free cash flow was negative $1.09 billion, with 2026 capital expenditure guidance exceeding $25 billion.
Tesla currently trades around $345.82, with a forward P/E of 181.67, a forward EV/EBITDA of 77.38, and a forward EV/revenue of 12.04. In comparison, the median for peers is only 0.66 times. Such a premium cannot be sustained by the traditional auto business; it essentially prices an option on autonomous driving and robotics.
The article offers a straightforward verification method: Don't just look at the approval cap; look at the actual number of unsupervised vehicles. If the national fleet remains near 20 vehicles by the Q3 earnings report in late October, the Nevada approval will look more like a publicity move. If the vehicle count starts compounding after the launch of FSD v15, then the autonomous driving story will truly be worthy of the current valuation.
