Signals Inbox·July 28, 2026·Autonomous Systems

When will Tesla’s Cybercab be everywhere?

Tesla’s Cybercab could become common across major US cities around 2029 to 2031, but broad international availability is more likely between 2032 and 2035—and only if today’s tiny fleets turn into dense, reliable driverless networks.

Q1Is Tesla’s Cybercab already giving public rides today?

Cybercab itself is still waiting for its first public passenger.

Tesla’s current autonomous ride service uses Model Y vehicles. The company’s live Robotaxi page lists Austin, Dallas, Houston, Miami, Orlando and Tampa, while describing Cybercab rides as something that will arrive later.

The San Francisco Bay Area creates some confusion. Tesla offers rides there under the Robotaxi name, including trips to San Francisco International Airport, but a human driver remains behind the wheel. California regulators classify it as a chauffeur service rather than an autonomous passenger operation.

Model Y gives Tesla a practical way to launch the software before Cybercab production reaches scale. The vehicle is readily available, has manual controls for testing and is already familiar to Tesla’s service teams. Cybercab should gradually replace it once the purpose-built vehicle is approved and reliable enough for public use.

Tesla Robotaxi exists today. Cybercab does not yet exist as a public transport service.

Q2What does “Cybercab everywhere” really mean?

For this question, Cybercab is “everywhere” only when ordinary riders can reliably find one across most large cities.

A company can launch a city with three cars, publish a large service map and technically claim that the market is live. Most residents would still struggle to get a ride.

Useful availability requires enough vehicles to produce reasonable waiting times throughout the day. The service also needs broad operating zones, highway access, airport pickups, charging sites, cleaning teams and support when a car gets stuck.

We consider Cybercab common when it works reliably across most major US metropolitan areas. We reserve “everywhere” for broad US coverage plus a meaningful presence in Europe, Asia and other large international markets.

Cybercab rollout stages

Q3Why does Cybercab feel much closer now?

Cybercab feels closer because Tesla finally has both a working ride service and physical vehicles coming off a pilot production line.

At the 2024 unveiling, Tesla demonstrated around twenty Cybercab prototypes in a controlled environment. It had no public driverless network, no production line and no commercial mileage to show.

Now the company can point to autonomous Model Y rides in multiple cities. Its latest quarterly update said paid Robotaxi mileage had almost doubled from the previous quarter. Tesla also classified Cybercab as being in pilot production and said volume production should follow.

Those are concrete advances. Cybercab has moved beyond a design exercise, and Tesla has shown that it can run a paid autonomous service.

The two programmes still have not come together. Current passengers receive Model Ys, while production Cybercabs remain outside the public network. The milestone that counts is a sustained fleet of Cybercabs carrying paying riders without onboard supervision.

Q4How large is Tesla’s Robotaxi network currently?

Tesla has reached six autonomous markets, but the fleet behind those maps still looks tiny.

Tesla does not publish active vehicle counts, completed weekly rides, average waiting times or mileage by city. Its statement that paid mileage nearly doubled is encouraging, although the company did not disclose the starting figure.

Independent vehicle trackers cited in recent reporting estimated roughly sixteen active vehicles in Austin, seven in Dallas and three in Houston. Tesla has not revealed the number of cars assigned to Miami, Orlando or Tampa.

Those estimates may miss vehicles, but the order of magnitude is clear. Tesla’s public autonomous fleet appears to number in the tens rather than the hundreds or thousands.

That is enough to test different road environments and build operating experience. It is far too small to serve more than a sliver of local ride-hailing demand.

A credible transport network needs fleet density. Ten cars spread across a large city can produce a service map. Hundreds are required before many residents can depend on it.

Q5Is Tesla expanding a real service or mainly expanding its maps?

Right now, Tesla is widening its service maps faster than it is building visible fleet density.

Austin shows the problem clearly. Tesla expanded its unsupervised operating zone from a small launch area to roughly 245 square miles, including more suburban roads and highway driving. Yet independent trackers have recently counted only around sixteen to twenty active vehicles.

A larger geofence gives Tesla more varied driving data. It exposes the system to highways, unfamiliar junctions, construction zones and longer trips. From an engineering perspective, that is useful.

For riders, the number of available cars matters more than the shape of the map. A huge operating zone served by a few vehicles can mean long waits, distant pickups and unpredictable availability.

Tesla is beginning to build the physical network behind the service. A San Antonio zoning panel recently approved a proposed charging site with capacity for up to 56 autonomous vehicles. Similar depots will be needed for charging, cleaning, inspections and repairs.

The facility is useful preparation, but it is not an active fleet. Tesla will look genuinely scaled when vehicle counts start growing as quickly as the maps.

Q6Do Tesla’s 12 billion FSD miles prove Cybercab is ready?

Tesla’s 12.1 billion supervised FSD miles are a serious advantage, but they do not answer the driverless safety question.

Tesla’s live safety page reports about 12.1 billion miles driven with Full Self-Driving (Supervised), including roughly 4.6 billion city miles. Few autonomous-driving companies have access to such a broad stream of real-world camera footage.

The data covers unusual intersections, temporary roadworks, pedestrians, emergency vehicles, bad lane markings and countless driving mistakes made by other road users. Tesla can use those examples to train its models and search for rare situations.

Every supervised mile still has a licensed driver responsible for the vehicle. That person can brake, steer or take control when the software behaves badly. A mistake can disappear into the intervention history without becoming a collision.

Cybercab has no steering wheel or pedals. It has to recognise when it is confused, stop safely, deal with passengers and handle emergencies without someone sitting in the driver’s seat.

Tesla’s supervised fleet is a major training advantage. Proof of Cybercab readiness has to come from unsupervised mileage, intervention rates and safety outcomes from the commercial Robotaxi fleet.

Q7Has Tesla solved driverless operation?

Tesla has proved that its system can run driverless trips in selected areas. Mass-market reliability is still unproven.

The latest federal filings bring Tesla’s disclosed automated-driving incidents to 22 unique cases. Many were minor, and several involved another vehicle hitting a stationary or slow-moving Tesla. The raw total cannot show whether Tesla is safer or more dangerous without the number of driverless miles completed.

Some incidents expose problems that fleet totals would hide. Tesla vehicles have driven into chains, poles, curbs and construction objects at low speed. More recently, a Dallas vehicle entered a parking-lot entrance and hit a thin metal chain, repeating a similar Austin incident from several months earlier.

Repeated mistakes deserve more attention than isolated ones. A system can encounter an unusual chain once. Hitting the same kind of obstacle again suggests the earlier event did not fully solve the weakness.

Tesla also withholds several figures needed for a proper safety assessment. We do not know how often cars stop unexpectedly, request help, alter their pickup point or fail to complete a ride.

The technology clearly works much of the time. Tesla has not released enough data to show how performance changes when the fleet grows, the roads become harder and human support is spread across more vehicles.

Q8Can remote operators support thousands of Cybercabs?

Remote support can scale only if a human is needed rarely and gives guidance more often than direct control.

Robotaxis will occasionally meet situations that software cannot easily resolve. A police officer may wave traffic through a red light. A construction worker may create a temporary lane. A delivery truck may block the only legal route.

A remote employee can look at the scene and provide context. One person could support dozens of cars when requests are brief and infrequent.

Directly driving a vehicle from another location is harder. The operator sees the road through cameras, relies on a network connection and cannot physically feel how the car is moving.

Tesla’s latest federal reports describe three collisions involving remote control. In each case, the autonomous system became stuck, a remote operator took over and the operator hit an object at low speed. The newest case involved a Houston vehicle driven into a hidden tree stump while being recovered from a dead-end road.

The damage was minor, but the recurring sequence deserves attention. Tesla does not disclose how often remote staff intervene, so we cannot calculate whether three crashes came from dozens of interventions or thousands.

Waymo takes a more cautious approach. Its remote teams generally provide information while the car remains responsible for driving. Tesla will need a similar model, or clear evidence that direct remote driving is extremely rare.

Q9Is Cybercab production now Tesla’s biggest bottleneck?

Today, autonomy and operations are tighter constraints than Cybercab manufacturing.

Tesla still describes Cybercab as being in pilot production. The company has not published a weekly output rate, installed annual capacity or the number of completed vehicles ready for commercial service.

Yet Tesla already knows how to manufacture cars at scale. It produced more than 450,000 vehicles in its latest quarter across its existing range. That industrial base gives Tesla an advantage over robotaxi companies that must rely on outside manufacturers.

Even a modest Cybercab line could initially outrun the service. Production of 1,000 vehicles per week would add 52,000 cars in a year, far beyond the number Tesla could deploy immediately under its current operating model.

The early challenge is finding productive work for the first few thousand vehicles. Each city needs permits, charging sites, cleaning capacity, maintenance teams, remote support and enough customer demand inside the approved zone.

Manufacturing becomes the critical bottleneck once Tesla can activate hundreds of vehicles in each new market. We are not there yet.

Tesla’s factory experience makes a large Cybercab fleet plausible. The company first needs an autonomous network capable of absorbing the cars.

Q10Will Tesla’s camera-only approach make Cybercab easier to scale?

Tesla’s camera-only design could cut cost and speed manufacturing, although city launches will still need local validation.

Cybercab avoids the large lidar units and elaborate external sensor arrays used by several competitors. Cameras are compact, inexpensive and already produced in automotive volumes. A simpler hardware package should reduce the purchase price and make damaged vehicles easier to repair.

Tesla also wants one general driving model to work across many locations. Successful software improvements could then move quickly from Austin to Dallas, Miami and later markets without rebuilding the entire system city by city.

Camera performance can weaken in darkness, glare, heavy rain, fog or when lenses become dirty. A current federal investigation is examining Tesla’s handling of reduced-visibility conditions and the system used to detect when its cameras have degraded.

Tesla’s own rollout shows that local testing remains necessary. The company still uses geofences, expands roads gradually and prepares each new city before opening rides.

The camera-only approach may eventually make Cybercab cheaper and easier to manufacture than rival vehicles. So far, it has not removed the need to validate weather, road design and local driving behaviour market by market.

Q11Why are Texas and Florida easier for Tesla than California?

Tesla chose the easiest large US markets first, and California shows how much regulation can slow the same product.

All six autonomous markets currently advertised by Tesla are in Texas and Florida. Both states have relatively friendly automated-vehicle rules, large ride-hailing markets and little exposure to severe winter weather.

Texas largely prevents cities from creating separate and contradictory autonomous-driving regimes. Tesla can move from Austin to Dallas and Houston under a broadly consistent state framework.

Florida also allows vehicles designed to operate without a human driver. Miami, Orlando and Tampa still present difficult traffic and heavy rain, but the legal path is easier than in many other states.

California separates safety-driver testing, driverless testing and commercial deployment. The state’s current permit list gives Tesla Robotaxi LLC permission to test with a human safety driver. Tesla does not appear among companies authorised for driverless testing or deployment.

That is why Tesla can offer Bay Area rides and airport trips with an employee driving, while Waymo can carry passengers there without anyone behind the wheel.

California will be a much stronger test than another launch in a permissive state. Approval there would show that Tesla can satisfy one of the country’s most demanding autonomous-vehicle regulators.

Q12Will regulation delay Cybercab outside the United States?

International approval will stretch Cybercab’s rollout by years, even if the software improves quickly.

Governments are slowly creating clearer autonomous-driving rules. A new global technical regulation for automated-driving systems was recently approved through the United Nations vehicle-regulation process. It establishes common principles for safety management, testing and post-deployment monitoring.

The global framework should reduce some duplication, but it does not give Tesla permission to launch a taxi service in every participating country. National vehicle approvals, commercial transport licences and local operating rules still apply.

The United Kingdom is building a permitting system for automated passenger services under its Automated Vehicles Act. The European Union has a framework for approving fully driverless vehicles, while work continues on harmonising testing and deployment between member states.

China is moving cautiously. It has conditionally approved its first Level 3 production vehicles for limited roads and is developing stricter Level 4 requirements. Tesla would also face strong domestic robotaxi competitors and local rules covering mapping, data and fleet operation.

What regulation currently means for Cybercab

Q13Can Cybercab make robotaxi rides much cheaper?

Cybercab can probably beat today’s premium robotaxis on vehicle cost, but Tesla has yet to prove the full cost per paid mile.

Tesla designed Cybercab around a price below $30,000 and previously discussed an operating cost near $0.20 per mile. A small two-seat electric vehicle without manual controls or expensive external sensors should cost less to build than a modified premium SUV.

Vehicle price is only one part of the calculation. Robotaxi operators must also pay for charging, cleaning, insurance, remote support, depot rent, repairs and customer service.

Empty driving can become especially expensive. A vehicle earns nothing while travelling to its next customer, returning to charge or moving from a quiet area to a busy one.

Utilisation will decide whether the cheap hardware produces a cheap ride. A Cybercab completing thirty paid trips per day can spread its fixed costs widely. The same vehicle completing five trips while frequently waiting for support would be much less attractive.

Tesla currently publishes no cost per paid mile, average daily utilisation or Robotaxi gross margin. Its cost advantage looks credible at the vehicle level. The economics of the full service remain open.

Q14Does Waymo make Tesla’s Cybercab timetable look realistic?

Waymo makes Tesla’s opportunity look more credible and Tesla’s timetable look less credible.

Waymo has shown that customers will repeatedly use driverless taxis. The company now operates across eleven cities and says its service areas cover more than 1,400 square miles. Recent reporting places its weekly ride volume above 500,000.

That equals more than 70,000 rides per day. It is a substantial transport business, although still small compared with the tens of millions of daily trips handled by conventional ride-hailing platforms.

Waymo reached that scale after years of driverless operations, thousands of vehicles and extensive work with local governments. It is now introducing a cheaper sixth-generation driving system and preparing more than twenty additional markets.

Tesla hopes to move faster by manufacturing its own lower-cost vehicle and using software trained on billions of supervised miles. That approach could shorten the rollout once it works consistently.

Current fleet numbers show no evidence of that acceleration yet. Tesla remains several orders of magnitude below Waymo in disclosed weekly rides and active commercial vehicles.

Waymo proves that robotaxis can become a real service. Its long path to eleven cities shows why “everywhere” remains a much bigger claim.

Q15Can we trust Tesla’s Cybercab timelines?

Tesla’s Cybercab forecasts deserve a heavy discount.

At its 2019 autonomy event, Tesla predicted that more than one million robotaxis would be operating the following year. Its first limited paid service eventually arrived around five years after that fleet was supposed to exist.

Elon Musk later predicted that autonomous ride-hailing would cover roughly half of the US population by the end of 2025. He also discussed around 500 vehicles in Austin and 1,000 in the Bay Area by that point.

Today’s operating footprint falls far below those forecasts. Austin appears to have fewer than twenty active driverless cars, while the Bay Area service still uses human drivers.

The miss is measured in both years and orders of magnitude. Another executive deadline tells us less than a visible increase in vehicles, rides and unsupervised mileage.

Tesla has eventually delivered difficult products after missing earlier dates. Model 3 production, large battery factories and the global Supercharger network all went through painful ramps.

Cybercab may follow the same pattern. Believe the operating data first and management’s calendar second.

Q16What would prove that Cybercab is finally scaling?

The next proof point is hundreds of Cybercabs carrying paying riders. Another city announcement adds little by itself.

A few employee rides would show that the vehicle functions. A few dozen public cars would confirm that Cybercab has joined the Robotaxi network. Hundreds operating daily would demonstrate the beginning of a commercial ramp.

Tesla also needs to publish enough data for outsiders to judge the service. Weekly rides, paid miles, active vehicles and average utilisation would reveal whether new cars are doing useful work.

Safety figures need the same treatment. A crash count without total driverless mileage is hard to interpret. Remote interventions without a denominator tell us little about how often cars need rescuing.

Regulatory progress will provide another clear test. A driverless deployment permit in California or commercial approval in a demanding international market would carry more weight than several additional launches in friendly states.

The evidence that would show real Cybercab scale

Q17When could Cybercab become common across the United States?

Our base case puts useful Cybercab coverage across most large US metros around 2029 to 2031.

Over the next year, Tesla’s realistic job is to place production Cybercabs into public service and show that the fleet can grow beyond a few dozen vehicles. More city launches are likely, especially in states with friendly rules, but many could remain thin.

The rollout could speed up during 2027 and 2028 if three things happen together: Cybercab production stabilises, remote intervention rates fall and the early fleets build a convincing safety record.

Tens of thousands of vehicles would already be enough to create noticeable fleets across dozens of large metropolitan areas. Tesla does not need millions of Cybercabs before the service becomes common.

Harder markets would probably follow later. California requires additional permits. Northern states bring snow, ice and dirty sensors. New York and other dense cities add stricter local transport rules and difficult street layouts.

In a fast scenario, Cybercab could reach dozens of useful US markets by 2028 or 2029. That would require a sharp break from Tesla’s current fleet growth.

A serious crash, weak economics or frequent remote intervention could push broad US coverage beyond 2032. The range remains wide because Tesla has revealed very little operating data.

Q18When will Tesla’s Cybercab be everywhere?

Cybercab is unlikely to become ubiquitous before the early 2030s, and literal worldwide coverage may never arrive.

Our base case is meaningful service across most large US cities between 2029 and 2031. Selective international expansion could follow between 2032 and 2035, starting with countries that have clear national rules and attractive ride-hailing markets.

Tesla has made real progress lately. Autonomous Model Y rides now span several markets, paid mileage is growing and Cybercab manufacturing has begun.

Cybercab is still in pilot production and has not carried public passengers. Tesla’s visible driverless fleet remains tiny, while a useful nationwide network would need tens of thousands of active vehicles.

International growth adds another layer. Tesla must satisfy different vehicle standards, taxi regulations, insurance systems, data rules and local authorities. Europe, Britain and China are creating clearer frameworks, although none offers an automatic route to continent-wide operation.

Tesla’s strongest advantage is its ability to manufacture lower-cost electric vehicles at automotive scale. Its biggest weakness is the autonomy timetable, which has repeatedly moved much more slowly than management predicted.

Cybercab has a credible path to becoming a major urban transport network. “Everywhere” will come much later than the first few city launches suggest.

Cybercab rollout scenarios