Countdown To Cybercab, Tesla’s Multi-Trillion Dollar Robotaxi Opportunity

On October 10th, Tesla will unveil Cybercab, its highly anticipated robotaxi, taking a crucial step toward launching its autonomous ride-hailing service. In our view, an autonomous taxi platform will unlock a multi-trillion dollar market and begin to dominate Tesla’s valuation, approaching ~90% of its enterprise value over the next five years.¹

How would Tesla’s robotaxi service work?

Through over-the-air updates to its Full Self-Driving (FSD) software, Tesla estimates it will enable full autonomy in the US during the next year, as shown below.² Once the FSD update is live, ARK estimates Tesla could launch a robotaxi service in the next year or two. Even before full autonomy is achieved, Tesla could launch a human-driven ride-hail service, which would collect valuable data on routes, customer behavior, and operational dynamics while also informing its service infrastructure and building its customer base. 

When Will Tesla Achieve Full Autonomy? 

Tesla Estimate:

“...based on the current trend, it seems as though we should get miles between interventions to be high enough that—to be far enough in excess of humans that  you could do unsupervised possibly by the end of this year. I would be shocked if we cannot. do it next year.”

 – Elon Musk, 7/23/24³

Source: Tesla earnings call, July 23, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

ARK Estimate:

*Note: Figures may not add to 100% due to rounding. Source: ARK Investment Management LLC, 2024. This ARK analysis draws on a range of external data sources, data as of 6/12/24, which may be provided upon request. Forecasts are inherently limited and cannot be relied upon. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

Importantly, production of a low-cost Cybercab should not be a bottleneck for the launch of Tesla’s robotaxi service. Tesla’s full autonomy solution should be compatible with vehicles equipped with Hardware 3 and Hardware 4—about 6.5 million cars globally, with ~2.5 million cars in the US.⁴ To jumpstart the service, it could deploy a fleet of existing Model 3 and Model Y vehicles, including those coming off lease and out of inventory,⁵ while customers opting to enroll their vehicles in the robotaxi service⁶ could supplement the fleet. In the long run, most of Tesla’s robotaxi fleet is likely to be owned and operated by third-party partners, with Tesla hosting the ride-hail platform itself and perhaps maintaining a small fleet of its own vehicles.

The importance of Tesla’s ability to scale the fleet quickly should not be underestimated. In contrast, with a fleet last reported to be ~700 vehicles,⁷ Waymo currently operates a robotaxi service in LA, San Francisco, Phoenix, and Austin, with wait times higher than Uber.⁸ Indeed, Waymo probably partnered with Uber in Phoenix, Austin, and Atlanta because of its scaling challenges.⁹

Like Waymo, Tesla is likely to launch its robotaxi service city by city; but unlike Waymo, it should be able to scale much faster, because it will not rely on HD maps or geofencing. Perhaps Tesla’s most important advantage over Waymo, however, is its data lake of real-world driving miles. 

According to ARK’s research, Tesla’s customer vehicles drive over 5 million miles per day in FSD, and over 87 million miles per day when including US-only non-FSD miles, generating valuable video clips for Tesla to train its autonomous software.¹⁰ ARK estimates that Waymo’s fleet drives ~70,000 miles per day, as shown below.¹¹

Note: FSD still requires driver attention, while Waymo does not have drivers in the above data. Source: ARK Investment Management LLC, 2024. This ARK analysis draws on a range of external data sources, data as of 9/27/24, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

Waymo’s fleet size and the limited driving conditions of the few cities in which it operates likely have restricted its access to real-world “corner cases.” In contrast, Tesla’s fleet has been collecting diverse camera data from all over the US.

Additionally, thanks to its vertically integrated production, Tesla has a clear path to scaling its robotaxi fleet, while Waymo must rely on Zeekr. Based in China, Zeekr could face hefty US tariffs,¹² which may be why Waymo has entered into another manufacturing relationship with Hyundai.¹³ ARK estimates that Waymo’s vehicles cost more than $100,000 to produce, its sensor set alone ~$40,000+,¹⁴ though it is working to reduce costs.¹⁵ Tesla’s Model 3 costs $40,000, sensors included.¹⁶ While it needs to build backend support for remote vehicle assistance and customer support, Tesla should be able to leverage its existing factory, charging, and service infrastructure to scale efficiently.

Regulatory approval should pose fewer obstacles than have the technical challenges to full autonomy. Competitors Waymo and Cruise have paved the regulatory path.¹⁷ In the US, 40 states either have approved autonomous vehicles explicitly or have not prohibited them.¹⁸

Moreover, its vast data library should help Tesla prove its safety benefits to regulators. ARK estimates that autonomous vehicles could reduce accident rates by more than 80%. Last year, Tesla’s data already had demonstrated that its FSD-equipped cars were ~5x safer from accidents than its cars without FSD and ~16x safer than the average car, adjusting for city street driving.¹⁹ Meanwhile, Waymo’s cars are roughly twice as safe as the national average with over 70% fewer injury crashes.²⁰

With higher utilization rates, autonomous ride-hail service could undercut the price of human-driven ride-hail services. New York City taxis currently operate at ~30-40% of their revenue utilization rate, or the amount of time that a passenger is in the car during any given 24-hour period. Autonomous vehicles could exceed a 50% utilization rate, which ultimately would lower the cost per mile. Lower prices should stimulate demand and expand the market over time, as shown below. Notably, even without full autonomy, a human-driven ride-hail service hosted by Tesla could undercut today’s ride-hail rates, as the operating cost of an electric vehicle is roughly one third that of a gas-powered car.²¹

*$11 Trillion is the addressable market, not the revenue we expect in 2030, as we do not expect autonomy to penetrate all addressable miles. Sources: ARK Investment Management LLC, 2024. This ARK analysis is based on a range of underlying data from external sources, which may be provided upon request. Forecasts are inherently limited and cannot be relied upon. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

Thanks to its lower cost structure and the price umbrella provided by traditional ride-hail services, we believe Tesla could command a take rate more than double the ~30% that of Uber. According to our research, as autonomy reduces costs and expands the market, robotaxi platforms could scale to ~$4 trillion in net revenue in 2030, as shown below.²² Targeting a total available market (TAM) much greater than Uber had envisioned, Tesla also is tackling the largest applied AI project on earth, one that could save more than 40,000 lives in the US and ~1.2 million globally per year.²³

Note: ”EBIT” denotes Earnings before interest and taxes. In conventional parlance, Enterprise value (EV) measures a company’s total value. Its calculation includes not only the market capitalization of a company but also short-term and long-term debt, as well as any cash or cash equivalents on the company’s balance sheet. It is often used as a more comprehensive alternative to market capitalization when valuing a company."²⁴ Source: ARK Investment Management LLC, 2024. This ARK analysis draws on a range of external data sources, data as of 1/31/24, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Forecasts are inherently limited and cannot be relied upon.