What would it take for a car manufacturer to evolve into a global autonomous mobility service provider? For Tesla, according to Morgan Stanley, the key may lie in a vision of a million fully autonomous Robotaxis by 2035, which would rewire the paradigm surrounding the use of AI in transport.
The prediction, made by analyst Andrew Percoco, presents a conservative but boldscale path. Tesla’s current count, estimated at between 50 to 150 units in its current limited operations, is to extend to 1,000 units by 2026. The critical mass will be reached in the next decade, in which the company’s vertically integrated approach, comprising both the vehicles and the Full Self-Driving (FSD) software stack, might help it dominate the market even in a world where companies are using licensed autonomy software.
There are three catalysts that are prominent in Morgan Stanley’s outlook. Firstly, the absence of the Safety Monitor, which is the human observer in the passenger seat, is the start of full-scale unsupervised driving. Elon Musk announced this month that Tesla is testing Model Y Robotaxis in Austin with no human being present at all, a step Musk predicted in late 2025. Secondly, the continuance of a positive trajectory in safety scores without the Safety Monitor is paramount. The latest FSD 14.2 beta release announced in Tesla includes a continued focus on gathering real-world driving experience to develop neural networks specifically built to be safe and effective in their performance. Thirdly, the start of production of Cybercabs in April 2026 will initiate a dedicated autonomous car without a wheel or pedals, made using Tesla’s groundbreaking “Unboxed” car-building method.
The Cybercab’s design approach is actually revolutionary in comparison with traditional car models. Components are produced in parallel, and subsequently integrated during final assembly stages, a production technique which, according to Musk, closely resonates with consumer electronics manufacturing techniques. He once said that this technique offered potential cycle time of about ten seconds for every vehicle produced, and that with this technology, a possible five million could be produced on an annual basis.
“Tesla’s FSD system combines inference hardware integrated into our vehicles with a vision-based neural net trained on several billion miles of data, while Uber’s system relies on an ecosystem business, partnering companies like Waymo, Aurora, and Motional.” Tesla has a full-stack system, meaning that it controls everything, right from sensor adjustment to software update, whereas Uber has an ecosystem business, which partners with firms that are also involved in the development of autonomous systems, firms such as Waymo, Aurora, and Motional. This gives a clear system to implement a revenue stream, but also allows it to license its FSD tech to car makers. However, as stated by
The performance of safety capability is an area of great importance for regulatory and investment communities alike. As of September, Waymo’s cars experienced one crash per 98,600 miles, while the nearest competitor, Tesla, experienced one crash per 62,500 miles. This significant disparity reflects the critical importance of Tesla’s ongoing validation before allowing unattended driving with paying clients. In the Austin deployment scenario, supervised experience identified scenarios that require human intervention, such as map problems on one-way streets. With the removal of the Safety Monitor function, the AI stack will assume all responsibility for operations, and a high level of safety performance will no longer be a consideration for regulatory compliance but a business imperative.
Investor sentiment is shifting according to Tesla’s technological roadmap. According to the insights of Dan Levy of Barclays, deliveries of vehicles are becoming less relevant as an valuation metric and are being increasingly followed by developments in AI, Optimus humanoid robots, and autonomous mobility solutions. This is already being reflected in the words of Tesla’s Elon Musk: “The future of the company is fundamentally based on large-scale autonomous cars and large-scale and large volume, vast numbers of autonomous humanoid robots.” Tesla’s Master Plan Part 4 reflects this by placing AI/robotics at its center rather than autos.
The Robotaxi initiative’s expansion outside of Austin and the Bay area of California is now under way with Robotaxi pilot launches set to take place in the cities of Houston, Dallas, Miami, Las Vegas, and Phoenix. Speed of deployment rather than geographic reach appears to be the name of the game at this point, wrote Mark Delaney of Goldman Sachs, continuing, We believe the focus will be on how quickly Tesla can deploy driverless networks (including how rapidly they can scale vs. rivals if the company truly has a competitive advantage with their approach to software and hardware capabilities). Tesla’s hand has been strengthened by the successful acquisition by the company of Musk’s fellow directors, Jakom reports. A settlement has been reached involving the Securities and Exchange Commission, Musk, and the company, although the terms are not yet publicly known.
Tesla’s chances of meeting Morgan Stanley’s predictions of a million units sold will depend on the firm’s ability to implement the company’s product milestones. The combined strengths that Tesla has developed in autonomous hardware, its mature AI software platform, and innovations in manufacturing have positioned the company to successfully compete in the industry for ride-sharing and AV platform solutions. For technology-centric investors, the story surrounding
