JPMorgan Says Tesla to Accelerate Robotaxi by 2027
JPMorgan's research indicates that Tesla will cautiously transition the Model Y into its Robotaxi fleet, reflecting management's increased confidence in the recent large-scale deployment of Cybercab; the main acceleration window for fleet expansion is expected between late 2026 and early 2027.
The report identifies FSD v15 as the core trigger for Robotaxi expansion. Tesla's management states that v15 includes seven core technologies, with approximately 40% currently being tested in the Robotaxi fleet and receiving initial positive feedback; its performance goals are described as a leap similar to the upgrade from v13 to v14.
In comparison, the report notes that the upgrade from v13 to v14 involved a larger model parameter count, a longer context window, and about a 20% reduction in latency. Tesla states that the existing AI/HW4 hardware can run v15 and support unsupervised FSD; the planned AI4.5 system is expected to provide about 10% higher computing power and approximately double the memory to accommodate model and context expansion.
The unit economic model shows that the total cost of ownership for Model Y and Model 3 at typical private car utilization rates is about $0.60 to $0.70 per mile; if calculated based on typical Robotaxi usage intensity, which is about 4 to 5 times that of private cars, this cost could drop to $0.50 to $0.60 per mile. The report cites traditional ride-hailing costs at about $2.50 to $3 per mile.
Tesla's long-term goal is to reduce the total cost of ownership to about $0.30 per mile through a dedicated Robotaxi platform. The Cybercab is positioned as the first model specifically designed for autonomous driving, rather than a final product; the company plans to introduce more dedicated models in the future. Recent English reports also indicate that Tesla is preparing to gradually integrate the Cybercab, which has no steering wheel or pedals, into its existing Robotaxi service in Austin, although the timeline remains adjustable.
The commercialization route for Optimus is divided into three phases: entering "Optimus Academy" training in the second half of 2026 to accumulate real-world data; subsequently deploying within Tesla's factories; and selling to external customers as early as the second half of 2027. The research also states that the design for Optimus Gen 3 has been completed, the supply chain is largely locked in, and production lines are being installed at the Fremont factory.
In terms of market mechanisms, Tesla is the supplier of the autonomous driving mobility network, dedicated vehicles, and AI infrastructure, while passenger and enterprise capacity demand are the buyers. If the Cybercab achieves operation without a safety driver under regulatory approval and reaches a cost of about $0.30 per mile, the cost structure for traditional ride-hailing drivers, as well as the commission structures for Uber and Lyft, will be under pressure; beneficiaries will include Tesla itself, the autonomous driving computing power and sensor supply chain, insurance, and high-utilization fleet operation systems. The aforementioned costs and timelines are derived from company management and sell-side research and have not yet been validated by large-scale public operating data.
Source: Public Information
ABAB AI Insight
Tesla's Robotaxi path differs from Waymo's. Waymo started with high-precision maps, LiDAR, multi-sensor redundancy, and geofenced operations in limited areas like Phoenix, initially pursuing a safe closed loop within specific cities; Tesla bets on vision dominance, fleet data, end-to-end neural networks, and general hardware, using FSD subscriptions and mass-produced vehicles as data entry points. The company's control of Model Y within a transitional fleet size indicates its desire to avoid sinking too many assets into retrofitted models while concentrating capital expenditures on the dedicated Cybercab platform without a steering wheel or pedals.
The key to the capital path is leveraging existing asset cycles. Tesla accumulates cash flow and driving data through vehicle sales, FSD subscriptions, the Supercharger network, onboard computing power, and proprietary insurance data; if the Cybercab expands deployment, it will extend manufacturing capabilities from one-time vehicle sales to self-operated mobility assets. Unlike Uber, which primarily connects drivers and passengers, and Lyft, which relies on external driver supply, Tesla aims to control vehicle manufacturing, autonomous driving software, energy supply, and dispatch networks simultaneously, thus creating a larger potential profit pool, but also bearing risks of vehicle idleness, accident liability, operational scheduling, and local regulatory compliance.
A historical analogy can be drawn to Amazon's shift from reliance on third-party logistics to building its own warehousing and distribution network. Amazon initially depended on external carriers, then restructured fulfillment costs through warehousing, sorting, air transport, and last-mile delivery; if Tesla achieves low-cost Robotaxi, it will also be vertically integrating "vehicles-software-energy-mobility" into a service network. The difference is that e-commerce packages can be delayed and compensated, while autonomous passenger services face real-time safety liabilities and public road permissions, with operational scale not solely determined by manufacturing capacity.
Essentially, this represents a restructuring of the industry chain. The core scarce resource of traditional mobility platforms is driver supply, thus commission rates need to be distributed among drivers, passengers, and platforms; autonomous driving transforms scarce resources into permissions, vehicle turnover rates, software reliability, charging replenishment, and capital costs. If the cost per mile approaches $0.30, pricing power will shift from labor platforms to operators with autonomous driving systems and dedicated fleets; however, before reaching the operational thresholds of no safety driver, intercity service, and low accident rates, Robotaxi remains a pilot business with high capital, regulatory, and technological uncertainties.
ABAB News · Cognitive Laws
The profit of autonomous driving lies not in selling cars but in increasing vehicle turnover.
Once the cost curve falls below that of human labor, platform rules will be rewritten.
The endgame of autonomous driving is not algorithmic victory, but operational permission.