Latest Research by J.P. Morgan at Tesla Factory on the Large-scale Deployment of Robotaxi and the Launch Pace of Optimus Robot

Tesla's two core narratives—scaled deployment of Robotaxi and commercialization of the Optimus humanoid robot—are moving from concept to a quantifiable execution path. JPMorgan recently completed an on-site investigation at Tesla's Fremont factory and met with the company's investor relations team. Overall, analysts gave positive evaluations of the progress on these two business lines, believing both are broadly in line with the previously disclosed management timetable.

According to JPMorgan's report, the key trigger for expanding the Robotaxi fleet is now clearly focused on the rollout of FSD v15 within the year. Management stated that technical validation for a small-scale test fleet of Cybercab continues, with about 40% of v15’s core technology modules already being tested in the Robotaxi fleet and initial feedback being positive. At the same time, the company intends to control the number of Model Y units converted to the Robotaxi fleet, reflecting confidence in Cybercab's near-term scalability. For Optimus, the Gen 3 design has been finalized, the supply chain is essentially secured, and production lines are being installed at the Fremont factory. The initial deployment goal after the start of production (SoP) is to enter the 'Optimus Academy' training phase in the second half of 2026, with the earliest external commercial sales expected in the second half of 2027.

JPMorgan maintains a Neutral rating on Tesla, with a price target of $445. At the time of the report's publication, the stock was trading at $339.30. Analysts believe expansion of the Robotaxi fleet will see a significant acceleration from the end of 2026 to early 2027, Optimus commercialization has a clear path, and rising FSD penetration along with a new model lineup together support the recent uptick in demand.

JPMorgan’s report notes that the next critical deployment point for the Robotaxi fleet is directly tied to the release of FSD v15. Management characterizes v15 as a step change in performance, comparable to the leap from v13 to v14—which involved significantly increasing parameters, expanding context windows, and reducing latency by about 20%. v15 consists of seven core technologies, with roughly 40% already tested in the Robotaxi fleet and receiving good feedback.

Tesla stated that the existing AI/HW4 hardware stack is already capable of running v15 and supporting unsupervised FSD, while the upcoming AI4.5 computing system is designed for future demands, offering roughly a 10% improvement in processing power (FLOPS) and about double the memory to meet growing computational needs from Robotaxi model scaling and increased context windows.

For vehicle strategy, Tesla explicitly stated the intention to limit further additions of Model Y into the Robotaxi fleet, based on management’s confidence in scaling Cybercab in the near term. Cybercab uses unboxed manufacturing, increasing assembly efficiency through independently parallel construction of major submodules that are ultimately merged. Technical validation and capacity building are proceeding simultaneously.

In terms of unit economics, Tesla revealed that the total cost of ownership (TCO) for Model Y and Model 3 under personal vehicle usage is about $0.60 to $0.70 per mile; with typical Robotaxi utilization (4 to 5 times higher than personal use), it can fall to $0.50 to $0.60 per mile—significantly lower than current ride-hailing platforms, which charge about $2.5 to $3.0 per mile. Management also emphasized that Robotaxi’s long-term addressable market far exceeds traditional ride-hailing (which accounts for only a low single-digit percentage of overall mobility), with the aim to push TCO down to about $0.30 per mile via a dedicated Robotaxi platform. Cybercab is only the initial form, with more models to follow later.

During the latest on-site visit at Fremont factory, the Optimus production area was covered by canvas, so JPMorgan analysts could not observe directly. However, Tesla management confirmed that the relevant production line is being installed at the former Model S/X line (retired in May 2026), with overall progress essentially on target for the planned about 4-month conversion cycle.

Tesla outlined a three-phase commercialization path for Optimus: Phase one, in the second half of 2026, will see deployment of the Optimus robot in the "Optimus Academy" to accelerate data accumulation through real-world interaction; the second phase will extend deployment internally to Tesla's factories for further data collection and to avoid third-party data compliance complexity; third phase anticipates launching external commercial sales as early as the second half of 2027.

For internal application prioritization, management indicated that stamping and body-in-white processes are most likely to benefit first from humanoid robots because these tasks are highly repetitive and dangerous, while general assembly lines—which still heavily rely on manual dexterity—will be longer-term application scenarios.

Tesla also disclosed that the Gen 3 design is finalized, the supply chain is locked in, but some aesthetic details are still being optimized and will be released closer to SoP. The official debut of Gen 3 will be intentionally delayed until close to mass-production to preserve competitive advantage; the scope and cost targets for Gen 4 will be set only after accumulating practical experience running Gen 3. The Fremont factory’s long-term capacity target is about 1 million units, while the future goal for the Texas Gigafactory is about 10 million units. In terms of computing power, Tesla stated that computing capacity for the first half of 2026 has doubled year-on-year.

JPMorgan’s report shows that FSD is increasingly becoming a core consideration for consumers when deciding to purchase a vehicle. Management has noticed more consumers visiting stores specifically to learn about FSD features—a trend observed in early markets such as Australia, South Korea, and parts of Europe, where demand saw a clear uptick once FSD was launched.

Regarding pricing strategy, Tesla ended the one-time purchase option for FSD in the U.S. and Canada as of February 2026 and will complete a global shift to subscription-only by August of the same year. Management said the current $99/month pricing is primarily aimed at expanding the subscription user base and boosting usage frequency, not short-term price hikes. This is based on the fact that historically, about 50% of Tesla owners have never tried FSD, and some users of older versions have not activated a new subscription. Management believes FSD has strong stickiness—a high renewal rate once experienced—which is the logic behind offering every new owner a one-month free trial.

For regulatory progress in Europe, Tesla has adopted a dual-track approach: direct engagement with the EU level (where the scheduled approval timeline has been repeatedly postponed, now expected in October), while also collaborating with individual member countries like the Netherlands. Once an appropriate regulatory framework is established, it could be adopted by others. Tesla disclosed that European FSD driving records indicate about a fivefold reduction in collision events over roughly 65 million kilometers of driving, a safety metric management believes will help accelerate regulatory approval. Once approved, activation cycles are expected to be counted in weeks, not months or quarters.

Tesla attributed the 25% year-over-year and 34% quarter-over-quarter increase in vehicle deliveries in Q2 2026 (the largest single-quarter increase since 2019) to two factors: the ongoing evolution of FSD features and an optimized new model lineup—including an updated base model, Model Y L, and revised performance versions—covering broader usage scenarios and price points.

As for gross margin, Tesla noted they have implemented targeted price adjustments for some Model Y variants and more generally for Model 3 globally, and revised interest rate subsidy plans to manage commodity cost pressures. Both Q1 and Q2 automotive gross margins were also impacted by the FSD business model shift from one-off sales to subscription. In-house cathode and anode material production began in January 2026 and is expected to gradually yield cost improvements, although their precise impact is hard to quantify individually due to multiple factors—typically, it takes about 18 months for a factory to reach reasonable scale and utilization rates.

JPMorgan expects Tesla to deliver about 1.8 million vehicles in 2026, while the current total capacity ceiling is around 3 million units—leaving management significant room to boost revenue through higher capacity utilization.