Tesla Full Self-Driving, Cybercab and Optimus Timeline Faces Skepticism

Tesla Updates Its Full Self-Driving and Robotics Timeline

Tesla recently outlined updated timelines for Full Self-Driving, the Cybercab robotaxi, and the Optimus humanoid robot during a meeting with JPMorgan at its Fremont factory. The bank's notes included several significant claims, including Tesla's description of FSD v15 as a major “step-change,” its assertion that the current HW4 computer can support unsupervised driving, and the possibility that Optimus could become available to outside buyers as early as the second half of 2027. However, many of these targets resemble promises Tesla has made previously, including claims that earlier hardware and software generations would eventually enable unsupervised driving.

FSD v15 Is Being Presented as Another Major Capability Jump

Tesla told JPMorgan that FSD v15 represents a substantial improvement in capability and is built around seven “core technologies.” About 40% of those technologies are reportedly already operating in the robotaxi fleet in Austin, while early feedback has been described as “encouraging.” The company shipped FSD v14 in October 2025, marking its first major update in roughly a year, while Elon Musk had described v14.3 in April as the “last piece of the puzzle.” That makes v15 at least the third consecutive major version presented as a potentially decisive advance toward the same long-promised destination: unsupervised Full Self-Driving in customer-owned vehicles.

The broader issue is that Tesla has promoted this destination for years without reaching it for customers. Since 2016, the company has repeatedly positioned Full Self-Driving as technology that would eventually operate without driver supervision. Each major software generation has been presented as a meaningful step toward that objective, yet customers continue to supervise their vehicles. The latest claims therefore carry a credibility challenge because the underlying promise has remained largely unchanged while the software versions, hardware platforms, and projected timelines have continued to evolve. FSD v15 may represent a genuine technical improvement, but the claim that it finally enables unsupervised driving remains unproven.

HW4 Faces the Same Unsupervised Driving Question

Tesla also told JPMorgan that its current HW4 computer, which the company calls AI4, is sufficient to run FSD v15 and unsupervised Full Self-Driving. At the same time, Tesla is introducing AI4.5, a newer chip version with roughly 10% more compute and approximately twice the memory. The coexistence of these claims raises an important hardware question because Tesla previously made similar assurances about HW3. The company maintained for years that HW3 would be capable of supporting the eventual Full Self-Driving system, but HW3 vehicles are now receiving a reduced FSD v14 “Lite” experience rather than the full system.

The hardware roadmap adds further uncertainty. Tesla has delayed its next-generation AI5 chip to mid-2027, while the Cybercab is expected to launch using AI4 or potentially AI4.5. That progression means the hardware platform associated with the promise of unsupervised driving continues to shift. The experience with HW3 has also become more complicated because reports have highlighted rising computer failures associated with the software. For customers and investors, the central question is therefore not simply whether HW4 can technically run FSD v15, but whether the hardware will remain adequate as the system develops. Tesla's history makes that distinction particularly important.

Cybercab Scaling Remains a Major Test

On the robotaxi program, Tesla told JPMorgan that it is confident it can scale the Cybercab in the near term. The company is deliberately limiting additional Model Y robotaxi conversions because of that confidence and plans to use an “unboxed” manufacturing process. Tesla is targeting a long-run operating cost of around $0.30 per mile, with fleet expansion expected to accelerate from late 2026 into early 2027. Those plans assume safety performance can support broader deployment, something Tesla's robotaxi program has been pursuing for more than a year. The company is therefore presenting manufacturing scale, lower operating costs, and fleet expansion as the next major stages.

Actual service performance, however, remains considerably smaller than the long-term ambition. Tesla's driverless service in Austin still operates with only a limited number of vehicles more than a year after launch, and the robotaxis continue to experience issues despite relatively low mileage. Tesla recently disclosed only about 380,000 unsupervised miles since the service began. By comparison, Waymo has accumulated more than 200 million miles. The difference in operating experience highlights why confidence in scaling should not be treated as evidence that large-scale autonomous service has already been demonstrated. Tesla is pursuing a rapid expansion, but the existing service remains much smaller than the stated long-term objective.

Optimus Gen 3 Could Reach Customers in 2027

Tesla also told JPMorgan that the Optimus Gen 3 design has been finalized, its supply chain is “essentially locked in,” and a production line is being established in the former Model S/X facility in Fremont. External commercial sales could begin as early as the second half of 2027. Tesla has also outlined long-term production ambitions of approximately 1 million units at Fremont and 10 million units in Texas. Those figures imply a manufacturing program on an enormous scale, but the immediate challenge remains demonstrating that the Gen 3 robot can perform useful work consistently. The timeline therefore depends not only on production readiness, but also on the technology reaching the required level of capability.

The Optimus timeline is also weakened by Tesla's previous delays and changing claims. The Gen 3 reveal has been pushed back repeatedly, while Musk has acknowledged that no Optimus robots are currently performing useful work inside Tesla after previously making claims about their factory capabilities. Musk had also previously said Tesla would have 5,000 Optimus robots last year, whereas the company still has only a small number of robots being used for testing. The current pitch is consequently ambitious: a robot that has not yet been fully demonstrated, has not established useful factory work at scale, and could nevertheless become available to external customers in less than two years.

What the Latest Timeline Means

The latest JPMorgan meeting reinforces Tesla's strategy of pursuing multiple ambitious technology milestones simultaneously. FSD v15 is being positioned as another major advance toward unsupervised driving, Cybercab is being presented as ready for accelerated scaling, and Optimus is being assigned an external-sales target beginning in the second half of 2027. Yet each program still faces a substantial gap between stated targets and demonstrated scale. The company's previous experience with FSD hardware, robotaxi deployment, and Optimus production makes those gaps particularly important. Investors and customers will ultimately need to judge these programs by demonstrated performance, regulatory readiness, production volumes, and sustained real-world operation rather than by projected milestones alone.

Frequently Asked Questions

When could Tesla begin selling Optimus to outside customers?
Tesla told JPMorgan that external commercial sales of Optimus Gen 3 could begin as early as the second half of 2027. The company says the design is finalized, its supply chain is essentially locked in, and production preparations are underway at Fremont. Tesla also has much larger long-term production ambitions for Fremont and Texas. However, the timeline remains subject to the robot reaching useful performance levels, completing production development, and overcoming the delays that have affected earlier Optimus milestones. Tesla's previous claims about factory use and production volumes also make the 2027 target difficult to evaluate solely from the latest projection.