- Elon Musk projects that Optimus will account for up to 80% of Tesla's value in the medium term.
- Tesla shifts training strategy to a vision- and video-based approach.
- Pilot production in Fremont in 2025, with possible delivery to third parties in the second half of 2026.
- Ambitious goal: scale Optimus to one million units annually in five years and present Optimus 3 prototypes before the end of the year.
While the car business is suffering, Tesla has made a move by placing its humanoid robot Optimus at the center of his strategyElon Musk maintains that, within a few years, up to 80% of the company's value could come from this line, a bet that comes alongside the most recent master plan, where robots appear for the first time as a key element.
The roadmap includes ambitious goals—such as achieve a production of around one million units per year in about five years and show new prototypes, called Optimus 3, before the end of the year—although with few fine details for now. The pitch is clear: move the AI to the physical world for industrial and domestic tasks, with a product that Tesla describes as potentially transformative.
Optimus as a value lever for Tesla
In the most recent quarter, vehicles still contributed a good portion of revenue, but Deliveries fell by around 13% in the first half of the year, a context that explains why Musk asks investors to look at robots and the autonomy as the next growth engines.
Tesla's latest master plan, shorter than the previous one, reinforces that narrative: The company wants to put AI to work in real environments And not just in driver assistance software. Although Musk has announced other plans in the past with mixed results, his current message positions Optimus as the great value lever.
Musk has repeated that Optimus could concentrate up to 80% of the corporate value in a few years. Even so, he acknowledges that implementation will be complex and that the deadlines are aspirational. There are precedents: some objectives of previous plans did not materialize in a timely manner.
In production, Tesla talks about a pilot line in Fremont during 2025 to begin deploying robots in useful roles within its own factories. Regarding sales to third parties, Musk himself noted that his "very rough estimate" points to the second half of 2026, subject to technical advances.
How Optimus Trains: From Video to Practice
Tesla has refocused the program towards an approach based on vision and video dataAfter employing teleoperation and motion capture suits, it now prioritizes recordings of people performing real-life tasks so the robot learns to imitate actions like picking up objects or folding clothes.
To scale up data collection, The team records with a set of five cameras developed by Tesla that is mounted on a helmet and a backpackMulti-view capture makes it easier for AI models to understand the scene and the positioning of hands and joints in greater detail.
The leadership change has also marked the transition: Ashok Elluswamy, head of AI, He took over the Optimus program after Milan Kovac's departure.During the change, hiring was briefly paused and then resumed with dozens of robot-focused openings.
Experts consulted warn that training only from video has its limits: without physical interaction, the robot loses signals that are obtained by touching and manipulating the environment. Teleoperation, used by companies such as Boston Dynamics, It remains a useful reference for generating contact and force data..
Other researchers point out clear advantages of multiview: it helps to infer hand and finger postures and can complement previous teleoperation data. Tesla has already shown clips of Optimus performing tasks learned from recordings, and Musk went so far as to claim that the robot will be able to learn by watching videos on public platforms, like YouTube.
Planned schedule, production and deployment
On the technical side, Tesla plans to present Optimus 3 prototypes before the end of the year, continuing with iterations focused on dexterity, mobility, and reliability. The medium-term goal is to scale to a high-volume series, with with an eye on a million a year in about five years.
As for uses, the first applications are factory tasks and domestic chores Simple tasks, where repetition and precision add value. Tesla has already released videos of the robot folding clothes or sorting objects, examples of “learning by demonstration” from video.
Regarding commercial deadlines, Musk noted that —if everything goes according to plan— Deliveries to third parties could begin in the second half of 2026First, the plan is to expand internal deployments and refine the product in controlled settings.
The shift towards robotics occurs in a demanding and competitive context pressure on key automotive marketsFor Tesla, Optimus is both a path to diversification and a vision for the future that must be backed up by solid technical and production milestones.
Technical challenges, feasibility and open questions
Musk himself has acknowledged that Optimus's training needs They are enormous, probably "at least 10 times larger" than those of the self-driving car. This requires data, simulation, and validation infrastructures on a scale uncommon in humanoid robotics.
Another issue is the generalizationGoing from watching a video to performing a task in the real world involves understanding, planning, and precise control. Researchers in AI and robotics emphasize that, in addition to observing, robots need to practice—in simulators and physical environments—to fine-tune their behavior.
Compared to the industry standard — teleoperation and motion capture — Tesla is attempting a “very Tesla path” based on massive vision data setsThe big question is whether this approach, combined with contact information when necessary, will be sufficient for more complex and evolving tasks.
It remains to be seen whether the company will reprioritize capture suits or will consolidate a hybrid approach. For now, The firm has not publicly detailed all the details of the data and training pipeline., and did not respond to requests for comment on certain aspects.
The goals—from next-generation prototypes to the full-scale leap—are ambitious. If Tesla manages to string together consistent advances, the bet on Optimus could change the company's profileOtherwise, the program will run into the practical limits that have held back other humanoids.
With a renewed focus on Optimus, a training that prioritizes the camera vision and with still tentative deadlines—pilot in 2025 and possible deliveries to third parties in 2026—Tesla is testing its ability to bring AI to the physical world at an industrial pace without losing the balance between promise and reality.
I am a technology enthusiast who has turned his "geek" interests into a profession. I have spent more than 10 years of my life using cutting-edge technology and tinkering with all kinds of programs out of pure curiosity. Now I have specialized in computer technology and video games. This is because for more than 5 years I have been writing for various websites on technology and video games, creating articles that seek to give you the information you need in a language that is understandable to everyone.
If you have any questions, my knowledge ranges from everything related to the Windows operating system as well as Android for mobile phones. And my commitment is to you, I am always willing to spend a few minutes and help you resolve any questions you may have in this internet world.