Quick summary: Tesla Optimus is Tesla’s humanoid robot, first announced in 2021. Version 3 — the model designed for actual production — is scheduled for reveal in late July or August 2026 and production start at Tesla’s Fremont factory the same summer, using the line freed up by ending Model S/X production in May. Musk has targeted 1 million units per year at Fremont by late 2026 and 10 million per year at Gigafactory Texas by 2027. Realistic 2026 output: low thousands. The headline V3 hardware upgrade is 22 degrees of freedom per hand (up from 11) with 50 total actuators across both hands. This guide covers what’s real, what’s marketing, and how Optimus compares to Figure, Boston Dynamics, and Apptronik. Written for beginners. Updated 2026-05-15.
Tesla revealed the original Optimus concept in August 2021 — a person in a robot costume on stage at AI Day. It was easy to dismiss. Three years later Optimus is real hardware, walking around Tesla factories, and is the centerpiece of Elon Musk’s argument that Tesla is an AI and robotics company that happens to make cars rather than a car company that’s experimenting with AI. The Version 3 robot, set for reveal in summer 2026 and production at Fremont using the line freed up by ending Model S/X production in May, is the model that’s supposed to validate whether Musk’s humanoid bet actually scales.
This is the explainer. What Optimus actually is, the credibility gap between announcement and shipping product, what the V3 changes, and whether the manufacturing scale Tesla is targeting is real or aspirational.
Get Smarter About AI Every Morning
Free daily newsletter — one story, one tool, one tip. Plain English, no jargon.
Free forever. Unsubscribe anytime.
What is Tesla Optimus in plain English?
Tesla Optimus (also called Tesla Bot) is a humanoid robot Tesla is developing in-house. The plan: build a robot shaped like a person, train it to do useful work using the same kind of AI that runs Tesla’s Full Self-Driving system, and produce it at automotive-industry scale so it becomes affordable for both businesses and eventually consumers. The robot is approximately 5’8″ tall and weighs around 125 lb, with the dimensions explicitly chosen so it can use tools and work in spaces designed for humans.
The strategic logic Tesla puts forward: a car factory and a humanoid factory require similar capabilities — supply chains for actuators, batteries, electronics, software, and large-scale assembly. Tesla has built the world’s most automated automotive production, and Musk argues that’s the right starting point for humanoid manufacturing scale. Whether the analogy holds is the open question, but it’s not a crazy argument.
What’s the history of Optimus so far?
| Date | Milestone |
|---|---|
| August 2021 | Tesla AI Day: original Optimus concept revealed; dancer in robot costume on stage |
| September 2022 | Tesla AI Day 2: first real Optimus prototype walks slowly on stage |
| December 2023 | Optimus Gen 2 reveal — substantially refined; better balance, faster walk, more dexterous hands |
| October 2024 | “We, Robot” event: Optimus units mingle with attendees; some interactions later confirmed to be partly teleoperated |
| 2024–2025 | Internal Tesla factory deployment begins; small numbers of Optimus units doing battery cell handling and other tasks |
| March 2026 | Abundance Summit: Musk states V3 reveal is “in final stages” and production starts summer 2026 |
| April–May 2026 | Last Model S/X rolls off Fremont line; line is converted for Optimus production |
| Summer 2026 (target) | V3 reveal (late July/August) and start of formal production at Fremont |
| 2027 (target) | High-volume production ramp; Gigafactory Texas humanoid lines |
Two things to note from this timeline. First, Tesla has consistently advanced Optimus through clearly demonstrable hardware generations — this is not vaporware. Second, the timelines have been pushed multiple times. The original Musk statement at AI Day 2021 implied a much faster timeline than what’s actually shipping. Treat every future Optimus date with the standard Tesla discount.
What changes with Optimus V3?
The headline V3 upgrade is the hand system. Optimus V2 had 11 degrees of freedom per hand. V3 doubles that to 22 degrees of freedom per hand, with 50 total actuators across both hands, using a biomimetic tendon-driven design. In plain English: V3’s hands can move and articulate more like human hands, which matters for tasks that require fine manipulation — picking up small parts, using tools, threading wires, manipulating soft or irregularly-shaped objects.
The V3 brain runs on Tesla’s new AI5 chip, which Tesla says delivers roughly 5x the memory bandwidth of its predecessor. This matters because the brain is the bottleneck for what a humanoid can actually do — better silicon means the robot can run larger and more capable foundation models on-device with lower latency. Tesla’s AI silicon advantage (built up over years of FSD work and the Dojo supercomputer) is one of the company’s two real competitive advantages, alongside its manufacturing scale.
What’s not known publicly about V3 as of mid-2026: full body specifications, exact battery life, the realistic price target, and where on the body the meaningful weight changes happen. Tesla has been less specific about V3 than Figure was about Figure 03 in the lead-up to that robot’s October 2025 reveal. The summer 2026 unveil is when most of the open questions will be answered.
How will Tesla actually manufacture Optimus at scale?
The Fremont factory’s Model S and Model X production line ended in May 2026 (Tesla announced this in April). The line is being converted to Optimus production with the V3 robot’s first manufacturing run targeted at late July or August 2026. Musk’s stated goal is to reach a 1 million-units-per-year run rate at Fremont by late 2026. Gigafactory Texas is planned to add humanoid production lines targeting 10 million units per year by 2027.
Sober context on those numbers: a 10-million-units-per-year humanoid factory would be larger than any car factory ever built. Tesla makes roughly 1.8 million cars per year globally as of 2025. The implication that Tesla would build five times as many humanoid robots per year as it builds cars within 18 months of starting humanoid production is the kind of stretch goal that’s worth taking seriously as ambition but not as a forecast. Realistic 2026 Optimus output is in the low thousands. Realistic 2027 output is more likely tens of thousands than millions.
That said, even “tens of thousands of Optimus robots in 2027” would be the largest humanoid production volume of any company. Tesla’s manufacturing scale is a real competitive advantage. Musk’s track record on timing is bad. Both can be true at once.
How will Tesla sell Optimus?
Musk has publicly stated the long-term price target is approximately $20,000 per robot. As of mid-2026 no Optimus units have been sold to external customers; deployment so far is internal at Tesla factories. The realistic price path: enterprise pilots first (likely industrial manufacturers, possibly some warehousing), at substantially higher prices than $20,000 — comparable to Figure or Boston Dynamics enterprise pricing in the $30,000–$150,000 range — followed by consumer availability much later at lower price points.
Tesla’s distribution advantage versus Figure or Boston Dynamics: Tesla has hundreds of stores worldwide and a customer base that already trusts Tesla products. If Optimus eventually becomes a consumer product, the existing retail and service infrastructure is real. Whether it ever becomes a consumer product at $20K is what determines whether Tesla’s humanoid bet pays off financially at the scale Musk has implied.
How does Optimus compare to Figure, Atlas, Apollo, and Digit?
| Robot | Production status (mid-2026) | External customers | Manufacturing scale ambition |
|---|---|---|---|
| Tesla Optimus V3 | Reveal summer 2026; production starts at Fremont same season | None yet; internal Tesla only | 1M/yr Fremont late 2026; 10M/yr Texas 2027 (stretch) |
| Figure 03 | BotQ factory online; in production | BMW Spartanburg (Figure 02); Figure 03 evaluation phase | 12K/yr year one → 100K/yr |
| Boston Dynamics Atlas | Serial production from March 2026 | Hyundai RMAC + Google DeepMind (only) | Hyundai factory targets 30K/yr by 2028 |
| Apptronik Apollo | Pilot phase; commercial scale targeted 2026 | Mercedes-Benz pilot, GXO R&D | Not yet specified at scale |
| Agility Digit | Shipping in production deployments | GXO (100K+ totes), Amazon, Toyota Mfg Canada | Not yet specified at scale |
Tesla’s manufacturing-scale ambition dwarfs every competitor. Tesla’s external customer count is zero. That asymmetry tells you exactly what kind of bet Tesla is making: that the manufacturing capacity will pull the customer demand, rather than the customer demand pulling the manufacturing capacity. This is the opposite of how Figure, Boston Dynamics, Apptronik, and Agility have approached the market.
What can go wrong for Optimus?
- Musk’s timeline track record. Robotaxi, Cybertruck, Roadster 2, FSD Level 5 — every Tesla project has shipped slower than Musk’s stated dates. Optimus V3 production-start in summer 2026 is the current claim. Treat it as ambitious.
- No external customer validation. Figure has BMW. Boston Dynamics has Hyundai. Apptronik has Mercedes. Agility has GXO. Tesla has Tesla. Without an external customer deploying Optimus in their own facility, the “this is a real product” narrative depends entirely on internal Tesla validation.
- The 22-DOF hand bet is hard. High-DOF biomimetic hands are mechanically complex, expensive to manufacture, and historically failure-prone. Tesla’s reputation for shipping ambitious hardware has limits.
- FSD-as-precedent is mixed. Tesla’s FSD has been “almost done” for nearly a decade. If Optimus’s brain follows the same trajectory, “almost a useful robot” could persist for years longer than the manufacturing buildout.
- Brand polarization. Musk is a polarizing public figure in ways that may or may not affect enterprise procurement decisions. Some customers may not want to deploy Tesla-branded humanoids in their factories.
- Internal organizational risk. Tesla has lost significant senior engineering talent in 2024–2025. The team building Optimus is smaller and less experienced than competitors’ teams in absolute terms.
Learn Our Proven AI Frameworks
Beginners in AI created 6 branded frameworks to help you master AI: STACK for prompting, BUILD for business, ADAPT for learning, THINK for decisions, CRAFT for content, and CRON for automation.
Frequently asked questions
Can I buy a Tesla Optimus?
Not yet. Optimus units have only been deployed internally at Tesla factories as of mid-2026. Tesla has not announced an external sales program. Pre-orders are not open. Pricing has been discussed publicly only as a long-term target (~$20,000), not as current pricing for a product you can buy now.
When will Optimus actually be ready?
Depends on definition. “Ready for internal Tesla factory use” — already happening at small scale. “Ready as an enterprise product for non-Tesla customers” — probably 2027–2028. “Ready for the consumer home market at $20K” — Musk says soon; sober analysts say 2030 at the earliest, possibly never at that price.
How does Optimus’s AI work?
Tesla has been deliberately quiet about the architecture. Strong signals point to a foundation-model approach related to but distinct from Tesla’s FSD system, running on the new AI5 chip in V3. The Dojo supercomputer, which Tesla has built to train FSD, is also believed to be used for Optimus training. Tesla controls more of the hardware-and-AI stack in-house than essentially any humanoid competitor.
Is Optimus part of Tesla’s $TSLA stock thesis?
Increasingly, yes. Musk and various Tesla bull analysts (Cathie Wood at ARK Invest most notably) include Optimus as a multi-trillion-dollar long-term opportunity in their TSLA valuation models. Whether to take those projections seriously is a question for individual investors, not this guide; the link between Optimus reaching scale and TSLA reaching ARK’s price targets is direct.
Will Optimus do household chores?
Eventually, in Musk’s vision. Realistically, factory work first (where the environment is structured), enterprise warehousing second (less structured but still industrial), and home use last (the hardest environment for any humanoid). Most credible robotics observers expect home humanoid deployment to remain difficult through at least 2028–2030 regardless of which company builds the robot.
How does Optimus compare to Boston Dynamics’ Atlas?
Different bets. Atlas is in serial production now, with deep engineering and a partnership with Google DeepMind for foundation models. Atlas’s customer pipeline is smaller but enterprise-validated. Optimus has bigger manufacturing-scale ambitions but no external customers yet. Atlas costs more per unit; Optimus aims for a much lower price point eventually. Both companies will likely succeed, in different parts of the market.
Sources
- Tesla — Tesla AI page (Optimus, FSD, Dojo)
- Electrek — Tesla Optimus V3 reveal pushed and Fremont line conversion
- Reuters Tesla coverage — Mainstream news on Tesla Optimus
- Tesla investor communications (Q1/Q2 2026 calls) — covering Optimus production timeline and AI5 silicon
- The Robot Report — Industry coverage of Optimus and competitors
- Humanoids Daily — Sector news