Who is Tesla’s biggest humanoid rival?

What does “Tesla’s biggest humanoid rival” mean?

Tesla’s biggest humanoid rival is the company most likely to challenge Optimus as a useful, general-purpose worker, not the company with the flashiest robot video.

Optimus is supposed to handle many jobs across factories, businesses, and eventually homes. That sets a much higher bar than walking well, lifting one type of box, or selling robots to universities. The winner needs capable hardware, broad AI, dependable work in real buildings, and a credible way to build thousands of machines at a price customers can justify.

With that test, the leaderboard splits. Agility has logged the most convincing paid work. AgiBot and Unitree ship the most robots and sell them cheaply. Boston Dynamics has the strongest body. Apptronik has gathered major partners across AI, manufacturing, and logistics. Figure has fewer proven working hours than Agility and far less manufacturing muscle than Tesla, but no other rival covers as much of the Optimus ambition. Each rival covers only part of the Optimus ambition, so we look at who is ahead in our humanoid robotics market report.

Is Tesla already leading the humanoid robot race?

Tesla leads the humanoid race in resources and possible future scale, but Optimus still trails several rivals in public evidence.

Tesla’s third-quarter results, published on October 21, 2026, moved Optimus from plan to early production. The Fremont line, built in the old Model S and Model X area, is listed as “early ramp”, and Tesla said the bots coming off it are its first production-quality robots. The Texas plant is listed as under construction. Musk tied the Fremont line to a goal of about one million robots a year, roughly 20,000 a week, and said reaching that rate would take at least a year. In July Tesla said the Texas line was being designed for ten million a year. Those are targets, not output. No other humanoid company plans on anything close to that level.

Tesla can also afford a slow start. Its quarterly filing shows about $45.4 billion in cash, cash equivalents, and short-term investments at the end of September 2026. Optimus can learn inside Tesla’s own plants, use Tesla-designed boards and AI chips, and draw on the company’s experience with motors, batteries, vision systems, and mass production.

The weak point has not moved. Even with the line running, Tesla has not shown V3 publicly, has given no count of robots built, and has no announced outside customer. The Information reported in September 2026 that Fremont was building several hundred Optimus robots a week, mostly for internal testing; Tesla’s results did not confirm that figure. On the October call Musk said most of the remaining design work is on the hand and that the ramp will be bumpy. He talked about an unscripted demonstration video in November or December 2026 and outside customers around late 2027; both are plans. Tesla has built the strongest launchpad, but the robot has not yet built the strongest record.

Which companies are serious Tesla Optimus rivals today?

Figure, Agility, Unitree, Apptronik, Boston Dynamics, and China’s leading manufacturers form the serious Optimus field, but each one threatens Tesla in a different way.

We compared them with a simple question: what could each one take away from Tesla? Figure could win the same general-purpose category. Agility could lock up industrial customers before Optimus arrives, now with public-market money behind it. Unitree could keep pushing hardware prices down. Apptronik could combine Google DeepMind’s AI with Jabil’s factories. Boston Dynamics could own the hardest factory jobs. AgiBot and UBTech show how fast China turns prototypes into product families and factory work.

Company	Main threat to Tesla Optimus	Strongest evidence now	Biggest gap
Figure AI	Competes for the same general-purpose vision	More than 350 Figure 03 robots delivered by late April 2026 (company-stated) and a long BMW run with Figure 02	Limited independent proof across many customers
Agility Robotics	Wins industrial work before Optimus is ready	65,000+ operating hours, nine committed customer sites, listing deal valued at $2.5 billion	$174,000 of sales in the first half of 2026 and a narrower task range
Unitree	Resets the market price for humanoid hardware	Listed in Shanghai in August 2026; robots sold from $4,900	Mostly a platform for developers, not a complete labor service
Apptronik	Builds a partner-led alternative to Tesla’s vertical model	Google DeepMind, Jabil, Mercedes-Benz, GXO; $935 million+ raised	Few published production or uptime figures
Boston Dynamics	Takes the premium industrial segment	Production Atlas and a training center inside Hyundai’s Georgia plant	No published throughput; main deployments planned for 2027-2028
AgiBot and UBTech	Turn China’s supply chain into a volume advantage	AgiBot about 8,400 units shipped in the first half of 2026 (estimate); UBTech 921 full-size humanoids sold (filed)	Product mixes and autonomy levels vary widely

Why is Figure AI the closest rival to Tesla Optimus?

Figure AI resembles Tesla Optimus more closely than any other humanoid company, from its general-purpose AI strategy to its factory-first path toward household robots.

Figure builds the body, hands, batteries, actuators, manufacturing software, and Helix AI around one robot. Figure 03 was designed for high-volume production and for tasks that need the whole body to work together. The company wants the same robot architecture to move from structured workplaces into homes. One robot architecture for factories and homes is an ambitious bet; we look at which technologies are gaining traction in our humanoid robots market report.

Figure is also copying Tesla’s habit of owning the inputs. In August 2026 it revealed Index, a platform that pays people to upload first-person video of everyday tasks; Figure says it has received more than 16 million uploads from 108 countries. In September 2026 it signed a deal with Nscale for up to 100,000 Nvidia GPUs in Texas starting in the second half of 2027, with an initial commitment of $3.5 billion. Up to 100,000 is a ceiling, not hardware in place.

The path looks very similar to Tesla’s: start where the environment is controlled, gather huge amounts of physical data, improve AI and hardware together, then widen the range of work. Apptronik follows a related path, but its AI and manufacturing strength come largely from partners. Agility chose a narrower commercial start. Unitree sells a flexible machine that customers often turn into a working system themselves. Figure is taking on the same broad product challenge as Tesla.

Has Figure done more real factory work than Tesla Optimus?

Yes. Figure has shown more measurable factory work at an outside company than Tesla Optimus, and BMW brought the newer Figure 03 back for a harder job.

Figure reported that Figure 02 worked ten-hour shifts, five days a week, over a ten-month deployment in Spartanburg. It moved more than 90,000 sheet-metal parts, ran for about 1,250 hours, and supported production of more than 30,000 BMW X3 vehicles. These are company figures, but ten months of logged work on a customer’s line says more than an edited demonstration. Figure retired its Figure 02 fleet in September 2026, so that record now belongs to a previous model.

The headline sounds broader than the job was. Figure handled one repetitive step, averaging roughly 72 parts per operating hour and about three parts per vehicle. Still, ten months on a live automotive line gave Figure something Tesla has not shown publicly: repeated exposure to breakdowns, awkward part positions, shift schedules, maintenance, and production pressure at an outside customer.

Since June 2026, BMW has been testing Figure 03 on a more variable sequencing task in Spartanburg. The robot picks thin metal parts that may have shifted, places them in the right slots, moves its feet while carrying them, and pulls a heavy cart. BMW has not published results from this second phase yet. Those numbers will matter more than another polished home video.

Do Figure’s 350+ robots matter against Tesla’s Optimus factory?

Yes, for now. Figure has counted real robots out of its factory, while Tesla has only described what its Optimus line should reach. If Tesla gets even a fraction of that target, it would leave Figure far behind.

Figure said in late April 2026 that BotQ had delivered more than 350 Figure 03 robots and had sped up its build pace from one robot a day to about one an hour. It has published no newer total, so 350 is the last published floor, not today’s fleet. Earlier, the company reported more than 9,000 actuators produced, a 99.3% first-pass yield on batteries, and a final robot first-pass yield above 80%. All of these are company figures, not audited ones, and Figure has not said how many of its robots work for customers rather than in its own training fleet.

They still describe a real manufacturing operation: supplier defects, inspection points, final tests, and parts that fail after repeated use. The pace needs context. One robot an hour is a demonstrated rate, not cumulative output; even without stopping, it equals 8,760 robots a year. Figure’s first line is designed for up to 12,000 a year, and the company has also talked about a 50,000-a-year target. Both are ceilings, not output.

Tesla’s one-million-a-year goal is roughly 83 times Figure’s first-line capacity. That compares two targets, not two outputs. On units actually counted, Figure has more than 350 and Tesla has disclosed none; the reported rate of several hundred a week at Fremont remains unconfirmed. Figure has the stronger proof of present production. Tesla has the far larger machine waiting behind Optimus.

Company	What is running or being built	Public scale figure	What remains unproved
Tesla	Fremont Optimus line in early ramp; Texas plant under construction	Goal of about 1 million robots a year, at least a year away (plan); no units disclosed	Stable production, yields, and customer deliveries
Figure	BotQ producing Figure 03	More than 350 delivered by late April 2026 (company-stated); first line designed for 12,000 a year	Sustained output near that annual capacity
Agility	RoboFab in Oregon	Peak capacity of 10,000 robots a year	Production near peak capacity
Boston Dynamics	Production Atlas in Boston	Hyundai plans a U.S. robot plant for up to 30,000 robots a year by 2028	How much of that capacity will be Atlas
Apptronik	Apollo production partnership with Jabil	No comparable unit figure published	Output, yields, and delivery pace

Does Helix 2.5 put Figure’s AI ahead of Tesla Optimus?

On public evidence, yes. Figure leads in broad humanoid manipulation, while Tesla may still have the stronger AI foundation behind closed doors.

Figure released Helix 2.5 in September 2026. It is a whole-body system pretrained on Index video. In 30 Bay Area homes the robot had never seen, on three jobs (tidying a living room, folding towels and making a bed), Figure says it completed 237 of 420 full-task trials, about 56%, against 9% for a version without the video pretraining. In August, Figure 03 also climbed up and down a 15-foot ladder on its own, a step toward industrial mezzanines.

A 56% success rate is a real step and also a clear limit. A robot that fails almost half of its household tasks is not ready to be sold as a home helper. Figure designed and ran the test itself, and company videos rarely show human preparation or remote help. Still, Figure published its trial count and failure rate, which is more than most rivals do.

Tesla reveals even less about Optimus. On its October call, Musk said Optimus would need an “Optimus Academy” of 10,000 to 20,000 robots learning through self-play. That is a proposed training fleet, not robots already built. Tesla’s strength lies around the robot: large-scale vision training, custom chips, data systems, and years of real-world driving data. That helps with perception, but it does not teach a hand to separate thin fabric or place a delicate part. On what outsiders can check today, Figure has shown the better humanoid-specific AI. We cannot tell whether Tesla’s internal system is better.

Could Apptronik become Tesla’s biggest industrial humanoid rival?

Apptronik could still overtake Figure as Tesla’s biggest industrial rival, but Apollo still lacks the hard operating numbers needed to claim that spot.

Apptronik has one of the strongest partner networks in the field. Google DeepMind is building the intelligence through Gemini Robotics, and in July 2026 those models gained whole-body control running on Apollo. Jabil can build Apollo at global manufacturing scale and test it in its own plants. Mercedes-Benz and GXO provide factory and logistics sites. It has raised more than $935 million, at a reported valuation near $5 billion in February 2026; we compare what companies are worth in our report covering humanoid robots.

The company’s most practical move is its Robot Park program. Its site in Austin covers nearly 90,000 square feet, and smaller setups run at partner and customer sites. Fleets of Apollo 2 robots collect data through a mix of autonomy and teleoperation across manufacturing, logistics, and retail tasks. A wheeled Apollo 2 can enter some workplaces under existing mobile-robot safety rules while the biped keeps improving. A new Apollo 3 is expected in 2027. In late October 2026 Apptronik published more detail on Apollo’s whole-body control, but no new fleet numbers.

That modular choice could help Apptronik sell useful systems sooner. It also makes the comparison with Optimus less direct, since Tesla wants one bipedal robot for the whole job. Apptronik’s missing piece is still public operating data: no robot count, uptime, parts moved, or customer hours on the level of Figure or Agility. The ingredients are excellent; the proof is still thin.

Is Unitree the biggest price threat to Tesla Optimus?

Unitree is Tesla’s biggest price threat because it sells humanoid hardware for less than many companies spend on one industrial robot arm, and it now has public-market money to keep doing it.

Unitree’s R1 Air starts at $4,900 and the standard R1 at $5,900, before tax and shipping. In September 2026 it launched a stronger G1+ at about $15,000. These are smaller, lighter robots than the full-size machines from Tesla, Figure, or Boston Dynamics. They still change the market by letting universities, startups, and corporate labs buy a walking humanoid instead of building one.

Unitree also proved the price attack can be profitable. Its prospectus showed 2025 revenue of about 1.7 billion yuan and a reported net profit of 278 million yuan, or about 591 million yuan excluding one-off items. It listed on Shanghai’s STAR Market on August 19, 2026, raising about 6.1 billion yuan, or roughly $900 million. Each robot it ships feeds an ecosystem of researchers, developers, spare parts, and shared know-how.

Buying a G1 or R1 still does not give a warehouse a dependable employee. Customers usually add software, task training, safety systems, and support. Unitree’s own chief executive said in August 2026 that its robots still work slower than people on factory lines. Unitree has made humanoid bodies cheap; it has not yet proved a complete autonomous labor product.

Are AgiBot and Unitree a bigger threat to Tesla than Figure?

As a group, yes, on cost and volume. As single product rivals, no. China is a bigger manufacturing threat to Optimus than any one Chinese company is a direct rival.

The volumes are now large. Smart Analytics Global, as reported by the South China Morning Post, estimates that AgiBot shipped about 8,400 humanoids in the first half of 2026, a 44% global share, against about 5,900 for Unitree, out of about 19,100 worldwide. Counterpoint gives higher figures for the same period, about 9,700 for AgiBot and more than 7,000 for Unitree, so the totals depend on who counts. AgiBot has begun a Hong Kong listing process and says its 2025 revenue topped 1 billion yuan. UBTech, whose results are filed in Hong Kong, sold 921 full-size humanoids in the first half of 2026 for about 590 million yuan, and keeps adding capacity for car factories.

The mixed product categories make those totals hard to compare with Optimus. A research robot, a reception robot, and an autonomous factory worker do not create the same value. China’s deeper advantage is the machinery around them: motors, batteries, power electronics, sensors, contract manufacturing, and a crowded field of domestic buyers and suppliers. China's depth in motors, batteries and contract manufacturing is a real edge; we weigh the likely future winners in our humanoid robots report.

That combination can push down the price of every key component, even if no Chinese brand becomes the global winner. Tesla faces Figure in the closest one-on-one contest, while China squeezes Tesla on cost, speed, and supply. Unitree’s fresh public-market capital, and AgiBot’s planned Hong Kong listing, make that squeeze harder than it was a year ago.

Why aren’t Agility Robotics or Boston Dynamics Tesla’s biggest rival?

Agility Robotics is ahead in paid industrial proof and Boston Dynamics is ahead in heavy-duty hardware, but neither is chasing the Optimus market as directly as Figure.

Agility is the strongest challenge to our Figure verdict. Digit has passed 65,000 operating hours across commitments at nine customer facilities, with work for GXO, Schaeffler, and Toyota Motor Manufacturing Canada. In September 2026 Agility unveiled Digit 5, which lifts up to 50 pounds and recharges in about nine minutes, with early-access deliveries expected in 2027. Agility says it has more than $300 million in committed Digit 5 orders, but they come from a single customer contract for 1,000 robots over three years, tied to milestones. That is a backlog, not revenue.

Money is the next test. Agility plans to go public by merging with Churchill Capital Corp XI in a deal valued at $2.5 billion; shareholders vote on November 2, 2026, and the deal could bring about $620 million in gross proceeds if no investors redeem. Its merger filings showed 2025 sales of about $1.78 million, a net loss of about $138 million, and a going-concern warning; first-half 2026 sales were about $174,000. Agility leads in proof of work, not in revenue scale.

Boston Dynamics threatens Tesla from the premium end. The production Atlas lifts up to 50 kilograms, swaps its own battery, and works across a wide temperature range. In September 2026 Boston Dynamics opened a training center inside Hyundai’s Metaplant in Georgia, where Atlas learns to sequence parts. Hyundai plans 25,000 Atlas robots across its plants and a U.S. factory for up to 30,000 robots a year by 2028. Those are plans, not deployments.

Figure stays closer to Tesla because it combines household manipulation, factory work, its own general-purpose AI, and its own robot factory in one plan. Agility has stronger proof of paid work. Boston Dynamics has more formidable industrial hardware. Figure is aiming at the same destination as Optimus.

Who is Tesla’s biggest humanoid rival now?

Figure AI is Tesla’s biggest humanoid rival, with Agility leading proof of paid work, UBTech leading filed humanoid sales, AgiBot and Unitree leading volume and price, and Boston Dynamics the toughest premium factory competitor.

Figure wins because it is strong across more parts of the race than any other rival. It has its own full-body AI, now tested in homes it had never seen. It has a robot factory that, by its own count, had delivered more than 350 units by late April 2026. It completed a long BMW production run with its previous model, and Figure 03 is back at BMW on a harder task. It is also lining up the data and compute that Tesla has always counted as its edge.

Tesla could still overwhelm every startup if Optimus works and its factory plans hold. It has more money, more compute, more production know-how, and its own plants as a planned first customer. For now, though, Tesla’s line is in early ramp with no disclosed output, and its claims are about what Optimus should become. Figure has published the most balanced evidence across AI, hardware, manufacturing, and outside work.

Agility deserves a sharper warning than it usually gets. Its hours, customers, orders, and coming listing give it the clearest route to an industrial business, although its revenue is still tiny. Apptronik could move up fast once it publishes fleet results. AgiBot and Unitree may set the economics of the whole category.

Our final judgment is firm: Figure is the closest rival to Tesla Optimus because both companies are trying to win the full general-purpose humanoid market. Agility is more proven at paid work right now, but Figure is competing for the bigger prize, and no rival matches its spread of evidence.