Top Humanoid Robot Stocks 2026: The Pure-Plays Are Here

For years, "humanoid robot stock" meant one thing: buy Tesla and squint.

Not anymore. In August 2026, Unitree became the first pure-play humanoid maker listed in mainland China, and its stock closed its first day up 460%, at a roughly $50 billion valuation, on an IPO oversubscribed more than 8,000 times.^[1] Agility Robotics is following via SPAC. The theme finally has real tickers.

Here's why the timing matters: humanoids stopped being demos. Unitree shipped over 5,500 humanoids in 2025 (more than anyone on Earth) at a 60% gross margin. Figure's robots work a BMW line. Agility's Digit has moved 100,000+ totes at GXO warehouses. Tesla runs Optimus in its own factories. TrendForce expects global shipments to blow past 50,000 units in 2026, a ~700% jump in one year.

Barclays pegs the whole category at ~$200B by 2035, which every bull cites and no bear believes. We think it plays out like early EVs: a handful of platform winners, a bloodbath of also-rans, and a component layer that gets paid no matter who wins. China already builds over 80% of the world's humanoids, so geography is a feature of this trade.

We're splitting humanoid robot stocks three ways: Pure-plays (the robot is the whole business), Integrated (giants with a humanoid program inside), and Picks-and-shovels (the joints, brains, and magnets every robot needs). Plus the private bellwethers you can't buy yet but should watch anyway.

These companies live or die by the robot. Highest upside, highest odds of a zero. Two trade in Asia today; the first US listing is en route.

Unitree RoboticsSHA: 688836

The Hangzhou company whose kung-fu-fighting androids went viral is now the poster child of China's robot push, and the numbers are shockingly un-startup-like. Unitree shipped 5,500+ humanoids in 2025 (plus 33,000 cumulative quadrupeds), booked ¥1.7B in revenue at a 60% gross margin, and is actually profitable. Its August 19 STAR Market debut raised ~$905M at ¥150.80/share and closed up 460%.

The catch: that pop stretched the valuation from ~$9B at pricing to ~$50B at the close. You're paying roughly 30x trailing revenue for the privilege of owning the leader, foreign access to STAR Market shares is limited (Stock Connect or China-focused ETFs like KOID are the practical routes), and the company sits under a US FCC Covered-List cloud.

UBTech RoboticsHKEX: 9880

The elder statesman of listed humanoids: UBTech has traded in Hong Kong since December 2023. Its Walker S2 industrial humanoid can swap its own battery in under three minutes and is deployed in "swarm" pilots at Chinese auto and electronics plants, with Beijing's explicit backing.

UBTech was the only clean listed pure-play for two and a half years, which earned it a scarcity premium. Now Unitree exists, out-ships it, and out-margins it. UBTech's edge is factory pilot depth with real manufacturers; the question is whether pilots convert to volume orders before the cash burn does.

Agility Robotics (NASDAQ: CCXI → AGLT, pending)

The one to watch if your brokerage account can't touch Shanghai. Agility announced a merger with Churchill Capital Corp XI in June 2026 at a ~$2.5B valuation. The deal is expected to close by end-2026, converting CCXI into AGLT and creating the first US-listed humanoid pure-play. Its Digit robot has moved 100,000+ totes at GXO under a robots-as-a-service contract, Amazon is both investor and pilot customer, and it runs the first dedicated humanoid factory in the US, with a "$300M+" order book disclosed.

Two caveats: over 90% of recent de-SPACs trade below their $10 reference price, and Amazon being both backer and main customer is concentration risk wearing a friendly face.

The robot program comes bundled with a much larger business. Less upside per robot, far less risk of ruin.

TeslaNASDAQ: TSLA

Tesla is the only major automaker building its humanoid in-house. Everyone else partnered: Hyundai→Boston Dynamics, BMW→Figure, Mercedes→Apptronik, Toyota→Agility. Optimus Gen 3 brings redesigned hands and better batteries, robots already do simple tasks in Tesla factories, and Musk targets $20,000–$30,000 per unit at scale, a price that, if hit, ends the debate.

But the program has scars: a mid-2025 production pause over overheating joint motors and weak hands, and a program-lead departure mid-ramp. Optimus is maybe 6–20% of the stock by most sum-of-the-parts math, so you're mostly buying a car and energy company with a robot lottery ticket stapled on. Great optionality, diluted exposure.

Hyundai MotorKRX: 005380

The stealth humanoid stock. Hyundai owns Boston Dynamics, the most famous robotics lab alive, whose fully electric Atlas debuted at CES 2026 with pricing reportedly above $150,000 and first deployments aimed at Hyundai's own Metaplant plus a Google DeepMind tie-up. You get three decades of legged-robot IP inside an automaker trading at automaker multiples. The flip side: Atlas revenue will be a rounding error on Hyundai's income statement for years.

Rainbow RoboticsKOSDAQ: 277810

Korea's humanoid bet, spun out of the KAIST team behind the HUBO robot. Samsung owns roughly 35% of Rainbow Robotics and has signaled robots are a core future business: a credible lab with a trillion-dollar manufacturing parent that can fund, build, and deploy at Samsung scale. The multiple is nosebleed relative to current revenue; you're paying for the Samsung relationship.

A gearbox from Harmonic Drive or a magnet from MP goes into a Tesla, a Figure, or a Unitree interchangeably. The suppliers win whichever robot wins.

NvidiaNASDAQ: NVDA

Every humanoid that isn't Optimus is likely trained on Nvidia hardware. Its Isaac simulation stack lets robots practice a task millions of times virtually before touching hardware, and Nvidia put money where the silicon is: it invested directly in Figure's $39B round. Humanoids are a rounding error on Nvidia's data-center revenue today; the point is you own the tollbooth if physical AI becomes the next compute wave, with a fortress business underneath.

Harmonic Drive SystemsTYO: 6324

The most concentrated component bet on the list. Harmonic Drive makes the precision strain-wave gears inside high-spec robot joint actuators. A single humanoid needs dozens of reducers, they're a large chunk of the bill of materials, and Harmonic is the default Tier-1 for Western and Japanese builds. The threat has a name: China's Leader Drive and Shuanghuan are localizing cost-competitive alternatives fast, and Chinese OEMs mostly buy domestic. Harmonic wins the Western ramp; it may lose the Chinese one.

MP MaterialsNYSE: MP

No rare-earth magnets, no actuators. No actuators, no robot. MP Materials runs the only integrated rare-earth mine-to-magnet operation in the Western Hemisphere, which makes it the geopolitical hedge of this watchlist. If US-China robot tensions escalate (see: FCC Covered List), every Western humanoid program needs a non-Chinese magnet supply, and there's basically one address. It's a commodity-adjacent business with commodity-adjacent volatility, so size accordingly.

You can't buy these yet, but watch them anyway. They set the sector's price signals, and they can sometimes prove technical capability months ahead of the public names.

Figure AI

The most valuable humanoid company on the planet, public or private: a $39B Series C in September 2025, backed by Nvidia, Microsoft, Bezos Expeditions, and Qualcomm, with Figure 03 robots working a BMW production pilot. Two things to track: secondary markets were quoting Figure below that $39B mark by August 2026, a read on sector risk appetite, and in July the company declared unauthorized share transfers void, naming four marketplaces. If someone offers you pre-IPO Figure stock, run.

1X Technologies

The home-robot bet. Norwegian-founded, OpenAI-backed, now headquartered in Palo Alto, 1X opened preorders for its NEO home humanoid in October 2025 at $20,000 (or $499/month) and runs a Hayward factory rated for 10,000 units a year. It also signed a deal with EQT to deploy up to 10,000 NEOs across EQT's portfolio companies through 2030, a quiet hedge into industrial work while the consumer story matures.

Two grains of salt. First, no verified customer delivery of a NEO had surfaced as of mid-2026, and complex chores still lean on human teleoperators. Second, the widely cited ~$10B valuation was a fundraising target that never closed; as of late August, SoftBank was reportedly in talks to buy a majority stake at roughly $6B. If that deal happens, it's a real price for the sector's home-robot thesis, and a markdown.

Apptronik

Austin-based Apptronik raised $935M in early 2026 at a ~$5.3B valuation with Google and Mercedes-Benz behind it, and its Apollo robot is aimed at Mercedes assembly work. It's the "automaker-partnered, reasonably priced" counterweight to Figure's mega-valuation. No IPO chatter yet, but with Unitree public and Agility closing, the window is open, and Apptronik is the most obvious next name through it.

Two failure modes matter most.

The robots don't earn their keep. Agility's Digit costs $10–$12/hour to operate today, and that number has to fall hard before humanoids beat human labor or purpose-built automation. Pilots are cheap PR; renewals are proof. Watch for pilot programs quietly converting to volume purchase orders (or quietly not). Roboticist Rodney Brooks calls the general-purpose humanoid vision "fantasy thinking," and the bear case has credentials.

Valuation air pocket. A 460% first-day pop, 8,000x oversubscription, and a BIS warning that AI asset valuations look overextended is a combustible mix. This sector already has a corpse to study: Vicarious Surgical, a 2021 robotics SPAC, voted on dissolution in July 2026 with common holders likely recovering nothing. Watch Unitree's H1 2026 numbers. The company itself guided to a year-on-year decline in adjusted net profit even as revenue grows, and margins compressing while stocks moon is how corrections start.

The next 12–18 months are catalyst-dense:

• The Agility/Churchill SPAC vote and AGLT listing (expected by end-2026): a live test of whether Western markets will pay Shanghai prices.
• Unitree's first earnings reports as a public company: the sector's benchmark tape from here on.
• Tesla's Optimus Gen 3 production ramp, and whether the hand-reliability problems stay fixed: the single biggest swing factor for the "robots at $25K" narrative.
• The reducer share fight: whether Leader Drive and Shuanghuan take meaningful share from Harmonic Drive.
• Washington's next move: whether robotics restrictions expand after tightening in August.
• The SoftBank/1X talks: a majority stake at ~$6B would put a real price on the home-robot thesis, below the numbers the sector has been quoting.

The shipping leaders (Unitree, and Agility if the SPAC closes clean) and the agnostic component layer (Nvidia, Harmonic, MP) are the highest-probability ways to be right. Everything else is a story stock until a customer renews.