Most Important Humanoid Robot Companies of 2026
Most lists of humanoid robot companies rank by funding, which tells you what investors believe rather than what exists. Others rank by demo quality, which tells you about video editing. This one ranks by evidence, using four criteria: verified commercial deployment, manufacturing throughput, capital position, and software depth. Where the evidence is thin, we say so. Where sources disagree—and on this topic they disagree constantly—we give the range instead of picking the convenient number. Two disclosures before the list. This reflects late August 2026, and the field moves monthly. And 3GEN Robotics does not manufacture humanoid robots, holds no position in any company named, and has no commercial relationship with any of them.
Three things that changed in 2026
The United States closed its market to foreign-built humanoids. On July 28, 2026, the FCC's Public Safety and Homeland Security Bureau issued Public Notice DA 26-786, adding foreign-produced advanced robotic devices to the Covered List [9]. The determination came from a White House–convened interagency body finding such devices pose cybersecurity and national security risks [8][10]. It's a forward-looking gate rather than a recall: it applies to new device models seeking FCC authorization, doesn't claw back existing units, and doesn't stop the sale of models authorized before that date [7]. Exemptions run through a conditional approval process, and federal purchases are carved out [7]. China holds roughly 85 percent of the global market [8], so the effect is substantial, and as one analyst noted, the best robots are next year's robots, so being frozen at already-certified models constrains foreign share more than current numbers suggest [7].
Shipments and valuations fully decoupled. The volume leaders and the valuation leaders are different companies on different continents, and neither ranking predicts the other.
A shared software layer emerged. NVIDIA's Isaac GR00T models are used across Agility, Boston Dynamics, NEURA, Mentee, and 1X, with releases validated on Unitree and AgiBot hardware [17]. Hardware differentiates; intelligence converges.
Tier 1: Verified commercial deployment
Paid work in customer facilities with published operating numbers. The smallest and most meaningful group here.
1. Agility Robotics — agilityrobotics.com
Pittsburgh, US · Founded 2015 · Digit
The clearest commercial record in the category. Digit operates under a multiyear robots-as-a-service contract at GXO, where Agility reported moving more than 100,000 totes across over a year of continuous full-time deployment [15][13], and now runs paid customer workflows at roughly nine commercial sites [19]. Commercial relationships extend to Toyota Motor Manufacturing Canada and Mercado Libre [15]. Its RoboFab facility scales past 10,000 units per year [17].
Why it's first: the company most credibly generating recurring revenue from productive work rather than pilots. It's also taking the first pure-play humanoid path to U.S. public markets, merging with Michael Klein's Churchill Capital Corp XI at roughly $2.5 billion [16][25].
Worth noting: CEO Peggy Johnson has been conspicuously measured about home robots and near-term capability while peers promise more [25]. In a category where forecasting discipline is rare, that's a signal rather than a weakness.
The caveat: Digit's proven envelope is narrow tote and material movement in structured warehouse flows. The focus is why the track record exists, but it isn't general-purpose capability.
2. Figure AI — figure.ai
San Jose, US · Founded 2022 · Figure 03
The most valuable pure-play humanoid company in the West and the holder of the industry's most granular published operating data: an eleven-month pilot at BMW Group Plant Spartanburg running ten-hour weekday shifts, more than 1,250 runtime hours, and more than 90,000 sheet-metal parts loaded, contributing to over 30,000 vehicles [11][14]. Its BotQ facility is specified at up to 12,000 units per year [17]. Figure develops its own vision-language-action model, Helix, now in a second generation capable of controlling two robots simultaneously [26]—one of the few humanoid companies with a genuinely proprietary AI stack rather than a licensed one.
Why it's here: top-tier capital paired with real, measured production work, plus in-house models.
The caveat, and it's important: the $39 billion valuation from the $1 billion Series C was self-reported [25]. Widely circulated claims of ten thousand Figure robots in the field don't trace to the company [11], and reported cumulative production is in the low thousands [22]. Valuation and deployment scale are further apart here than anywhere else on this list.
3. Apptronik — apptronik.com
Austin, US · Founded 2016 · Apollo
Deeper engineering roots than most: the team's work dates to a NASA contract that produced the Valkyrie humanoid, and Apollo descends from that lineage [113]. Apptronik closed a $935 million round valuing the company above $5.5 billion [25]. Apollo runs intralogistics at Mercedes-Benz — moving assembly kits, performing initial quality checks — with Mercedes holding an equity stake [13][16]. Contract manufacturer Jabil both builds Apollo and tests it on sorting, kitting, inspection, and subassembly [15], which is the most credible route to automotive-grade volume of anyone here.
Why it's here: the deepest integration into an existing industrial customer's operations, plus a manufacturing partner that already builds at scale.
The caveat: analysts who tier claims carefully classify the Mercedes program as evaluation rather than scaled autonomous production [12].
Tier 2: The volume leaders
Chinese manufacturers ship the overwhelming majority of the world's humanoids. Volume isn't validation, but manufacturing learning curves are real.
4. AgiBot (Zhiyuan Robotics) — agibot.com
Shanghai, CN · Founded 2023 · A-series, X-series, G-series
The world's largest humanoid vendor by shipments in the first half of 2026: roughly 8,400 units and 44 percent global share, up 562 percent year over year, across full-size bipedal, compact, and wheeled lines [1]. Revenue exceeded €130 million last year, with 15,000 cumulative units delivered by June 2026 [6]. Founded by former Huawei engineers, mass production began December 2024, and a Hong Kong listing process launched in July 2026 targeting up to €5.6 billion [6].
Why it's here: it outsells everyone, and the product breadth—bipedal, compact, wheeled—suggests a company optimizing for markets rather than demos.
The caveat: unit counts include research, education, and data-collection buyers, and no comparable profitability figures are public [6].
5. Unitree Robotics — unitree.com
Hangzhou, CN · G1, H2, R1
The price disruptor and now public. Unitree priced its Shanghai STAR Market IPO at about $9 billion in August 2026, raised roughly $905 million, then closed its first session up around 460 percent — briefly touching a valuation near $66 billion [3][4][5]. Unusually, it's profitable: roughly €220 million revenue and €36 million net profit last year, with 5,632 humanoids and more than 33,000 quadrupeds sold between 2023 and 2025 [6]. First-half 2026 shipments were about 5,900 units, a 31 percent global share [1]. DeepSeek participated in the IPO; Tencent was a prior investor [3].
Why it's here: the only company here demonstrating that humanoids can be a profitable hardware business today.
The caveat: shipments concentrate in the G1 for education, research, and performance rather than industrial work [1]. One analyst framed the challenge precisely—converting hardware and manufacturing leadership into application and workflow leadership [4].
6. UBTech Robotics — ubtrobot.com
Shenzhen, CN · Founded 2012 · Walker S2
The first humanoid company to list on a major exchange, going public in Hong Kong in late 2023 [5]. Walker S2 targets industrial deployment with a stated production plan of 5,000 units per year [17].
Why it's here: the longest operating history of any pure-play humanoid company, plus an established multi-line robotics business beyond humanoids.
The caveat: public deployment evidence is thinner than the order announcements suggest, and it now faces the US import gate [7].
Tier 3: Heritage and scale ambition
7. Boston Dynamics — bostondynamics.com
Waltham, US (Hyundai-owned) · Atlas
Building legged robots since 1992, long before this funding cycle [16]. The electric Atlas—56 degrees of freedom, 50 kg lift—replaced the hydraulic research platform, with a product version unveiled in January 2026 [12][16]. Hyundai became the sole owner in July 2026, buying SoftBank's remaining stake at roughly a $3.3 billion valuation [16]. At CES 2026, Google DeepMind and Boston Dynamics announced a partnership running Gemini Robotics models on Atlas and Spot, with testing planned at Hyundai plants [17].
Why it's here: unmatched locomotion and controls engineering, a captive industrial customer, and a frontier-lab AI partner.
The caveat: the most-watched company with the least deployment evidence. Atlas work at Hyundai's Georgia plant is targeted for 2028 [16], and a scheduled fleet is announced until shipment and operation are confirmed [12].
8. Tesla — tesla.com/AI
Austin, US · Optimus
Here because manufacturing ambition at Tesla's scale would reshape the category if realized, and no higher because the evidence isn't there. As of mid-July 2026, Optimus production at Fremont had not begun; the Q2 report contained no Optimus figures, and Tesla has never published a production count [11]. Analysts counted eight production milestones delayed or revised before their original dates [11]. Where Optimus is deployed internally, it's described as primarily generating training data rather than productive output [13].
Why it's here: vertical integration, in-house actuator development, and automotive manufacturing capability no startup can match.
The caveat: treat all near-term Optimus dates as probabilities [11].
Tier 4: The consumer bet
9. 1X Technologies — 1x.tech
Moss, Norway / Palo Alto, US · Founded 2014 · NEO
The only company seriously attempting the home market. NEO is priced at $20,000 or $499 per month, and 1X reported 10,000 pre-orders within five days of opening them [16][21]. It opened a 58,000-square-foot Hayward, California factory in April 2026 targeting 100,000 units per year by 2027 [16]. Reported valuation is around $10 billion [20].
Why it's here: if domestic humanoids arrive this decade, 1X is furthest along, and consumer volume would fund a data flywheel no industrial player could match.
The caveat: the hardest problem in the category. Early home systems operate in tidy, well-defined homes under tight safety rules with teleoperation backup, and no public demonstration has shown a full household task completed autonomously start to finish on novel inputs [23]. Household capability remains narrow — meal preparation means microwaving and pouring [21].
Tier 5: Europe
10. NEURA Robotics — neura-robotics.com
Metzingen, Germany · 4NE-1
Europe's best-capitalized entrant, having raised up to $1.4 billion led by Tether at roughly a $7 billion valuation [17][16]. Around 1,200 employees, a stated goal of five million robots by 2030, and a 4NE-1 Mini at €19,999 [16][21]. An early Isaac GR00T adopter [17].
Why it's here: the only European company with capital comparable to US leaders and a cognitive-robotics platform predating its humanoid work.
The caveat: deployment evidence lags the funding, and the shipment targets are extraordinarily aggressive.
11. PAL Robotics — pal-robotics.com
Barcelona, Spain · TALOS, REEM, ARI
Building bipedal humanoids since 2004 — longer than most of this list has existed. Its platforms are widely used across European research institutions, and it operates in specialized markets with regional strength [19].
Why it's here: the research platform installed base matters. Companies whose robots train a generation of roboticists shape the field disproportionately, and surviving multiple hype cycles is itself a signal.
The caveat: research and specialized deployment rather than industrial scale, not competing on volume or capital.
Tier 6: The company that isn't a manufacturer
12. NVIDIA — nvidia.com
Santa Clara, US · Isaac GR00T, Isaac Lab, Jetson Thor
Not a humanoid manufacturer, and arguably more important to the category than most companies that are. Isaac GR00T N1, released March 2025, was the first open foundation model for humanoid robots; adopters now include Agility, Boston Dynamics, NEURA, Mentee, and 1X, with releases validated on Unitree G1 and AgiBot hardware [17]. Isaac Lab's GPU-parallel simulation compresses months of training into hours [17], and the GR00T synthetic pipeline generated 780,000 trajectories—roughly nine months of human demonstration—in 11 hours of compute [17].
Why it's here: when most serious humanoid programs run on one company's foundation models, simulation stack, and edge silicon, that company is a structural participant.
The caveat: shared infrastructure compresses hardware differentiation. If everyone's intelligence comes from the same place, a durable advantage moves to proprietary data and deployment capability, the argument we make in our post on automated training data collection.
Which company for which use case
The most common question this list gets asked, answered directly. Not endorsements — starting points for a shortlist.
Warehouse tote and material movement: Agility is the only company with a published multi-year commercial record in this exact workflow [15].
Automotive intralogistics and line-side kitting: Apptronik and Figure both have automotive OEM programs with measured hours [11][13][14].
Research and university platforms: Unitree on price and availability [21]; PAL Robotics for European institutions and mature ROS integration.
Data collection for your own model development: Unitree G1 and Apptronik's data-collection configurations are both explicitly positioned for this [12][21].
Home and domestic: 1X is the only serious answer, and the honest one is not yet [23].
Precision assembly, welding, or high-speed work: none of them. This remains firmly with traditional fixed industrial arms [13], and any vendor suggesting otherwise is selling a roadmap.
The incumbents circling
Worth watching separately because large companies entering change the category's economics rather than its rankings.
Hyundai owns Boston Dynamics outright as of July 2026 and is the intended first customer [16]. Google DeepMind supplies models to Boston Dynamics and partners with Apptronik [15][17]. Samsung holds a position in Rainbow Robotics, and LG, XPeng, and Xiaomi all have humanoid programs [21]. Amazon absorbed Fauna Robotics in March 2026 and tests Digit in fulfillment centers [17]. The pattern: automakers and logistics operators are buying access to the technology rather than waiting to purchase the product. For a manufacturer evaluating this space, that's the more telling signal: the people closest to the work are hedging by acquiring capability.
Also worth watching
China: Fourier Intelligence (fourierintelligence.com, GR series, Saudi sovereign wealth relationship); AI2 Robotics (Shenzhen, wheeled humanoids, raised roughly $735 million at nearly $3 billion in mid-2026 [25]); EngineAI; Galbot; Kepler; RobotEra; LimX Dynamics; MagicLab; and Astribot. Several closed nine-figure rounds within twelve months [17].
Elsewhere: Mentee Robotics (Israel), Kinisi Robotics and Humanoid (UK), Dexmate (notable for actually listing a commercially available product [18]), Engineered Arts (Ameca), Wandercraft (France), and Clone Robotics (musculoskeletal designs).
Software-only: Physical Intelligence builds robotics foundation models without hardware—increasingly a viable position and the one Sanctuary AI moved toward.
Components: Japanese precision suppliers, particularly Harmonic Drive, sit underneath much of the industry [21]. Actuators have become the strategic in-house capability that used to be a commodity purchase.
The shakeout nobody mentions
Lists like this imply a rising tide. The 2025–26 record says otherwise: K-Scale Labs shut down in November 2025; Cartwheel Robotics in February 2026; Amazon absorbed Fauna Robotics in March 2026; and—most tellingly—Sanctuary AI, long a fixture on lists like this one, pivoted from hardware to software in June 2026 [17]. Sit with that last one. Sanctuary had a distinctive general-purpose humanoid, a serious cognitive architecture, and a Magna partnership. It concluded the defensible position was above the hardware. With more than 140 companies manufacturing humanoids globally [21] and roughly $8.6 billion raised in 2026 alone [22], consolidation is arithmetic, not speculation.
How to read this market
Ask for verified operating hours, not units shipped. Shipment counts include research buyers, entertainment, and state-backed data-collection purchases [2]. Hours in a customer facility is the number that means something [22].
Distinguish the four claim tiers. Demo, announced, piloted, and verified operational. A scheduled fleet is announced until operation is confirmed [12], and corporate ownership isn't customer validation.
Check who published the number. Many widely cited figures don't trace to the companies they describe [11], and at least one headline valuation is self-reported [25].
Watch the reliability gap, not the capability gap. Bessemer characterized this as robotics' "GPT-2.5 moment"—real capability and real scaling, with a wide distance to the 99.9 percent reliability production demands [24]. That last stretch is where timelines slip.
Remember the tasks are narrow. Material handling, transfer, kitting, machine tending, and simple assembly support [13].
Frequently asked questions
Who is the leading humanoid robot company in 2026? It depends on the metric, which is why this list is tiered rather than ranked. AgiBot leads on shipments, Unitree on profitability and public market value, Figure on private valuation, and Agility on verified commercial deployment. No company leads on all four [12][18].
Which humanoid robots can I actually buy? Very few are generally available. Unitree publishes real prices and ships at volume, 1X takes NEO orders, NEURA lists 4NE-1 Mini pricing, and Dexmate lists a commercially available product [18][21]. Most Western industrial humanoids are sold through pilot agreements or robots-as-a-service contracts rather than off a price list.
Can US companies still buy Chinese humanoid robots? The FCC ban applies to new device models seeking authorization; it doesn't affect already-owned units or models authorized before July 28, 2026 [7]. Exemptions exist through conditional approval, and federal purchases are exempt [7]. Anyone with an active procurement should get current legal advice—this is recent and evolving.
Are any humanoid companies profitable? Unitree reported roughly €36 million net profit last year [6]. That appears to be the exception; most companies here are venture- or parent-funded and pre-profit. Revenue across the sector remains minimal, with much of the activity in letters of intent and pilots rather than recurring contracts [27].
How much does a humanoid robot cost? Manufacturer-published prices span roughly $13,500 for a Unitree G1 to about $100,000 for an H2 Plus [21], with 1X's NEO at $20,000 and NEURA's 4NE-1 Mini at €19,999 [16][21]. Most industrial humanoids have no public price and are sold as service contracts.
Should my facility buy a humanoid robot? For most manufacturers, not yet, and possibly not ever in preference to a better-suited form factor. The proven task envelope is narrow, and a fixed arm or AMR is often faster, cheaper, and more reliable for the same work. The exception is genuinely human-shaped environments—stairs, human-scale workstations, equipment designed for people—where the form factor earns its cost.
The variable that outlasts the ranking
Read this as a hardware ranking, and it's obsolete within a year. Read it for structure, and something more durable appears: the shipment leaders don't have the deployment records, the shipment leaders don't have the volume, and a growing share of everyone's intelligence comes from the same foundation models. Which puts a durable advantage somewhere other than the robot. It sits with whoever can teach a machine a new task quickly, capture what happens when it fails, and push the improvement back out to a fleet. That's a data and deployment problem, and it applies identically to a bipedal robot, a cobot, and the fixed arm already on your floor. It's why 3GEN builds Bellwether — a deployment and operations platform — rather than a humanoid. When general-purpose hardware matures, the constraint moves entirely to who can deploy and improve it. That constraint exists today. If you have a task on your floor you want automated, request a demo or reach us at dan@3genrobotics.com.
Notes
Bellwether is in alpha. Public demo material illustrates the pipeline conceptually; screens, flow, and terminology are subject to change in the shipped product. Performance figures cited above are drawn from published third-party research and vendor documentation, not from 3GEN internal benchmarks. Figures reflect information available as of August 29, 2026 and change rapidly. Valuations, shipment counts, and funding totals come from secondary reporting and company statements rather than audited disclosures, and sources frequently disagree: Unitree's first-day close was reported at valuations between roughly $50 billion and $66 billion depending on whether the closing or intraday price was used [4][5]. Figure AI's $39 billion valuation was self-reported [25]. Shipment estimates from Counterpoint Research and Smart Analytics Global differ in both magnitude and application mix [1][2]. Company inclusion reflects editorial judgment against the four stated criteria, not a quantitative score, and reasonable analysts rank this field differently. The use-case section is a starting point for a shortlist, not a recommendation; no company named has been evaluated by 3GEN in a customer deployment. 3GEN Robotics does not manufacture humanoid robots, holds no financial position in any company named, and has no commercial relationship with any of them. Regulatory information regarding FCC Public Notice DA 26-786 is general context, not legal advice.
Changelog
August 29, 2026 — Initial publication. All company URLs verified. Includes FCC import ban (July 28, 2026), Unitree STAR Market listing (August 19, 2026), and AgiBot H1 shipment leadership.
References
[1] South China Morning Post. "AgiBot Overtakes Unitree as Top Global Humanoid Robot Vendor in First Half Amid IPO Push." August 2026, citing Smart Analytics Global. https://www.scmp.com/tech/tech-trends/article/3363544/agibot-overtakes-unitree-top-global-humanoid-robot-vendor-first-half-amid-ipo-push
[2] Koetsier, John. "Humanoid Robot Shipments Up 300%: Up To 30,000 So Far In 2026." Forbes, August 20, 2026, citing Counterpoint Research. https://www.forbes.com/sites/johnkoetsier/2026/08/20/humanoid-robot-shipments-up-300-up-to-30000-so-far-in-2026/
[3] CNBC. "Chinese Humanoid Robot Maker Unitree Prices IPO at $9 Billion Valuation." August 6, 2026. https://www.cnbc.com/2026/08/06/chinese-humanoid-robot-maker-unitree-prices-ipo-at-9-billion-valuation.html
[4] Al Jazeera. "China's Unitree Soars in Market Debut as Investors Bet on Humanoid Robots." August 19, 2026. https://www.aljazeera.com/economy/2026/8/19/chinas-unitree-soars-in-market-debut-as-investors-bet-on-humanoid-robots
[5] Fortune. "Unitree, Famous for Its Dancing Robots, Surges by 460% on Its Trading Debut." August 19, 2026. https://fortune.com/2026/08/19/unitree-china-dancing-robots-ipo-trading-surge-valuation/
[6] China Business Spotlight. "85% Market Share: China's Humanoid Takeover." August 2026. https://chinabusinessspotlight.substack.com/p/85-market-share-chinas-humanoid-takeover
[7] Koetsier, John. "United States Bans Chinese Humanoid and Quadruped Robots, Citing National Security." Forbes, July 28, 2026. https://www.forbes.com/sites/johnkoetsier/2026/07/28/united-states-bans-chinese-humanoid--quadruped-robots-citing-national-security/
[8] CBS News / Associated Press. "Humanoid Robot Imports Banned as U.S. Targets Chinese Products Over National Security Concerns." July 29, 2026. https://www.cbsnews.com/news/humanoid-robots-imports-us-ban-china-national-security-concerns/
[9] There's A Robot For That. "Humanoid Robot Import Ban: Complete 2026 FCC Rule Breakdown." July 2026, citing FCC Public Notice DA 26-786. https://theresarobotforthat.com/blog/humanoid-robot-import-ban/
[10] CNN Business. "China's Humanoid Robots Have Been Taking Over the Global Market. Now the US Is Banning Them." July 29, 2026. https://www.cnn.com/2026/07/29/tech/us-china-robot-ban-intl-hnk
[11] Technology.org. "Humanoid Robots in 2026: What Is Actually Deployed." July 18, 2026. https://www.technology.org/2026/07/18/humanoid-robots-in-2026-what-is-actually-deployed/
[12] TheMimic. "Every Humanoid Robot Company in 2026: The Complete Guide." July 2026. https://themimic.io/articles/humanoid-robot-companies-2026
[13] iFactory. "Humanoid & Quadruped Robots for Manufacturing Plants 2026: Complete Industry Deployment Guide." May 2026. https://ifactoryapp.com/industries/manufacturing-plant/humanoid-quadruped-robots-manufacturing-plant-2026-guide
[14] New Market Pitch. "Humanoid Robots in Factories: For When?" June 2026. https://newmarketpitch.com/blogs/news/humanoid-robotics-factories-when
[15] The AI Insider. "The State of Humanoid Robotics in 2026: Trends, Challenges and Opportunities." August 21, 2026. https://theaiinsider.tech/2026/08/21/the-state-of-humanoid-robotics-in-2026-trends-challenges-and-opportunities/
[16] Tech Funding News. "Top 10 Humanoid Robot Startups to Watch in 2026, Ranked by Total Funding." July 29, 2026. https://techfundingnews.com/top-humanoid-robot-startups-2026-funding/
[17] RoboZaps. "38 Best Humanoid Robots in 2026 (Evidence-Ranked)." July 2026. https://blog.robozaps.com/b/best-humanoid-robots
[18] RoboZaps. "53 Humanoid Robot Companies Compared (2026 Directory)." July 2026. https://blog.robozaps.com/b/humanoid-robot-companies
[19] The Robotic Life. "Top 10 Humanoid Robot Startups Pulling Ahead in 2026." July 11, 2026. https://theroboticlife.com/top-10-humanoid-robot-startups-2026/
[20] Value Add VC. "Humanoid Robots 2026: Figure vs Apptronik vs 1X vs Tesla Optimus vs Unitree." June 2026. https://valueaddvc.com/blog/humanoid-robots-in-2026-figure-apptronik-1x-and-tesla-optimus-compared
[21] There's A Robot For That. "Humanoid Robot Companies 2026: 140 Makers Compared." August 2026. https://theresarobotforthat.com/blog/humanoid-robot-companies-2026-complete-guide/
[22] Value Add Pulse. "Humanoid Robots: Follow the Manufacturing Data, Not the Demos." August 2026. https://valueaddvc.com/pulse/humanoid-robot-manufacturing-data-analysis-2026
[23] Winss Solutions. "Innovative Humanoid Robots in 2025–2026: Reality or Hype?" December 2025. https://www.winssolutions.org/humanoid-robots-2025-2026-reality-hype/
[24] Chourasia, Mehul. "The Truth About Humanoid Robots in 2026." May 2026, citing Bessemer Venture Partners. https://medium.com/@mehul.chourasia28/the-truth-about-humanoid-robots-in-2026-3ae82e9061b1
[25] TechCrunch. "This Humanoid Robotics Company Is Going Public, but Its CEO Isn't Promising a Robot in Your Home Anytime Soon." July 5, 2026. https://techcrunch.com/2026/07/05/this-humanoid-robotics-company-is-going-public-but-its-ceo-isnt-promising-a-robot-in-your-home-anytime-soon/
[26] Wikipedia. "Figure AI." Accessed August 2026. https://en.wikipedia.org/wiki/Figure_AI
[27] Sacra. "Figure vs Apptronik vs Agility Robotics." January 2026. https://sacra.com/research/figure-vs-apptronik-vs-agility-robotics/