The humanoid robot market is worth roughly $2.9–6.2 billion in 2026 depending on the research firm, and Wall Street expects it to reach $5 trillion (Morgan Stanley) to $7 trillion (Citi) by 2050. Between 13,000 and 18,000 humanoid robots were sold worldwide in 2025 (Omdia/IDC), and Bank of America expects 2026 shipments to hit 90,000 units — a 4.5x jump in a single year. Humanoid startups raised $6.1 billion across 139 deals in 2025, up more than 300% (PitchBook), while a China-built humanoid's bill of materials fell to about $35,000. We aggregated data from Goldman Sachs Research, Morgan Stanley, Citi, Bank of America Institute, TrendForce, IFR, Omdia/IDC, PitchBook, Crunchbase, and official company disclosures — every stat below links to its primary source.
Key Takeaways
- Wall Street's 2050 humanoid market forecasts span $5 trillion (Morgan Stanley, The Humanoid Economy 2025) to $7 trillion (Citi Institute, The Rise of AI Robots 2024).
- Goldman Sachs projects a $38 billion humanoid TAM and 1.4 million annual shipments by 2035 — a 6x upward revision (Goldman Sachs Research, 2024).
- Global humanoid shipments were 13,000–18,000 units in 2025, roughly 90% of them in China (Omdia/IDC, 2026).
- BofA Global Research forecasts shipments rising from 20,000 (2025) to 90,000 (2026) and 10 million by 2035 — an 86% CAGR (BofA Institute, Physical AI Part 2, 2026).
- TrendForce forecasts 2026 global shipments above 50,000 units, over 700% year-on-year growth (TrendForce, December 2025).
- Unitree shipped 5,500+ humanoids in 2025 — 32.4% of global unit shipments, more than any other maker (Unitree IPO prospectus, 2026).
- Figure has delivered 350+ Figure 03 robots and raised BotQ output from one robot per day to one per hour in under 120 days (Figure, April 2026).
- A China-built humanoid's bill of materials was about $35,000 in 2025 and is projected to fall below $17,000 by 2030 (BofA Global Research, 2026).
- Humanoid robotics startups raised $6.1 billion across 139 deals in 2025, a 300%+ increase over 2024's $1.5 billion (PitchBook, 2026).
- Figure's Series C topped $1 billion at a $39 billion post-money valuation (Figure, September 2025).
- Chinese companies account for 63% of the global humanoid robot supply chain (Morgan Stanley, The Humanoid 100, 2025).
- Bank of America projects 3 billion humanoid robots in operation by 2060 — 62% of them in households (BofA Institute, 2026).
01Market Size Forecasts: $2.9 Billion Today, $5–7 Trillion by 2050
The spread in humanoid market forecasts is itself the story: research firms disagree by more than 2x on where the market stands in 2026, and the banks' 2050 numbers differ by $2 trillion. The disagreement is definitional — hardware-only versus hardware plus software, data, and services — not a dispute about direction.
Every firm that revised its humanoid forecast between 2024 and 2026 revised it upward. Goldman Sachs lifted its 2035 TAM sixfold in a single revision after AI progress and cost declines beat its model.
| Metric | Value | Source |
|---|---|---|
| Humanoid robot TAM by 2035 | $38B | Goldman Sachs Research 2024 |
| Humanoid market by 2050 (incl. software & services) | $5T ($4.7T hardware) | Morgan Stanley 2025 |
| Humanoid market by 2050 | $7T | Citi Institute 2024 |
| Market size 2024 → 2030 | $1.55B → $4.04B | Grand View Research 2025 |
| Market size 2025 → 2030 (39.2% CAGR) | $2.92B → $15.26B | MarketsandMarkets 2025 |
| Market size 2026 → 2034 (50.6% CAGR) | $6.24B → $165.13B | Fortune Business Insights 2025 |
| Market size 2026 | $3.93B | Mordor Intelligence 2026 |
| Hardware share of market 2026 / Asia-Pacific share | 65–75% / >50% | MarketsandMarkets 2025 |
Definitions differ: Goldman and the research firms size hardware revenue; Morgan Stanley and Citi include software, data, services, and secondary industries. Goldman's figures date to its February 2024 revision — the firm's most recent public humanoid TAM.
02Unit Shipments and Installed Base: How Many Humanoid Robots Exist?
Roughly 13,000–18,000 humanoids were sold in all of 2025 — fewer robots than a single large auto plant builds cars in a week. The forecasts, not the current base, are what moves capital: every major model has 2026 as the inflection year.
Shipment growth is gated less by demand than by manufacturing yield and the supply of real-world training data — the same constraint driving the numbers in our Vision-Language-Action model statistics. Beyond humanoids, Citi's model puts the broader AI-robot population at 1.3 billion by 2035 and 4 billion by 2050 (Citi Institute, The Rise of AI Robots 2024).
| Year | Annual shipments |
|---|---|
| 2024 | 3,000 |
| 2025 | 20,000 |
| 2026 | 90,000 |
| 2027 | 290,000 |
| 2028 | 500,000 |
| 2029 | 800,000 |
| 2030 | 1,200,000 |
| Metric | Value | Source |
|---|---|---|
| Humanoids sold globally, 2025 | 13,000–18,000 | Omdia / IDC via Rest of World 2026 |
| Annual shipments 2024 → 2025 → 2026E | 3,000 → 20,000 → 90,000 | BofA Institute 2026 |
| Annual shipments 2030E / 2035E (86% CAGR) | 1.2M / 10M | BofA Institute 2026 |
| 2026 global shipment forecast | >50,000 (+700% YoY) | TrendForce, Dec 2025 |
| Shipments by 2030 / 2035 | 250,000+ / 1.4M | Goldman Sachs Research 2024 |
| Units in service 2035 / deployed 2050 | ~13M / 1B+ | Morgan Stanley 2025 |
| Humanoids working by 2050 | ~650M | Citi Institute 2024 |
| Installed base 2040 / 2060 | 300M / 3B | BofA Institute 2026 |
Shipments vs installed base matters: Goldman's 1.4M and BofA's 10M are annual shipments in 2035; Morgan Stanley's 13M is cumulative units in service. They are not directly comparable.
03Production Targets: Figure, Tesla, Unitree, AgiBot, Agility, UBTech
Announced factory capacity already exceeds 100,000 units a year across the major players — 5x what the entire industry shipped in 2025. The gap between capacity claims and delivered units is the metric to watch: Unitree and AgiBot delivered four-digit unit counts, while Tesla, the most-cited name in the category, sold on the order of 150 robots in 2025.
| Company metric | Value | Source |
|---|---|---|
| Unitree 2025 shipments / global share | 5,500+ / 32.4% | Unitree prospectus via Rest of World 2026 |
| Unitree planned annual humanoid capacity | 75,000 | TrendForce, Apr 2026 |
| AgiBot cumulative output (Mar 2026) / built in 2025 | 10,000 / 1,000 | TrendForce, Apr 2026 |
| Figure 03 delivered / production rate | 350+ / 1 per hour | Figure, Apr 2026 |
| Figure BotQ first-gen line capacity | 12,000/year | Figure 2025 |
| Tesla Optimus sold 2025 / 2026 output target* | ~150 / 50K–100K | Omdia via Rest of World 2026 |
| Agility Robotics RoboFab capacity | >10,000/year | Agility Robotics 2023 |
| UBTech Walker S2 delivered / 2026 capacity plan | Several hundred / 5,000/yr | UBTECH, Nov 2025 |
*Tesla's 2026 target is widely reported but the original source is unverified: multiple outlets cite a 50,000–100,000-unit goal and a 1M-unit/year Fremont capacity ambition, but Tesla has published no official production guidance. Tesla missed its reported 5,000-unit 2025 target.
04Cost Curves and Bill of Materials
Cost is falling faster than every published model expected. Goldman Sachs logged a 40% manufacturing-cost decline in a single year against a forecast 15–20%.
Actuation — not compute — is where the money sits: more than half of a humanoid's bill of materials is actuators and dexterous hands (see our robot actuator market statistics). That makes actuator and camera sourcing the single biggest cost lever a robotics team controls.
| Metric | Value | Source |
|---|---|---|
| Manufacturing cost per unit (2024) / one-year decline | $30K–$150K / −40% | Goldman Sachs Research 2024 |
| Pilot-stage BOM cost, 2025–26 | $90K–$100K | BofA Institute 2026 |
| China-built BOM 2025 → 2030E | $35K → <$17K | BofA Institute 2026 |
| Actuation share of BOM 2030E (linear 27% + rotary 24% + hands 19%) | >50% | BofA Global Research 2026 |
| Unitree R1 / R1 Air list price | $5,900 / $4,900 | Unitree official store 2026 |
| 1X NEO home robot price | $20,000 or $499/mo | 1X Technologies 2025 |
| Figure in-house actuators produced / battery first-pass yield | 9,000+ / 99.3% | Figure, Apr 2026 |
List prices are not BOM: the $4,900 Unitree R1 Air is a 27kg research platform, not an industrial worker. Full-size industrial units still land in the $90K–$100K pilot band.
05Funding and Investment
Capital concentrated hard in 2025–2026: two US companies — Figure and Apptronik — absorbed roughly a third of all disclosed humanoid funding, while China's category leader went public instead of raising private rounds.
Valuations run far ahead of revenue. Figure's $39 billion valuation is roughly 22x Unitree's entire 2025 revenue — and Unitree is the one that shipped 5,500 robots at a 60% gross margin.
| Metric | Value | Source |
|---|---|---|
| Humanoid VC funding 2025 / 2024 | $6.1B (139 deals) / $1.5B (65) | PitchBook 2026 |
| All robotics VC funding, H1 2025 | >$6B | Crunchbase News, Jul 2025 |
| Figure Series C / post-money valuation | >$1B / $39B | Figure, Sep 2025 |
| Apptronik total Series A / valuation | $935M+ / ~$5.3B | Apptronik, Feb 2026 |
| Sector investment 2018 → 2025 | $0.7B → $4.3B | BofA Institute 2026 |
| Unitree IPO raise / 2025 revenue / gross margin | RMB 4.2B / RMB 1.71B / 60.27% | Unitree prospectus via CNBC 2025–26 |
| Companies building humanoids (Jan 2026) / commercial launches | 50+ / 150 | BofA Institute 2026 |
Unitree is the profitability outlier: RMB 600M net profit on RMB 1.71B revenue in 2025, while most Western humanoid companies remain pre-revenue.
06China vs. United States: Two Ecosystems
China ships units; the US raises capital. Roughly 90% of 2025 humanoid sales happened in China, backed by a supply chain in which Chinese firms hold a 63% share — while US companies hold the valuation records and the frontier AI stack.
The labor math pressures both economies: US manufacturing alone may leave 1.9 million jobs unfilled by 2033 (Deloitte & The Manufacturing Institute), which is the demand-side argument in every bank model above.
| Metric | Value | Source |
|---|---|---|
| China share of 2025 global humanoid sales | ~90% | Omdia / IDC via Rest of World 2026 |
| Chinese share of humanoid supply chain | 63% | Morgan Stanley, The Humanoid 100, 2025 |
| Morgan Stanley China shipment forecast 2026 / 2030 | 50,000 / 446,000 | Morgan Stanley via SCMP, Jun 2026 |
| China humanoid output growth 2026 / Unitree+AgiBot share | +94% / ~80% | TrendForce, Apr 2026 |
| MIIT policy targets (issued 2023) | Mass production 2025; world-leading 2027 | China MIIT via SCMP 2023 |
| Industrial robot installs 2024 (context) / China share | 542,000 / 54% | IFR World Robotics 2025 |
| US manufacturing jobs needed by 2033 / potentially unfilled | 3.8M / 1.9M | Deloitte & Manufacturing Institute 2024 |
| Industrial robots in operation worldwide / 2025 install forecast | 4.66M / 575,000 | IFR World Robotics 2025 |
MIIT's guideline dates to October 2023; the 2026 output data (TrendForce, +94%) shows the policy largely tracking to plan. Morgan Stanley raised its China 2026 forecast twice — from 14,000 to 28,000 to 50,000 units — inside twelve months.
07Deployments and Pilots by Industry
Deployments in 2026 are real but narrow: totes, bins, and parts. Counterpoint expects 72% of 2027 installations in logistics, automotive, and manufacturing — structured environments where tasks repeat and demonstration data is cheap to collect.
Every pilot doubles as a data operation: a robot's shift output is measured in totes moved and demonstration hours captured. That is why deployment-scene access — the kind a sourcing and data partner provides across factories, hotels, and kitchens — has become a competitive input, and why we track the numbers separately in our robotics training data statistics.
| Metric | Value | Source |
|---|---|---|
| 2027 installations by industry | Logistics 33% · Auto 24% · Mfg 15% · Retail 12% | Counterpoint Research via BofA 2026 |
| Agility Digit totes moved at GXO (live deployment) | 100,000+ | Agility Robotics, Nov 2025 |
| UBTech Walker series 2025 order book (BYD, Foxconn, FAW-VW) | RMB 800M+ (~$112M) | UBTECH, Nov 2025 |
| Apptronik Apollo pilot partners | Mercedes-Benz, GXO | Apptronik, Feb 2026 |
| 1X NEO home robot deliveries | US deliveries start 2026 | 1X Technologies 2025 |
| Household share of installed base by 2060 | 62% (~2B units) | BofA Institute 2026 |
| Demand under 5–15% labor-substitution scenario (mfg & hazardous work) | 1.1M–3.5M units | Goldman Sachs Research 2024 |
Home deployment is the long tail, not the near-term market: BofA does not see households dominating the installed base until the 2050s.
Summary: The Humanoid Robot Market by the Numbers
| Metric | Value | Source |
|---|---|---|
| Humanoid market by 2050 (Morgan Stanley) | $5T | Morgan Stanley 2025 |
| Humanoid market by 2050 (Citi) | $7T | Citi Institute 2024 |
| Humanoid TAM by 2035 (Goldman Sachs) | $38B | Goldman Sachs 2024 |
| Market size 2025 → 2030 | $2.92B → $15.26B | MarketsandMarkets 2025 |
| Humanoids sold globally, 2025 | 13,000–18,000 | Omdia / IDC 2026 |
| 2026 shipment forecasts | 50K (TrendForce) – 90K (BofA) | TrendForce 2025; BofA 2026 |
| Annual shipments by 2035 | 1.4M (GS) – 10M (BofA) | Goldman 2024; BofA 2026 |
| Humanoids in operation by 2050 | 650M (Citi) – 1B+ (MS) | Citi 2024; Morgan Stanley 2025 |
| Installed base by 2060 | 3B (62% household) | BofA Institute 2026 |
| Unitree 2025 shipments / share | 5,500+ / 32.4% | Unitree prospectus 2026 |
| Figure 03 delivered / production rate | 350+ / 1 per hour | Figure, Apr 2026 |
| Tesla Optimus sold in 2025 | ~150 | Omdia via Rest of World 2026 |
| China-built BOM 2025 → 2030 | $35K → <$17K | BofA Institute 2026 |
| Cheapest commercial humanoid (Unitree R1 Air) | $4,900 | Unitree 2026 |
| Humanoid VC funding 2025 | $6.1B (+300%) | PitchBook 2026 |
| Figure post-money valuation | $39B | Figure, Sep 2025 |
| China share of 2025 humanoid sales | ~90% | Omdia / IDC 2026 |
| Chinese share of humanoid supply chain | 63% | Morgan Stanley 2025 |
| China humanoid output growth, 2026 | +94% | TrendForce, Apr 2026 |
| 2027 industrial share of installations | 72% | Counterpoint via BofA 2026 |
Frequently Asked Questions
How big is the humanoid robot market in 2026?
Research firms place the 2026 humanoid robot market between roughly $2.9 billion and $6.2 billion, depending on definition: Mordor Intelligence estimates $3.93 billion, Fortune Business Insights $6.24 billion, and MarketsandMarkets $2.92 billion for 2025 growing at 39.2% annually. Long-term bank forecasts are far larger — $5 trillion by 2050 per Morgan Stanley and $7 trillion per Citi — because they include software, services, and secondary industries.
How many humanoid robots exist in the world?
Roughly 13,000–18,000 humanoid robots were sold globally in 2025 according to Omdia and IDC data, about 90% of them in China. Counting earlier deliveries, the cumulative worldwide fleet in mid-2026 is on the order of a few tens of thousands of units. Bank of America projects the installed base will reach 300 million by 2040 and 3 billion by 2060.
Who ships the most humanoid robots?
Unitree led 2025 with 5,500+ units shipped — 32.4% of global humanoid shipments per its IPO prospectus — followed by AgiBot at 5,168 units. The largest Western producers (Figure, Agility Robotics, Tesla) each sold around 150 robots in 2025 per Omdia data, though Figure had delivered 350+ Figure 03 units by April 2026 at a rate of one robot per hour.
How much does a humanoid robot cost in 2026?
List prices span $4,900 for Unitree's R1 Air research platform to about $20,000 for 1X's NEO home robot, while full-size industrial humanoids cost $90,000–$100,000 at pilot stage per Bank of America. BOM costs are falling fast: a China-built humanoid's bill of materials was about $35,000 in 2025 and is projected to drop below $17,000 by 2030.
Will 2026 be the year humanoid robots go commercial?
2026 is the consensus inflection year for early commercialization, not mass adoption. TrendForce forecasts shipments above 50,000 units (+700% year over year), BofA models 90,000 units, and Morgan Stanley raised its China forecast to 50,000 units. Deployments remain concentrated in logistics and manufacturing pilots — Counterpoint expects 72% of 2027 installations in logistics, automotive, and manufacturing. Teams typically move spec → sample → pilot → scale over multiple quarters.
What is the biggest bottleneck to scaling humanoid robots?
Cost and capability, in that order. Actuators alone are more than 50% of the bill of materials per BofA, and Goldman Sachs' base case still assumes key technical barriers must be cleared before a mass-market general-purpose humanoid is proven. On the capability side, the binding constraint is high-quality manipulation training data — BofA notes progress remains limited by the scarcity of real-world demonstration data.
Methodology and Sources
This article aggregates 54 statistics from 24 primary sources, including bank research (Goldman Sachs, Morgan Stanley, Citi, Bank of America), market research firms (TrendForce, Omdia, IDC, Counterpoint, Grand View Research, MarketsandMarkets, Mordor Intelligence, Fortune Business Insights), industry bodies (IFR), funding databases (PitchBook, Crunchbase), and official company disclosures (Figure, Agility Robotics, UBTECH, Apptronik, Unitree, 1X). We cite original research over secondary blog coverage, distinguish annual shipments from installed base, and present conflicting forecasts side by side rather than picking a single number.
Verification notes: figures from the Unitree IPO prospectus, PitchBook's humanoid funding data, Morgan Stanley's Humanoid 100 supply-chain share, and Omdia/IDC 2025 shipment counts are consistently reported across three or more independent outlets but sit behind paywalled or restricted primary documents; they are labeled with their originator throughout. Tesla's 2026 production target is flagged inline as widely reported but unconfirmed by any official Tesla source. Goldman Sachs' forecast dates to February 2024 and Citi's to December 2024 — the most recent public versions from both banks.
- Goldman Sachs Research — The global market for humanoid robots could reach $38 billion by 2035 (2024)
- Morgan Stanley — The Humanoid Economy: $5 Trillion by 2050 (2025)
- Citi Institute — The Rise of AI Robots (2024)
- Bank of America Institute — Physical AI, Part 2: Humanoid Robots (March 2026)
- TrendForce — 2026 Humanoid Commercialization Forecast (December 2025)
- TrendForce — China Humanoid Output to Surge 94% in 2026 (April 2026)
- Rest of World — China is winning the humanoid robot race (Omdia/IDC data, February 2026)
- Figure — Ramping Figure 03 Production (April 2026)
- Figure — BotQ: A High-Volume Manufacturing Facility (2025)
- Figure — Series C at $39B Post-Money Valuation (September 2025)
- Apptronik — Closes Over $935 Million Series A (February 2026)
- Agility Robotics — Digit Moves Over 100,000 Totes (November 2025)
- Agility Robotics — Opening RoboFab (2023)
- UBTECH — Walker S2 Mass Production, Orders Exceeding 800 Million Yuan (November 2025)
- Unitree — R1 official pricing (2026)
- CNBC — Unitree plans $7 billion IPO valuation (September 2025)
- 1X Technologies — NEO order page (2025)
- PitchBook — The limits of VC's humanoid bet (2026)
- Crunchbase News — Robotics Startup Funding Rises H1 2025 (July 2025)
- IFR — World Robotics 2025 press release
- SCMP — Morgan Stanley raises China humanoid shipment forecast to 50,000 units (June 2026)
- SCMP — China MIIT humanoid robot guideline (2023)
- Deloitte & The Manufacturing Institute — 3.8 Million New Employees Needed by 2033 (2024)
- Grand View Research — Humanoid Robot Market Report (2025)
- MarketsandMarkets — Humanoid Robot Market (2025)
- Fortune Business Insights — Humanoid Robots Market (2025)
- Mordor Intelligence — Humanoids Market (2026)
Last updated July 2026. We update this page quarterly as new data becomes available.