100 million humanoid robots ship in a year

Where things stand (October 2026)

Current output is three to four orders of magnitude below the threshold.

• IFR, World Robotics 2026 (30 Sep 2026): about 7,000 full-size humanoids were sold worldwide in 2025. Consumer service robots came to about 34.2M, but these are almost all floor cleaners and lawnmowers, which the question excludes. IFR also stresses that its figures are supplier-sample data. https://ifr.org/ifr-press-releases/news/global-sales-of-professional-service-robots-surge-24-percent
• Analyst tallies for H1 2026: IDC puts it near 25,000 units, with AgiBot the largest maker at more than 8,600 (35%). AgiBot reached 20,000 cumulative units in Sep 2026. Counterpoint says more than 22,000. Smart Analytics says 19,100; Pandaily's "1.91 million" headline is a typo, and its article body says 19,100. https://roboticsandautomationnews.com/2026/10/02/idc-data-shows-agibot-led-global-humanoid-robot-shipments-in-h1-2026/105463/
• Tesla: per The Information, reported by Electrek on 25 Sep 2026, Tesla built several hundred Optimus robots a week in August. It aims for more than 1,000 a week by year-end and an eventual ~20,000 a week, and the robots still can't generalize. The long-term targets are a ~1M/yr Fremont line and a 10M/yr Giga Texas line (Teslarati, 31 Jul 2026). Even the aspirational 10M is a tenth of the threshold.
• Forecasts for 2030: IDC expects more than 750k units, Morgan Stanley ~446k, and Goldman Sachs' base case is 250k+.

Reference class and arithmetic

The closest reference class is how fast production of complex electromechanical products has ever ramped up. Epoch AI ("How fast could robot production scale up?") estimates 2025 humanoid output at ~16k, doubling every 6 months. It finds that supply of high-precision reducers caps humanoid output near 500k/yr. Its wartime-style demand-shock scenario reaches only 10–30M humanoids/yr by end-2030, worldwide.

My calculations:

• Going from ~80k (global 2026) to 100M in CY2030 requires ×1,250, or 5.95× a year (doubling every 4.7 months) for four straight years.
• For the largest single maker (~25k in 2026) it requires ~8× a year.
• Epoch's naive trend gives CY2030 ≈ 16.4M and CY2031 ≈ 65.5M (16k×4^n).

So even if today's anomalous trend continues, the answer is NO. The fastest historical ramps (the iPhone, wartime production) also fall short.

Timing

CY2026 data can't appear before January 2027. CY2030 is the last year that can be reported in time: manufacturer reports would come in Q1–Q2 2031, and IFR's World Robotics in late Sep/early Oct 2031. CY2031 can't be complete and published by 31 Dec 2031. The quarterly risk therefore rises gradually and spikes in each Q1 (manufacturers' annual figures) and around each Q3/Q4 (IFR).

Paths to YES and the case against my estimate

1. An AI-driven industrial explosion. In the AI 2027 "race" scenario, output reaches 1M robots a month by end-2028, and robots help build robot factories. I give the chance that something like this delivers ≥100M qualifying units in one calendar year, reported by one manufacturer or IFR in time, at about 0.2–0.3%. Current evidence points the other way: Reuters (Aug 2026) reports Chinese humanoids can't yet do factory work, and China is slowing humanoid IPOs.
2. Misresolution. IFR's consumer-robot total could plausibly pass 100M in reports covering CY2029–2030. The criteria explicitly exclude it (vacuums and lawnmowers; mixed totals must establish ≥100M qualifying units), so this adds only about 0.1%.

A thin Manifold market (10 bettors) gives 33% that 100M humanoids will have been made cumulatively by 2032. That is a far easier bar, it accepts analyst syntheses, and it still rests on the same trend that leaves CY2030 near 16M. It does not move me much.

The hazards are 0.003–0.008% a quarter through 2028, 0.025% in Q1 2029, 0.06% in Q1 2030 and 0.11% in Q1 2031. Accumulated, they give ≈0.39% by 31 Dec 2031.

Question. By the last day of each quarter, has any manufacturer reported, or has the International Federation of Robotics (IFR) reported for the world, production or shipments of at least 100 million humanoid or general-purpose mobile robots in a calendar year?

Definitions.

• Humanoid robot: a powered, programmable physical robot with a human-like torso and two articulated arms capable of manipulating objects, and either articulated legs or a powered mobile base. A human-like face is not required. Humanoid robots need not independently satisfy the general-purpose test below.
• General-purpose mobile robot: a powered, programmable physical robot that can move itself between work locations and manipulate objects to perform multiple materially different physical tasks using the same basic hardware, with software changes, instructions, or interchangeable tools permitted. Different routes, payloads, or variants of the same task do not establish general-purpose capability. Wheeled, tracked, legged, and other mobile forms can qualify. Product documentation must describe present capabilities, not merely future plans.
• Manufacturer: any robot manufacturer worldwide, including a consolidated corporate group reporting its own output and subsidiaries' output. Contract-manufactured units reported as that manufacturer's products count. Unrelated manufacturers' totals cannot be added together under the manufacturer route.
• Production: completed physical robot units, including units made for internal use. Components, unassembled kits, software agents, and model installations are not units.
• Shipments: completed physical robot units dispatched for sale, lease, or deployment. Unit sales explicitly reported as delivered or recognized sales of completed robots count; bookings and orders do not. Each unit counts once within the chosen metric for the year; returns, repairs, reshipments, and transfers within the reporting corporate group cannot create additional units.
• Calendar year: January 1 through December 31 of one year. The Observation period convention applies. Non-calendar fiscal years, trailing twelve-month periods, and annualized rates do not qualify unless the source also supplies figures from which the calendar-year total can be calculated exactly.

Resolves YES if an eligible source reports at least 100,000,000 qualifying units under either route:

1. A manufacturer's own production or its own shipments in one calendar year, across any number of its qualifying models and facilities; or
2. IFR's worldwide production or worldwide shipments in one calendar year, across manufacturers.

Humanoid and other qualifying general-purpose mobile robots may be combined, but a unit meeting both definitions counts only once. Production and shipments cannot be added together. Nonoverlapping quarterly, monthly, or model-level figures may be summed if all are from the same eligible reporting body, use the same metric, and cover the same calendar year. IFR estimates of realized annual output or shipments count. A worldwide IFR supplier-sample count also counts if the observed qualifying units alone reach the threshold, without extrapolation by the resolver. A mixed category counts only when the source establishes that at least 100 million units satisfy these definitions. An explicitly reported rounded figure of 100 million counts; a range counts only if its lower bound reaches the threshold, and “nearly 100 million” does not.

Exclusions (resolve NO).

• Capacity, factory design targets, forecasts, orders, reservations, planned production, cumulative lifetime output, installed fleets, or monetary revenue without qualifying annual unit counts.
• Fixed-base industrial arms, software-only robots, ordinary autonomous vehicles, transport-only warehouse robots, delivery-only robots, robotic vacuums, lawnmowers, and other single-purpose mobile machines, unless they independently meet the humanoid definition or the general-purpose mobile definition.
• Decorative human-shaped machines lacking functional manipulation, or products called “humanoid” or “general-purpose” without supporting physical and capability documentation.
• Aggregates assembled by analysts or the resolver across unrelated manufacturers, except totals published by IFR.
• No qualifying report by the quarter-end.

Resolving source. Manufacturer-authored annual reports, regulatory filings, production or delivery reports, or official company website publications; or IFR's annual World Robotics — Service Robots, World Robotics — Industrial Robots, their successor publications, or IFR-authored statistical releases on ifr.org. These are the exclusive reporting bodies; third-party press reports and executive interviews do not independently qualify. Public confirmation (A2) applies subject to this narrower source restriction. Technical product documentation may establish which units qualify but cannot replace the required manufacturer's or IFR's unit report. IFR's 2025 Service Robots release identifies its figures as supplier-sample data, not whole-industry projections; general service-robot totals are not automatically qualifying totals. Apply the Resolution window and Resolution date (A3) conventions; a qualifying revised figure resolves on its publication date, and a later downward revision does not undo an earlier YES.