Humanoid Orders vs. Actual Deployments: How Much Commercial Progress Is Really Happening?
Humanoid Orders vs. Actual Deployments: How Much Commercial Progress Is Really Happening?
The article examines the discrepancy between announced humanoid robot orders and real-world deployments, noting that while companies report significant orders, actual deployment numbers are much lower. It outlines different stages of commercial activity, from MOUs to active deployments, and stresses the importance of operational metrics over order volume in assessing commercial success.
Source: Robotics 24/7 — Industry reporting · Read original article ↗
Article text · Original source · English
On this page
Humanoid robot makers shipped about 20,000 units in 2025, according to Bank of America figures reported by Reuters. Chinese manufacturers built 95% of them. Smart Analytics Global counted 19,100 shipments in the first half of 2026, with AgiBot and Unitree supplying three of every four. Order announcements arrive weekly. Deployment counts arrive far less often, and the two rarely match.
The Robotic Life tracks humanoid robot companies through a business lens. To evaluate announcements, it distinguishes purchase commitments from deployment status. A signed order, a delivered robot, and a robot doing useful work are separate milestones.
What Does a Humanoid Robot Announcement Establish?
Humanoid robot announcements can describe four different types of activity. These categories are a reading guide, not mutually exclusive stages: a paid pilot can also involve a purchase contract, and a commercial order may still await delivery.
- MOUs and letters of intent: These often express an intention to cooperate without establishing a binding purchase commitment. Check the actual terms for payment obligations, delivery schedules, and conditions.
- Framework agreements and future orders: Humanoid, the London startup, announced an agreement involving at least 1,000 robots for Schaeffler, alongside Bosch-backed production capacity for 100,000 units over five years. Forbes reports that three-to-six-month customer pilots are planned for late 2026, with a Beta robot due in the fourth quarter. Production capacity is not customer demand, and the deployment agreement does not establish that those robots have already been delivered or accepted.
- Pilots and proofs of concept: Figure reports that Figure 02 ran more than 1,250 hours during an 11-month program at BMW Group Plant Spartanburg and loaded more than 90,000 parts for X3 production. Figure announced the results in November 2025 and returned Figure 02 to headquarters as part of its retirement. The program demonstrates sustained work on a defined task; it does not establish broad fleet deployment.
- Active commercial deployments: Agility Robotics reports more than 65,000 operating hours and deployment commitments across nine customer facilities, naming Schaeffler, GXO, Toyota Motor Manufacturing Canada, and Mercado Libre among its commercial customers. Those disclosures provide evidence beyond an order announcement, although aggregate operating hours do not reveal productive hours per robot, autonomy, or customer economics.
Long-term production targets describe intended factory output rather than customer demand. Bosch-backed capacity for 100,000 Humanoid units over five years and Figure’s goal of 100,000 robots over four years belong in this category. Reuters also reports that a Chinese industry ministry official said China expected to build more than 100,000 humanoids in 2026. These are capacity plans, goals, or expectations, not delivered customer orders.
Where Does the Headline Order Diverge from the Deployment Ledger?
Three 2026 cases show how far a headline order can sit from a deployed humanoid robot.
- Agility: Conditional future orders versus historical revenue. Agility’s June 24, 2026 announcement reports more than $300 million of multi-year Digit v5 orders, subject to contractual milestones. The cited merger filing reports approximately $1.8 million in 2025 net sales and a $140 million operating loss, with nearly two-thirds of sales from related parties. These figures cover different periods and measures: future multi-year commitments versus revenue recognized in one historical year. The comparison illustrates the scale of the planned expansion; it does not establish that orders have failed to convert into revenue.
- UBTech: Contracts include more than robot hardware. Reuters reports that UBTech won at least $123 million of fourth-quarter 2025 contracts to supply state-owned enterprises with humanoid hardware and related technology, compared with $122 million of full-year humanoid revenue. The contracts are not necessarily robot-only sales and may be fulfilled over time. The company shipped 1,079 full-size humanoids in 2025. Reuters also reports that manufacturing capacity outpaces commercial demand in China and that promotional dancing-robot rental rates declined from roughly $1,500 a day to $440–$600.
- Figure: Valuation and capacity alongside disclosed deployment evidence. Figure announced its Series C at a $39 billion post-money valuation and has not publicly disclosed revenue in the sources reviewed for this article. Its BotQ factory targets capacity of 12,000 robots a year. The BMW program is the Figure deployment examined here with published operating hours; this does not establish that it is the company’s only deployment.
What Does Commercial Availability Tell Us?
Funding totals and product availability help describe the sector, but neither establishes the scale of productive deployment.
As of October 1, 2026, The Robotic Life’s funding tracker reports $11.8 billion in disclosed capital raised across 26 tracked companies. That is a funding sample, not a census of every humanoid manufacturer or a count of companies with paying deployments.
![]()
The Robotic Life
The Robotic Life’s directory distinguishes R&D, prototypes, pilots, and products available to order and ship. This article describes the last category as broad commercial availability.
Selective enterprise sales can still constitute genuine commercial deployment: Agility, for example, reports active commercial customers even though TRL’s directory labels Digit as Pilot under its existing availability-based definitions. A research robot sold openly, and an industrial robot sold through a negotiated agreement can both generate revenue, while serving very different markets. Availability, payment, and productive use should therefore be assessed separately.
Who Deploys the Humanoid Robot After the Order Is Signed?
The deployment channel matters more than the order book, and systems integrators are becoming that channel for humanoid robots.
A NEURA Robotics engineer told me at Automate 2026 in Chicago that systems integrators will be the primary deployment channel for many humanoid manufacturers. That view points to the work required after an order is signed. Systems integrators can help assess a site, connect a robot to existing workflows, and establish its operating economics.
TechRepublic presents an illustrative Digit 5 purchase price of $200,000 and five-year ownership cost of approximately $400,000. These are estimates rather than a universal price or return-on-investment threshold. The business case depends on productive utilization, local labor costs, maintenance, integration, and the performance of alternative automation.
Agility’s Customer Acceleration Program and Arc fleet platform address parts of the work required to move from purchase intent to operational deployment.
How Do You Read the Next Humanoid Order Announcement?
Five questions separate a humanoid robot deployment from a press release:
- What is the commitment? Is it an expression of interest, a conditional framework, or a binding purchase agreement?
- Is the headline number units, dollars, or manufacturing capacity?
- Does a milestone condition or an "up to" qualifier apply?
- How many robots are installed and accepted, and what productive hours, uptime, and human interventions has the customer confirmed?
- Who installs and supports the fleet after delivery?
Smart Analytics Global forecasts approximately 60,000 humanoid shipments in 2026. Its estimates include full-size and smaller bipedal robots, wheeled systems, and ultra-bionic humanoids; they should not be read as a count of autonomous factory workers. Shipments measure hardware volume. Productive hours per unit, uptime, human intervention rates, and operating cost are needed to assess useful deployment.
Consistent reporting of those measures would make commercial progress much easier to judge.
Lars Talbert is the founder of The Robotic Life, which tracks humanoid robot companies, funding rounds, and deployment stages.
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
operating hours
>90,000
View original evidence
gure reports that Figure 02 ran more than 1,250 hours during an 11-month program at BMW Group Plant Spartanburg and loaded more than 90,000 parts for X3 production. Figure announced the results in November 2025 and returned Figure 02 to headquarters as part of its retirement. The program demonstrates sustained work on a defined task; it does not establish broad fleet deployment. Active commercial deployments: Agility
Open source S6operating hours
>65,000
View original evidence
Robotics reports more than 65,000 operating hours and deployment commitments across nine customer facilities , naming Schaeffler, GXO, Toyota Motor Manufacturing Canada, and Mercado Libre among its commercial customers. Those disclosures provide evidence beyond an order announcement, although aggregate operating hours do not reveal productive hours per robot, autonomy, or customer economics. Long-term production tar
Open source S7operating hours
>1,250 hours
View original evidence
gure reports that Figure 02 ran more than 1,250 hours during an 11-month program at BMW Group Plant Spartanburg and loaded more than 90,000 parts for X3 production. Figure announced the results in November 2025 and returned Figure 02 to headquarters as part of its retirement. The program demonstrates sustained work on a defined task; it does not establish broad fleet deployment. Active commercial deployments: Agility
Open source S6parts loaded
>90,000
View original evidence
gure reports that Figure 02 ran more than 1,250 hours during an 11-month program at BMW Group Plant Spartanburg and loaded more than 90,000 parts for X3 production. Figure announced the results in November 2025 and returned Figure 02 to headquarters as part of its retirement. The program demonstrates sustained work on a defined task; it does not establish broad fleet deployment. Active commercial deployments: Agility
Open source S6
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
gure reports that Figure 02 ran more than 1,250 hours during an 11-month program at BMW Group Plant Spartanburg and loaded more than 90,000 parts for X3 production. Figure announced the results in November 2025 and returned Figure 02 to headquarters as part of its retirement. The program demonstrates sustained work on a defined task; it does not establish broad fleet deployment. Active commercial deployments: Agility
Open source S6
Robotics reports more than 65,000 operating hours and deployment commitments across nine customer facilities , naming Schaeffler, GXO, Toyota Motor Manufacturing Canada, and Mercado Libre among its commercial customers. Those disclosures provide evidence beyond an order announcement, although aggregate operating hours do not reveal productive hours per robot, autonomy, or customer economics. Long-term production tar
Open source S7
Humanoid robot makers shipped about 20,000 units in 2025, according to Bank of America figures reported by Reuters . Chinese manufacturers built 95% of them. Smart Analytics Global counted 19,100 shipments in the first half of 2026 , with AgiBot and Unitree supplying three of every four. Order announcements arrive weekly. Deployment counts arrive far less often, and the two rarely match. The Robotic Life tracks human
Open source S2
ers reports that UBTech won at least $123 million of fourth-quarter 2025 contracts to supply state-owned enterprises with humanoid hardware and related technology, compared with $122 million of full-year humanoid revenue. The contracts are not necessarily robot-only sales and may be fulfilled over time. The company shipped 1,079 full-size humanoids in 2025. Reuters also reports that manufacturing capacity outpaces co
Open source S11
full-size and smaller bipedal robots, wheeled systems, and ultra-bionic humanoids; they should not be read as a count of autonomous factory workers. Shipments measure hardware volume. Productive hours per unit, uptime, human intervention rates, and operating cost are needed to assess useful deployment. Consistent reporting of those measures would make commercial progress much easier to judge. Lars Talbert is the foun
Open source S20
Source:Robotics 24/7 — Industry reporting · robotics247.com