Mass-production roadmap versus shipment record: China's humanoids, 2023~2026

See how the roadmap's 2025 mass-production goal, set in 2023, checks out against 2025 shipments and first-half 2026 tallies, separating orders and deliveries from actual deployment. Shipment leadership and commercialization leadership are different records.

Updated Oct 17, 2026

Does shipment leadership mean the same as commercialization leadership?
Verified facts

Documented facts

On 2023-10-20, the Ministry of Industry and Information Technology published its guiding opinions on the innovation and development of humanoid robots. The targets were an innovation system, breakthroughs in key technologies, a safe and effective supply of core components, and realized mass production by 2025, plus a world-advanced level, a safe and reliable industrial supply chain, and a major new engine of economic growth by 2027. Bank of America figures cited by Reuters put 2025 global humanoid shipments at about 20,000 units, 95% of them built by Chinese manufacturers, while Smart Analytics Global counted 19,100 shipments in the first half of 2026, with AgiBot and Unitree supplying 3 of every 4. The same analysis treats Smart Analytics Global's 2026 forecast of about 60,000 units and a Chinese official's expectation of more than 100,000 builds in 2026 as capacity and goals rather than delivered orders, and cites UBTech's quarter 4 of 2025 contracts worth at least 1.23 hundred-million dollars with state-owned enterprises, 1,079 full-size humanoids shipped in 2025, and falling dancing-robot rental rates from 1,500 dollars a day to 440~600 dollars as cases of the order-deployment gap.

Sources [1][2][3]
Interpretation

What the numbers say

On shipments alone, China's mass-production goal is on track. The 95% share in 2025 and the 19,100-unit tally for the first half of 2026 point in the same direction as the video's 97% estimate for the first half of 2026. But contracts and deliveries are separate records. UBTech's contracts worth at least 1.23 hundred-million dollars may not be hardware-only sales and may be fulfilled over time, while 1,079 physical units delivered in a year and yearly humanoid revenue of 1.22 hundred-million dollars describe a different scale than the contract headlines. Records that capacity outpaces commercial demand and that event rental rates have fallen suggest production leadership is not the same as deployment leadership. The expected 100,000 builds and the 60,000-unit shipment forecast for 2026 should likewise be read as goals and estimates, not delivered orders.

Sources [3]
Concept

Separating orders, deliveries, and deployment

When reading humanoid announcements, separate the signed order, the delivered robot, and the robot doing useful work. Capacity and framework agreements state future ceilings and intent, pilot operating hours evidence a bounded task, and commercial deployment is confirmed with operating figures such as customers, operating hours, autonomy, intervention rates, and cost. Shipments measure hardware volume, not the number of autonomous factory workers.

Limitations

Limits

Shipment figures are estimates whose scope varies by firm. Even for 2025, readings of China's share of 85~95% coexist, and depending on the tally, full-size and smaller bipedal, wheeled, and ultra-bionic machines are counted together, so they cannot be read directly as autonomous factory workers. The video's 97% figure for the first half of 2026 is likewise a cited partial estimate, not a confirmed total. Funding and contract figures are samples and announcements, so they should not be stretched into censuses or confirmed deliveries.

Sources

  1. Akin Intelligence — November 2023 (Guidance on the Development of Humanoid Robots)Accessed Oct 17, 2026
  2. China targets mass humanoid-robot rollout by 2025 — UrgentComm / IoT World TodayAccessed Oct 17, 2026
  3. Humanoid Orders vs. Actual Deployments — Robotics 24/7, 2026-10-12Accessed Oct 17, 2026