China Told Its Robotics Industry It Has Six Months. The US Is Still Writing an Executive Order
- David Borish

- 7 days ago
- 8 min read

In late June, China's Ministry of Industry and Information Technology and its State-owned Assets Supervision and Administration Commission issued a joint directive to local governments and state-owned enterprises across the country. By the end of 2026, key humanoid robot products must complete application verification and regular deployment in a number of representative scenarios, entering what the ministry called "work mode." Local authorities had until the end of June to submit implementation plans and until the end of November to report progress.
That is a government deadline attached to actual state-owned enterprises. It lands at a moment when the US and Chinese robotics industries are diverging in both scale and posture.
A Deadline, Not a Vision Statement
The scheme covers manufacturing, warehousing, logistics, retail, healthcare, workplace safety, equipment inspection, maintenance, emergency response, and disaster relief. Shao Hao, senior director of the robotics lab at smartphone maker Vivo, described the intent to the South China Morning Post: the policy is meant to push the industry from a demonstration-driven logic to a task-oriented logic, moving away from showcasing individual capabilities and toward building integrated systems that can perform real-world tasks. Shao added that six months is a tight window, but a concentrated push of this kind can help the industry converge faster on workable engineering solutions.
The commercial orders already reflect that shift. State Grid Corporation of China, which operates across 26 of mainland China's 31 provincial-level regions, committed 6.8 billion yuan, roughly $1 billion, to a 2026 procurement covering 500 humanoid robots, 3,000 dual-arm robots, and 5,000 quadruped units. The order spans more than 600 specialized tasks, from routine inspections to complex, high-risk maintenance. China Southern Power Grid is running a parallel program, and combined sector spending on embodied intelligence in China is projected to exceed 10 billion yuan, or about $1.46 billion, in 2026.
The deployments are not confined to controlled industrial settings. In Fangchenggang, a border city in Guangxi facing Vietnam, customs authorities are using Ubtech Robotics' Walker S2 humanoids to help manage the daily backlog of cargo trucks, buses, and commuters passing through the crossing, in a contract valued at roughly $40 million.
Logistics carriers SF Express and China Post have begun deploying humanoid robots from manufacturer Robotera at their sorting centers for high-frequency handling tasks. In Hangzhou, city authorities placed humanoid units alongside human officers directing traffic at intersections. These are not pilot programs designed to generate press coverage. They are operational tests of whether the machines can survive contact with unstructured, public environments.
Morgan Stanley has revised its shipment forecasts upward in response. The bank raised its 2026 China humanoid shipment estimate from 28,000 to 50,000 units, and its 2030 projection to 446,000 units, in a market it now expects to reach $15 billion. The upgrade rests on accelerating commercial validation, including the State Grid order, sustained policy support, and aggressive capacity expansion across the supply chain. Component makers are backing that up with real production numbers. Harmonic reducer manufacturer Leaderdrive increased monthly production capacity from 50,000 units in the first quarter of 2026 to roughly 70,000 currently, with plans to reach 100,000 to 120,000 by year end.
The Manufacturing Base Underneath the Robots
\
China's speed advantage did not appear this year. It rests on infrastructure the country has been building for a decade. China's installations reached 295,000 industrial robot units in 2024, a global market share of 54 percent. The country's national robotics strategy launched ten years ago, and the newly published 15th Five-Year Plan for 2026 to 2030 places robotics at the center of its industrial modernization effort.
A TechCrunch analysis captured the mechanism behind the current pace. China has built a more robust hardware supply chain, much of it developed through its electric vehicle sector, covering everything from sensors to batteries, combined with the world's strongest manufacturing base. That combination lets companies iterate faster than Western competitors. The advantage shows up directly in shipment volume. Chinese leader Unitree shipped roughly 36 times more units last year than US rivals Figure and Tesla combined.
This is also why China's humanoid rental market has become a real commercial category rather than a novelty. AGIBOT's rental subsidiary, SHAREBOT, said the robot rental market could reach $1.5 billion by the end of 2026, with customers able to rent a humanoid for as little as 3,500 yuan, about $517, a day, including shipping and a human operator to help control the machine. The company logged more than 5,500 orders in its first three months.
SHAREBOT's CEO, Li Yiyan, framed the model as a bridge technology, since renting the robots out helps move machines from exhibition pieces toward large-scale deployment.
CNN's on-the-ground reporting from Beijing's Robot Big World exhibition captured the honest state of the technology behind the marketing. A short walk from the choreographed public displays, more than 120 humanoids stood in rows at a separate state-backed facility, each repeating a single task such as sorting packages, changing diapers, or scooping popcorn, guided by human trainers with handheld controllers standing next to them. The scale of these training facilities points to the real bottleneck. Despite rapid advances in AI, robots still lack the volume of physical-world data needed to become genuinely capable workers, and China's approach is to generate that data through sheer deployment volume rather than wait for it.
Where the US Actually Stands
The US robotics picture is more mixed than either the boosters or the doomsayers suggest. On traditional industrial automation, the country is recovering. US industrial robot installations rose 11 percent year over year to reach 38,000 units in 2025, according to preliminary IFR data presented at Automate 2026 in Chicago. Automotive remained the largest adopting sector at 13,500 units, just 1 percent below the prior year, while food industry adoption surged 30 percent. IFR President Jane Heffner called it great news for the region, and the installation count marked the third strongest year on record for the US, behind only 2018 and 2022.
Robot density tells a more sobering story. The US stands at 307 industrial robots per 10,000 manufacturing employees, placing it eighth globally, two ranks higher than the previous year, behind South Korea's 1,220, Germany's 449, and Japan's 446, but ahead of China's 166. China's low density figure is a reminder that its dominance is a story of absolute scale and manufacturing breadth rather than automation intensity per worker. Even so, on total volume the gap remains enormous. IFR estimates that Chinese installations in 2025 ran about ten times higher than the US total, before even accounting for the humanoid segment separately.
The humanoid segment specifically, which draws the most capital and headlines, tells a more complicated story for the US. Interact Analysis research manager Blake Griffin offered a blunt assessment to Manufacturing Dive: in 2025 and 2026 there were virtually no real-world applications for humanoids, with most units serving academic research, development, or entertainment purposes, and China accounting for almost all deployments. His firm is forecasting fewer than 100,000 humanoids globally by 2030, with meaningful ramp not expected for roughly a decade.
Tesla's Optimus program illustrates the gap between announcement and delivery. Musk has confirmed, most recently at the Abundance Summit in March, that Optimus Gen 3 production begins in low volume in summer 2026, with high volume output targeted for 2027, after the company missed its original goal of showing a production intent Gen 3 in the first quarter.
As of April 2026, one detailed technical assessment noted that Optimus robots are not yet doing useful productive work, and that production targets have repeatedly slipped, with the gap between impressive demos and reliable general purpose autonomy still real and large. On the fourth quarter 2025 earnings call, Musk himself conceded that Optimus is not in usage in Tesla's factories in a material way. Specific figures claiming over 1,000 Optimus units already active on Tesla's Fremont line in January 2026 have circulated widely, but they trace back to marketing and industry analysis sites rather than primary Tesla disclosures or independent verification. That claim should carry a caveat until better sourced.
Other US humanoid efforts are further along on deployment, even if smaller in scale. Boston Dynamics' electric Atlas has begun commercial deployments, with its entire 2026 production allocation committed to Hyundai and Google DeepMind, and the company demonstrated Atlas working at a Hyundai plant in Georgia in January. BMW has expanded its use of Figure AI's humanoids at its Spartanburg, South Carolina plant following an earlier successful deployment there. Foundation, a US startup, has set a goal of building 50,000 humanoid robots by the end of 2027, a target that would still leave it well behind China's near term shipment volumes if Morgan Stanley's revised forecasts hold.
Washington's Policy Response Is Still in Draft Form
The US policy response to China's robotics push has been reported extensively but has not yet turned into action specific to robotics. Politico's reporting, cited across multiple outlets, describes Commerce Secretary Howard Lutnick as actively courting CEOs from companies including Boston Dynamics and Apptronik as the administration drafts a 2026 executive order aimed at robotics specifically. Ideas reportedly under discussion include tax breaks and federal funding for robotics firms alongside tougher trade measures targeting Chinese subsidies and intellectual property practices. A Department of Transportation robotics working group is also being formed to integrate automation into infrastructure and logistics planning.
That robotics specific order has not been signed as of this writing. What has been signed is a broader June 2 executive order titled Promoting Advanced Artificial Intelligence Innovation and Security, which directs federal agencies to build a framework for secure deployment of frontier AI models, including a process for developers to voluntarily grant early government access to models for up to 30 days before wider release. That order addresses frontier AI systems broadly, including cybersecurity and infrastructure protection, and does not set robotics hardware or deployment policy.
The more useful comparison than any single number is institutional posture. China's directive names specific state-owned enterprises, sets a six-month planning window, and requires progress reports by a fixed date. The comparable American robotics order remains, per reporting, still under evaluation inside the Commerce Department, with no signed text and no fixed deadline attached to industry.
What the Numbers Actually Support
Several figures now circulating in coverage of this story deserve a caveat. Claims about exact humanoid unit counts on any specific US factory floor, including widely repeated figures about Tesla's Fremont deployment, originate largely from industry blogs and marketing content rather than audited or independently verified sources, and Tesla itself has not made a comparable disclosure. Chinese shipment and revenue figures, including Morgan Stanley's forecasts and vendor reported order values, are the bank's own projections and company reported contract values rather than independently audited figures. The State Grid procurement and the MIIT directive itself are sourced to primary government and corporate disclosures reported by the South China Morning Post, China Daily, and Jiemian.
What is well supported across multiple independent sources is the directional story. China has moved from treating humanoid robots as demonstration technology to mandating operational deployment on a government timeline, backed by a manufacturing and component supply chain built over a decade of industrial policy. The US has a rebounding traditional industrial robotics sector, a humanoid segment still dominated by prototypes and pilot programs even among its most funded companies, and a robotics specific federal policy response that remains in the drafting stage while a broader AI security order has already been signed.
\
The next concrete marker to watch is the November 2026 progress reporting deadline that Chinese local governments and state enterprises face under the MIIT directive. That date will show whether China's six-month mandate produced durable operational deployments or simply accelerated procurement without matching gains in reliability. On the US side, whether Tesla's Optimus Gen 3 reaches even low-volume production on schedule this summer, and whether Washington's robotics executive order gets signed with specific deployment targets attached, will indicate whether the American policy response moves from courting industry CEOs to the kind of binding mandate China has already issued.