Whether you're looking for an angel investor, a growth advisor, or just want to connect — I'm always open to great ideas.
Get in TouchChina's Ministry of Industry released the world's first national standard system for humanoid robots covering 330+ models across 140+ manufacturers. The US and EU have zero comparable frameworks.
AI, startups & growth insights. No spam.
TL;DR: On March 1, 2026, China's Ministry of Industry and Information Technology released the world's first national standard system for humanoid robots — a six-component framework covering everything from brain-like computing to safety and ethics, developed by 120+ institutions across an industry that already has 140+ manufacturers and 330+ active robot models. The United States has no equivalent. The European Union has broad AI regulation but nothing that touches humanoid robot standards directly. China has effectively written the rules for an industry the rest of the world is still trying to define.
On March 1, 2026 — the eve of China's Two Sessions political meetings — the Ministry of Industry and Information Technology published what it is calling the national standard system for humanoid robots. This is not a regulatory guidance document or a voluntary industry code. It is a structured, government-issued standard system that establishes technical and safety specifications across the entire humanoid robot stack, from the computation layer to the physical hardware to the deployment context.
The standard system was developed with input from more than 120 research institutions, enterprises, and industry users. It covers an industry that has grown rapidly enough to produce 140 domestic manufacturers and over 330 active humanoid robot models in the Chinese market alone. The framework did not arrive in a vacuum — it arrives at the moment the Chinese humanoid robot industry is transitioning from R&D showcase to mass production, and the timing is not accidental.
What makes this release historically significant is not just that China did it — it is that no one else has. There is no equivalent national standard system for humanoid robots in the United States, the European Union, Japan, South Korea, or any other country with a developed robotics sector. China has done what it has done in several other technology domains in recent years: moved from fast follower to standard-setter by acting before anyone else organized around the problem.
The MIIT framework covers six component areas, each addressing a distinct layer of the humanoid robot system. The breadth of the framework signals that this is not a narrow technical specification document. It is an attempt to establish comprehensive national governance of an entire product category before that category reaches the scale where governance becomes reactive rather than proactive.
The MIIT standard system is structured around six components that together span the full scope of what a humanoid robot is and does. Understanding each component helps clarify both the technical ambition of the framework and the governance philosophy behind it.
1. Basic commonality. This is the foundational layer — the definitions, terminology, and classification standards that allow the industry to speak a common language. Before you can regulate or standardize anything, you need shared definitions of what the thing is. Humanoid robots span a wide range of form factors, capability levels, and intended use cases. Basic commonality standards establish the taxonomy: what counts as a humanoid robot, how models are classified by capability, and what terms mean consistently across the industry.
2. Brain-like computing. This component addresses the AI and computational architecture that powers humanoid robot cognition. It covers the software systems, inference frameworks, training standards, and computational requirements that govern how a humanoid robot's "brain" processes sensory input, makes decisions, and issues motor commands. Given that the intelligence layer is where most of the meaningful differentiation between humanoid robots happens, this component is arguably the most technically complex in the framework.
3. Limbs and components. The physical hardware layer — joints, actuators, sensors, structural materials, and the interfaces between them. Standardizing components at this level enables interoperability between manufacturers, simplifies supply chain development, and establishes baseline performance requirements for the mechanical systems that allow a humanoid robot to move, grasp, and interact with physical environments.
4. Complete machines. Standards for the integrated system — the humanoid robot as a complete product. This component covers how the brain, limbs, sensors, and power systems are assembled into a functioning machine, including interface standards between subsystems, integration testing requirements, and performance benchmarks for the complete robot.
5. Application. Deployment standards covering the contexts in which humanoid robots operate: manufacturing floors, logistics centers, healthcare settings, consumer environments. Application standards govern how robots are integrated into workflows, what they are permitted to do in different settings, and how their performance should be evaluated in real operational conditions.
6. Safety and ethics. The component that will draw the most international attention. This section addresses human-robot interaction safety, failure mode behavior, the handling of sensitive data collected during robot operation, and the ethical constraints on how humanoid robots should be designed and deployed. The inclusion of ethics as a named component in a government standard system is notable — it positions humanoid robot governance not just as a technical and commercial question but as a question of values.
| Component | Scope |
|---|---|
| Basic commonality | Definitions, taxonomy, classification |
| Brain-like computing | AI architecture, inference, decision systems |
| Limbs and components | Hardware, actuators, sensors, interfaces |
| Complete machines | System integration, benchmarks |
| Application | Deployment contexts, workflow integration |
| Safety and ethics | Human-robot safety, data, ethical constraints |
The humanoid robot industry's situation before March 1, 2026, is unusual even by the standards of fast-moving technology sectors. Across 140 manufacturers in China alone, 330 distinct humanoid robot models were being developed and, in some cases, deployed — without any common technical vocabulary, without interoperability requirements between manufacturers, without agreed safety specifications for human-facing deployment, and without any regulatory structure governing what these machines should and should not do.
That situation creates problems at multiple levels simultaneously.
For manufacturers, the absence of standards means every company is solving the same foundational engineering problems independently. There are no common component interfaces, which means supply chain development is duplicated across hundreds of companies. There are no common testing protocols, which means every manufacturer defines "safe" and "functional" according to its own internal criteria.
For enterprise buyers, the absence of standards creates procurement risk. When a logistics company wants to deploy humanoid robots in a warehouse, it needs confidence that those robots will operate safely, that they meet some minimum performance threshold, and that the manufacturer's safety claims have been validated against something other than the manufacturer's own judgment. Without standards, that confidence is impossible to establish systematically.
For the Chinese government, which has identified humanoid robotics as a strategic industry priority, the absence of standards creates a coordination problem. State investment in the sector — through subsidies, research grants, and procurement — is more effective when directed at manufacturers who are building to common specifications rather than pursuing hundreds of incompatible architectures.
Standards solve all of these problems simultaneously. They reduce duplicated engineering effort, enable supply chain efficiency, give buyers a basis for evaluation, and allow government investment to coordinate on shared infrastructure rather than divergent technical approaches.
The fact that China acted first means Chinese manufacturers will shape what those standards contain. International manufacturers who want to enter the Chinese market — or who want to partner with Chinese companies — will need to comply with a framework they had no hand in writing.
The scale of China's humanoid robot industry is frequently underestimated outside China, in part because the most visible humanoid robot announcements in Western media tend to feature US companies: Figure, Agility Robotics, Boston Dynamics, Tesla's Optimus program.
The actual production and deployment numbers tell a different story.
Chinese industrial robot output grew 14% year-over-year in 2024, generating $33.4 billion in revenue. In 2025, that growth rate doubled to 28% — a sign that the sector moved from steady growth into accelerating expansion. Chinese government and industry observers named 2025 "Year 1" of humanoid robot mass production in China, marking the transition from prototype and small-batch manufacturing to genuine volume production.
The 140 domestic manufacturers and 330+ models in the current market are a product of both aggressive private investment and coordinated state support. China's approach to strategic technology sectors — combining state-directed research funding, preferential procurement policies, and rapid private-sector scaling — has produced a market that is broader in terms of number of players than any other country's humanoid robot ecosystem.
The patent picture reinforces the scale advantage. China holds approximately 3 out of 5 of the world's AI patents and 2 out of 3 of the world's robotics patents. These figures represent accumulated investment over more than a decade and reflect a deliberate national strategy to dominate the intellectual property landscape in technologies identified as strategically important.
| Metric | Figure |
|---|---|
| Domestic humanoid robot manufacturers | 140+ |
| Active humanoid robot models | 330+ |
| 2024 industrial robot revenue | $33.4 billion |
| 2024 industrial robot output growth | +14% YoY |
| 2025 industrial robot output growth | +28% YoY |
| Share of global AI patents | ~60% |
| Share of global robotics patents | ~67% |
The year-over-year doubling of growth rate from 2024 to 2025 is the most significant data point. Industries in genuine mass production transitions typically show this kind of acceleration before plateauing. China's humanoid robot sector appears to be in the early phase of that transition, which means the current 330-model figure is likely to grow substantially in the next two to three years.
The United States has no national standard system for humanoid robots. This is not a recent oversight — it reflects a structural difference in how the US and China approach technology governance.
The US regulatory approach for emerging technology is characteristically fragmented. Existing standards bodies — NIST, ANSI, IEEE, UL — develop voluntary technical standards through industry-led processes that operate on multi-year timescales. Government agencies regulate specific harms (workplace safety through OSHA, product safety through the CPSC, data privacy through sector-specific frameworks) rather than comprehensive product categories. There is no mechanism in the US system that produces a top-down, government-issued national standard system for a product category the way MIIT just did for humanoid robots.
The executive order on AI safety signed in 2023 directed NIST to develop AI risk management frameworks, and NIST published its AI Risk Management Framework. But the NIST AI RMF is a voluntary guidance document, not a standard. It applies to AI systems broadly, not to humanoid robots specifically. It has no binding force and no compliance mechanism.
The Biden administration's AI executive order, which the Trump administration subsequently revised, included provisions about AI safety evaluation and frontier model reporting requirements. None of these touch humanoid robot standards directly.
The result is a gap that is not primarily about competence or interest — there are serious robotics safety researchers in the US — but about institutional structure. The US system produces voluntary standards slowly, through industry consensus, without the ability to mandate compliance. China's system produced a comprehensive national standard in the time it takes a US standards committee to complete its first round of public comment.
For US manufacturers, this gap creates a practical problem. Companies like Figure AI and Agility Robotics are building humanoid robots without a domestic standard to build toward. When they want to export to or partner in China, they are building toward a foreign government's standard. When US enterprise customers ask them to demonstrate compliance with recognized safety standards, there is no US standard to reference.
The European Union's position is different from the United States' and in some ways more complicated.
The EU AI Act, which entered force in 2024, establishes a risk-based regulatory framework for AI systems operating in Europe. Humanoid robots, particularly those deployed in human-facing contexts like healthcare and logistics, would qualify as high-risk AI systems under the Act and would therefore be subject to conformity assessment requirements, technical documentation obligations, and registration requirements before deployment.
This sounds like a comparable framework. It is not.
The EU AI Act regulates the AI software systems that power products. It does not establish technical standards for the physical hardware — the joints, actuators, sensors, and integrated machine specifications — that define what a humanoid robot actually is. The Act references "harmonized standards" that manufacturers should comply with, but those harmonized standards for humanoid robots do not yet exist. They are being developed by European standards bodies (CEN, CENELEC, ETSI) through the same slow, industry-consensus process that characterizes US standards development.
The EU also has the Machinery Regulation (replacing the Machinery Directive), which covers safety requirements for machines including robots. The Machinery Regulation is broad enough to apply to humanoid robots in some deployment contexts, but it is designed for traditional industrial machinery and was not written with humanoid robots in mind. Applying it to humanoid robots requires interpretive work that varies by member state and by application context.
What the EU has is a regulatory framework that creates compliance obligations but has not yet produced the technical standards that would tell manufacturers precisely how to meet those obligations. China's MIIT framework does both: it establishes the requirement and specifies how to meet it in sufficient technical detail that manufacturers can actually build to it.
The March 1 release of the humanoid robot standard system was timed deliberately to precede the opening of China's National People's Congress on March 4, 2026 — the annual political meeting where national policy priorities are formally set and legislative direction is given to government agencies.
The NPC session is expected to formalize what Chinese government and industry sources have been signaling for months: that "AI-plus" — the integration of artificial intelligence into manufacturing, logistics, energy, and other strategic sectors — will be a central pillar of the 15th Five-Year Plan covering 2026 through 2030.
The 15th Five-Year Plan represents the Chinese government's most consequential policy document of this decade. Five-Year Plans are not predictions or aspirations — they are coordination mechanisms that align state investment, regulatory development, procurement preferences, and research funding around specific targets. When humanoid robotics and embodied AI appear in a Five-Year Plan, it means budget allocations, state-owned enterprise procurement mandates, research grant priorities, and regulatory timelines are all being organized around those targets simultaneously.
The "AI-plus" framing is significant because it positions humanoid robots not as an isolated product category but as infrastructure for industrial modernization. In manufacturing, AI-plus means humanoid robots performing the flexible, dexterous tasks that traditional industrial robots cannot. In logistics, it means autonomous humanoid systems handling last-mile movement in environments not designed for wheeled robots. In energy, it means robots capable of operating in hazardous environments that are currently human-only.
The national standard system released on March 1 is the regulatory prerequisite for that deployment vision. You cannot mandate that state-owned enterprises procure humanoid robots at scale without a standard that defines what qualifies. You cannot provide preferential financing for humanoid robot manufacturers without a framework that distinguishes qualified manufacturers from unqualified ones. The standard comes before the scale.
The 140 domestic manufacturers who constitute China's humanoid robot industry will not all benefit equally from the new standard system. Standardization characteristically accelerates market consolidation — companies with the engineering resources to meet formal standards quickly pull ahead of those who cannot.
Several Chinese manufacturers have emerged as leaders in the current pre-standard phase. UBTECH, one of China's oldest humanoid robot companies, has been deploying robots in education and hospitality for years and has the organizational infrastructure to navigate a standards compliance process. Unitree Robotics has gained international attention for its four-legged robots and is now developing humanoid variants. AgiBot (formerly Zhiyuan Robotics) raised significant funding in 2025 specifically targeting humanoid robot deployment in manufacturing. Fourier Intelligence focuses on rehabilitation and healthcare robotics, a high-stakes application context where safety standards are particularly consequential.
The new standard system benefits these established players in several ways. It creates a formal certification process that signals quality to enterprise buyers, reducing the need for extensive individual buyer evaluation. It gives larger manufacturers a framework for differentiating from smaller competitors who may not meet the new specifications. And it establishes China's technical approach as the domestic reference standard, meaning the manufacturers who built to these specifications are better positioned than international competitors who built to different assumptions.
International manufacturers — Boston Dynamics, Figure AI, Agility Robotics, and others — now face a decision. If they want to participate in the Chinese market or partner with Chinese manufacturers, they need to understand and eventually comply with a standard system they did not design. If they want to influence future revisions to that standard, they need presence in the Chinese industry ecosystem.
The release of China's national humanoid robot standard system is most usefully understood not as a single regulatory event but as a data point in a pattern that has repeated across several strategic technology sectors over the past decade.
The pattern: China identifies a technology as strategically important, invests heavily in domestic industry development through both state funding and market incentives, builds out manufacturing and patent position faster than competitors expect, then establishes technical standards that reflect its own ecosystem's assumptions and become the default for any company wanting access to the Chinese market.
This sequence has played out in electric vehicles, where Chinese battery and vehicle standards are now a reference point for global manufacturers even in markets where Chinese EVs are not sold. It has played out in telecommunications infrastructure, where Huawei's equipment specifications became de facto standards in markets that deployed its networks. The humanoid robot standard system follows the same logic.
The consequence for the global industry is that the technical choices embedded in the MIIT framework — the definitions in the basic commonality component, the computational architecture specifications in the brain-like computing component, the safety thresholds in the safety and ethics component — will influence how humanoid robots are designed globally, not just in China. Companies that want to operate in or with China will build toward these specifications. Companies that do not will diverge, potentially creating incompatibilities that make collaboration harder.
The US and EU have a window to develop their own standards frameworks before the Chinese framework becomes entrenched as the global default. That window is probably measured in months, not years, given the pace of Chinese humanoid robot deployment. The question is whether either system has the institutional capacity to act quickly enough to matter.
History in adjacent sectors suggests the answer may be no. But humanoid robots are different from most prior strategic technology domains in one important respect: the US defense and commercial robotics ecosystems are substantial and have been investing in humanoid robot capability for years. If that investment is organized around a standards effort with sufficient urgency, it is not too late to produce a competing framework. What it requires is the kind of coordinated, government-industry action that neither the US nor EU has historically done well in technology domains where speed matters more than consensus.
MIIT is China's Ministry of Industry and Information Technology, the government body responsible for industrial policy, technology development standards, and telecommunications regulation. It functions as both a standards-setting authority and an industrial policy ministry, giving it the ability to issue binding technical standards for product categories it designates as strategically important. In China's governance structure, MIIT has the authority and the institutional mandate to issue a national standard system like the humanoid robot framework — there is no equivalent single-agency authority in the US or EU systems.
The MIIT framework establishes a national standard system, which in the Chinese regulatory context carries significant weight. While the specific compliance timeline and enforcement mechanisms will be clarified in subsequent regulatory guidance, national standards in China typically become prerequisites for government procurement and preferential financing, effectively making compliance mandatory for manufacturers who want access to the largest customer segments. The 120+ institutions involved in developing the standard — including both research institutions and enterprises — suggests the framework was built with practical enforceability in mind.
First-mover advantage in standards is real but not permanent. The most consequential aspect of China's move is not the specific technical content of the standards — those will evolve — but the demonstration that it can coordinate an industry-spanning governance framework faster than its competitors. If US and EU standards bodies respond with comparable urgency, the first-mover advantage narrows. If they respond at typical voluntary-standards-body pace, China's framework becomes entrenched before competing frameworks exist. The window for a meaningful competitive response is open but closing.
Embodied AI refers to artificial intelligence that operates in and through a physical body — as opposed to AI that processes data and produces outputs in purely digital form. Humanoid robots are the most prominent current example of embodied AI: they perceive their physical environment through sensors, process that information through AI systems, and act on the world through physical actuators. The distinction matters for regulation because embodied AI introduces physical safety risks — a robot that makes a wrong decision can cause physical harm in ways that a purely digital AI system cannot. China's standard system addresses this by including safety and ethics as a named component alongside the technical specifications.
The timing reflects the relationship between administrative action and legislative direction in China's governance system. MIIT releasing the standard on March 1 — just before the NPC opens on March 4 — positions humanoid robot standardization as an accomplished fact before the NPC session formalizes the broader "AI-plus" policy direction. It demonstrates implementation capacity to the legislative body that will be setting five-year priorities. It also signals to domestic industry that the regulatory framework is moving ahead of the NPC session rather than waiting for it, which gives manufacturers clarity to proceed with compliance planning.
The US humanoid robot market is substantially smaller in terms of number of active models and manufacturers. US companies like Figure AI, Agility Robotics, Boston Dynamics, and Tesla's Optimus program represent perhaps a dozen serious humanoid robot development programs, compared to 330+ models across 140+ manufacturers in China. The US programs are in many cases technically competitive or superior in specific capability dimensions — Boston Dynamics' Atlas program has demonstrated capabilities that no Chinese manufacturer has matched publicly. But the breadth and scale of China's industrial base, particularly in manufacturing volume and component supply chains, means China's aggregate production capacity for humanoid robots is substantially greater than the US equivalent.
German robotics startup Neura Robotics closed approximately €1 billion in funding from Tether Holdings, valuing the company at €4 billion as it prepares to fill nearly €1 billion in existing orders for cognitive humanoid machines.
Boston Dynamics partners with Google DeepMind to integrate frontier AI into the next-generation Atlas humanoid robot, accelerating the physical AI revolution.
China's 15th Five-Year Plan (2026–2030) names AI in 50+ directives and designates quantum computing a strategic technology — a direct reply to US chip export controls.