Quick Answer: China dominates humanoid robot manufacturing with 137 active producers and commercially available units starting at $16,000. The US has no commercially available humanoid robots as of early 2026. Buyers face a genuine tradeoff: Chinese robots offer 10x lower cost and immediate availability, while US-origin robots promise stronger data sovereignty and fleet learning architectures when they ship. Your decision should hinge on data sensitivity, deployment timeline, and whether fleet intelligence compounds in your use case.
The Manufacturing Gap Is Larger Than Most People Realize
The numbers tell a stark story. China has at least 137 companies actively manufacturing humanoid robots. The United States has a handful of well-funded startups — Figure, Apptronik, Agility Robotics — none of which have units available for general commercial purchase in 2026.
This is not a temporary gap. China's robotics industrial policy has been compounding for over a decade. State-backed investment, integrated supply chains for motors and sensors, and a massive domestic market for labor automation have created a manufacturing ecosystem that the US cannot replicate in two or three years.
For buyers, this means something practical: if you need a humanoid robot deployed in your facility this year, your realistic options are overwhelmingly Chinese-manufactured.
The Price Gap Changes the Economics Entirely
The Unitree G1, a capable bipedal humanoid from one of China's leading robotics companies, starts at approximately $16,000. Fully configured units with enhanced manipulation capabilities and extended battery life run in the $25,000 to $40,000 range.
Figure 02, the most anticipated US-origin humanoid robot, does not yet have public commercial pricing. Based on manufacturing costs, partnership structures with BMW, and comparable US robotics pricing, industry analysts project pricing above $150,000 per unit when it becomes commercially available.
| Factor | Chinese Humanoid (e.g., Unitree G1) | US Humanoid (e.g., Figure 02) | |--------|--------------------------------------|-------------------------------| | Base price | ~$16,000 | $150,000+ (projected) | | Commercial availability | Now | Late 2026 - 2027 (projected) | | Fleet size for $500K budget | 12 - 30 units | 1 - 3 units | | Manufacturing ecosystem | 137+ companies | < 10 companies | | Data architecture | Varies by vendor | US-hosted cloud + edge | | Fleet learning | Limited / proprietary | Advanced neural sharing | | Firmware update origin | China-based servers | US-based servers |
That is not a marginal difference. A $500,000 automation budget buys you a fleet of Chinese humanoids or a pilot of two to three US units. The fleet economics alone change the ROI calculation fundamentally.
Data Security Is the Real Decision Point
The most important question is not about hardware capability. It is about data.
Humanoid robots operating in your facility collect everything: floor layouts, workflow patterns, employee movement data, product dimensions, inventory levels, process timing. A robot with cameras, LiDAR, and microphones operating eight hours a day generates a detailed operational map of your business.
With Chinese-manufactured robots, you need to evaluate three data vectors:
1. Data residency. Where does telemetry go? Some Chinese robotics companies route data through servers in mainland China. Others offer configurable endpoints. You need contractual guarantees, not marketing claims.
2. Firmware update chains. Every firmware update is a potential vector for data exfiltration or capability changes. If your robot's firmware is pushed from servers you do not control, in a jurisdiction with mandatory government data-sharing laws, that is a risk you must size explicitly.
3. Network architecture. Can the robot operate fully air-gapped? Some Chinese humanoid robots require persistent cloud connectivity for navigation updates. Others can run entirely on-premise. The difference matters enormously for sensitive environments.
US-origin robots from companies like Figure and Apptronik are building data architectures designed for enterprise data sovereignty from the ground up. Figure's fleet learning system processes data through US-hosted infrastructure with configurable data retention policies. That is a meaningful differentiator for defense-adjacent, healthcare, or IP-sensitive operations.
Fleet Learning: The Differentiator That Compounds
Fleet learning is where the US-origin humanoid robots are building their strategic moat. When one Figure 02 robot learns to handle a new object or navigate a tricky environment, that knowledge propagates across the entire fleet through shared neural network updates.
This is not a minor feature. It is the difference between buying a static tool and buying into an intelligence network that gets better every day. A fleet of 50 Figure robots learns 50 times faster than a single unit. Six months after deployment, your robots are materially more capable than the day they arrived.
Most Chinese humanoid manufacturers have not demonstrated comparable fleet learning architectures. Some, like Unitree, are developing cloud-based learning systems, but the data-sharing implications loop back to the security concerns above. If fleet learning runs through Chinese cloud infrastructure, every operational improvement also means more of your facility data flowing to servers outside your jurisdiction.
For a deeper dive into how fleet learning changes your automation ROI, read our full analysis in Fleet Learning in Robotics: Why One Robot's Experience Benefits Your Entire Investment.
Compliance Considerations by Sector
Not every buyer faces the same regulatory landscape. Your industry determines how much the origin question matters.
Defense and government contracting: ITAR and CMMC requirements effectively rule out Chinese-manufactured robots with cloud connectivity. Even air-gapped Chinese units face scrutiny in classified or controlled environments. Wait for US-origin options or explore non-humanoid domestic alternatives.
Healthcare: HIPAA data handling requirements demand clear data residency guarantees. If a humanoid robot operates in patient-facing areas, its sensor data may constitute protected health information. Chinese vendors rarely offer HIPAA-compliant data processing agreements.
Warehousing and logistics: Generally the most permissive environment. If your facility handles standard commercial goods, the data sensitivity is lower. Chinese humanoids can be deployed with appropriate network segmentation and monitoring.
Manufacturing: Depends entirely on what you manufacture. Consumer goods facilities have more flexibility. Aerospace, semiconductor, or defense supply chain operations should apply heightened scrutiny.
The Recommendation Framework
Use this decision tree to determine which origin makes sense for your deployment.
Choose Chinese-manufactured humanoids when:
- Your primary goal is cost-effective labor augmentation at scale
- Your facility handles non-sensitive commercial operations
- You need robots deployed in the next 90 days
- You can implement network segmentation and monitoring
- Fleet learning is not a critical requirement
Choose US-origin humanoids when:
- Your operations involve sensitive, classified, or regulated data
- Fleet learning that compounds over time is central to your ROI model
- You operate in defense, healthcare, or IP-intensive manufacturing
- Data sovereignty and firmware provenance are board-level concerns
- You can wait 12 to 18 months for commercial availability
Consider a hybrid approach when:
- You have both sensitive and non-sensitive operational areas
- You want to pilot Chinese units now and transition to US units as they ship
- Different facilities in your portfolio have different risk profiles
Explore both categories in our humanoid robot comparison tool or get a personalized recommendation from the Robotomated Advisor.
The Uncomfortable Truth
The US vs China framing oversimplifies a more nuanced reality. There are Chinese robotics companies building serious, enterprise-grade platforms with thoughtful data architectures. There are US companies whose robots are not yet proven in real-world deployments.
Origin is a useful proxy for data governance and supply chain risk, but it is not a substitute for due diligence on a specific vendor. Ask for data processing agreements. Demand network architecture documentation. Test air-gapped operation. Verify firmware update provenance.
The robot that works best in your facility is the one that meets your operational requirements, fits your budget, and aligns with your risk tolerance. Nationality is one input to that decision, not the whole decision.
Key Takeaways
- China has 137+ humanoid robot manufacturers with commercially available units; the US has none shipping to general buyers in early 2026.
- The price gap is roughly 10x: $16,000 for a Unitree G1 vs $150,000+ projected for a Figure 02.
- Data security, not hardware specs, should be the primary decision driver for most enterprise buyers.
- Fleet learning architectures give US-origin robots a compounding advantage that grows over time.
- Compliance requirements vary by sector: defense and healthcare face the most restrictions on Chinese-origin robots.
- A hybrid approach — piloting Chinese units now while planning for US units — is viable for many operations.
- Due diligence on the specific vendor matters more than a blanket country-of-origin policy.
Ready to evaluate specific humanoid robots for your operation? Start with our humanoid category explorer or let the Robotomated Advisor match you to the right platform based on your requirements.