The race to mass-produce humanoid robots accelerated dramatically this week as Unitree filed for a $610 million IPO, UBTech locked in Siemens to hit 10,000 units, and Tesla confirmed Optimus Gen 3 production starts this summer. China's dominance is solidifying—and the West is scrambling to respond.

Here's everything that mattered in humanoid robotics from March 16-22, 2026.

Unitree Files $610M Shanghai IPO—The First Major Humanoid Robotics IPO

In what could be a defining moment for the industry, Unitree Robotics filed an IPO application with the Shanghai Stock Exchange on Friday, seeking to raise 4.2 billion yuan (approximately $610 million). This marks the first major public offering from a company primarily focused on humanoid robots.

The Hangzhou-based startup, best known for its Unitree G1 and Unitree H1 humanoid robots, has grown explosively since launching its first quadruped robots in 2016. The company went viral during the 2025 Spring Festival Gala when its robots performed autonomous martial arts, reaching 679 million viewers. We covered that breakthrough moment extensively.

Why it matters: A successful IPO would validate humanoid robotics as a standalone investment category and unlock significant capital for R&D and manufacturing scale. It also sets a benchmark valuation for competitors like Figure AI, 1X, and others still in private funding rounds. Watch this one closely.

UBTech + Siemens: 10,000 Robots by Year-End

Chinese robotics firm UBTech signed a strategic cooperation agreement with Siemens Digital Industries Software on March 16, targeting annual production capacity of 10,000 humanoid robots in 2026.

The partnership integrates UBTech's full-stack robotics capabilities with Siemens' expertise in industrial digitalization. Siemens will provide its Xcelerator platform for end-to-end digital workflow—from design and simulation to manufacturing management.

"Mass production of tens of thousands of units has become a goal that we must achieve," said UBTech CEO Zhou Jian, noting the company has seen a surge in orders this year.

UBTech's Walker S2 industrial humanoid is already being deployed in manufacturing, with total orders exceeding 1.4 billion yuan in 2025. The company also recently signed an Airbus deal to expand into aviation manufacturing.

Why it matters: This is the clearest signal yet that humanoid robots are transitioning from prototype demonstrations to industrial-scale production. When a company partners with Siemens specifically for manufacturing scale, they're serious. UBTech is positioning itself as China's answer to Tesla's Optimus program.

Tesla Optimus Gen 3: Production Starts This Summer

At the 2026 Abundance Summit, Elon Musk provided the most detailed Optimus roadmap yet. Key revelations:

• Optimus Gen 3 is in final development stages and will begin initial production this summer
• High-volume production targeted for summer 2027
• 10 million square feet of factory space dedicated to robot manufacturing
• Tesla will not lay off humans—robots will boost per-worker output instead
• The Gen 3 will feature 50 actuators per hand, representing a significant dexterity improvement

Musk also predicted that recursive AI self-improvement (models training better models without human intervention) will be fully automated by end of this year or early next. His vision: robots will create massive deflation and eventually "Universal High Income," where governments issue money because AI-driven output exceeds human desire.

Tesla's Fremont factory is being repositioned to mass-produce Optimus, with the Gen 3 expected to be the "most advanced robot in the world." For deeper analysis on Tesla's robotics ambitions, see our coverage of Tesla's Model S sunset.

Why it matters: Tesla's scale advantage could make it the most significant player in humanoid robotics—if they can execute. Summer production start puts them behind UBTech and Unitree in timeline, but Tesla's manufacturing capabilities and capital reserves make them a formidable long-term competitor.

NVIDIA Partners with European Chipmakers for Humanoid Hardware

On March 16, NVIDIA announced partnerships with three major European semiconductor firms—Infineon, NXP, and STMicroelectronics—to develop and sell hardware specifically for humanoid robots.

The partnerships position NVIDIA as the central AI brain while European chipmakers provide the motion control, power management, and sensor integration hardware that humanoid robots require. This vertical integration strategy mirrors NVIDIA's success in data centers.

Why it matters: Hardware has been a bottleneck for humanoid robot commercialization. Better chips mean better battery efficiency, more responsive motor control, and lower costs. NVIDIA is betting that owning the full stack—from AI training to edge inference to motor control—will make them indispensable to every humanoid manufacturer.

Boao Forum 2026: China Sets Global Standards

The Boao Forum for Asia (BFA) Annual Conference 2026 kicks off next week in Hainan Province, with humanoid robots as both service providers and a central discussion topic. Ahead of the forum, China's Ministry of Industry and Information Technology released its first standard system covering the entire industrial chain and lifecycle of humanoid robots and embodied AI.

Market forecasts presented at pre-forum briefings suggest China's embodied AI market could reach:

• 400 billion yuan (~$55 billion) by 2030
• 1 trillion yuan (~$138 billion) by 2035

The Guardian's in-depth feature this week, "Inside China's Robotics Revolution," revealed that China now accounts for over half of the world's new factory robot installations annually, with some estimates suggesting Chinese companies control 90% of humanoid robot shipments.

Why it matters: Standards define markets. If China sets the global standards for humanoid robots—just as they did for solar panels and batteries—Western manufacturers will be forced to comply or face market exclusion. This is industrial policy playing out in real time.

The Contrarian View: Mark Cuban Says Humanoids Will Fail

Not everyone is bullish. At a TBPN live stream Thursday, Mark Cuban predicted humanoid robots will "fail miserably" within 5-10 years.

"Everybody's making this push for humanoid robots. I think they might have a 5-year lifespan, and then they'll fail miserably. Maybe 10," Cuban said.

His alternative vision: robots and spaces will be co-designed. Instead of making robots that fit human environments, we'll redesign environments to optimize for robot efficiency. "The robots aren't going to be full-form humanoids. They're going to be whatever the optimal shape is."

He pointed to Amazon's 1 million+ warehouse robots—none of which are humanoid—as evidence that form follows function.

Our take: Cuban isn't entirely wrong about industrial applications. Purpose-built robots often outperform general-purpose humanoids in controlled environments. But his argument misses the core value proposition: humanoids are designed for unstructured environments built for humans—homes, construction sites, disaster zones. You can't redesign a grandmother's house or a hospital corridor. The real question is whether the economics work, and this week's news suggests major players believe they will.

Market Pulse

• Robotics funding 2026 YTD: On pace for $20B+, with February's Figure AI momentum and SkildAI's $1.4B round leading the charge
• Humanoid market projection: $35 billion by 2030 (Business Research Company)
• China market share: 90%+ of humanoid robot shipments (Commonwealth Union)
• Production targets: UBTech (10K), Tesla (ramping), Unitree (post-IPO scale)

What to Watch Next Week

1. Boao Forum sessions on humanoid robotics—expect policy announcements and new partnerships
2. NVIDIA GTC 2026 aftermath—more hardware partnerships likely in the pipeline
3. Unitree IPO pricing details—valuation benchmarks for the entire sector

The humanoid robot industry just entered its industrial era. Mass production isn't coming—it's here. See our complete guide to the best humanoid robots to understand who's leading and why, or explore the full marketplace to see what's actually available today.

The Xpeng Iron is what happens when an $18 billion EV company decides humanoid robots are the next frontier. With 82 degrees of freedom, 22-DOF dexterous hands, three proprietary Turing AI chips delivering 3,000 TOPS, and a 110,000-square-meter factory breaking ground in 2026, Xpeng isn't building a prototype — it's building an army. This comprehensive Xpeng Iron review covers everything: verified specifications, AI capabilities, mass production timeline, and how it stacks up against Tesla Optimus, Figure 03, and China's other humanoid contenders.

Key Takeaways

• Price: Not officially announced; early estimates suggest ~$150,000 for enterprise deployments, with consumer pricing expected to decrease as production scales.
• EV Manufacturing Advantage: Xpeng is leveraging its automotive supply chain and high-volume production expertise — the same infrastructure that produces 300,000+ EVs annually.
• Compute Power: Three in-house Turing AI chips deliver 3,000 TOPS — among the highest compute in any humanoid robot.
• Production Timeline: Factory groundbreaking Q1 2026, mass production targeted late 2026. That's ground-to-robots in ~9 months.
• Best For: Retail service, industrial inspection, guided tours, and eventually home assistance as costs decrease.
• Key Limitation: No confirmed pricing or delivery dates for external customers; initial deployments will be Xpeng's own facilities.

Xpeng Iron Specifications

The Xpeng Iron — a full-size humanoid from China's third-largest EV maker with industry-leading compute and dexterity.

Xpeng Iron Price: What Will It Cost?

Xpeng has not officially announced pricing for the Iron humanoid robot. Based on industry estimates and competitor benchmarking, enterprise deployments are expected to start around $150,000 — comparable to Fourier GR-2 and significantly below Boston Dynamics Atlas ($420K).

However, Xpeng's stated strategy is to leverage automotive manufacturing scale to drive costs down rapidly. The company produces over 300,000 EVs annually with established supply chains for motors, batteries, sensors, and compute hardware — all components shared with humanoid robots. CEO He Xiaopeng has publicly committed to consumer-grade pricing as production scales.

Here's how the estimated Xpeng Iron price compares to the market:

The EV-to-robot strategy positions Xpeng similarly to Tesla with Optimus — both are betting that automotive manufacturing expertise translates directly to humanoid production at scale. If Xpeng hits its late-2026 mass production target, pricing could drop substantially by 2027.

Company Background: Why Is an EV Company Building Robots?

Xpeng Inc. (NYSE: XPEV) is China's third-largest electric vehicle manufacturer, valued at approximately $22 billion. The company produces the popular G6, G9, and P7 electric vehicles, with annual production exceeding 300,000 units. Xpeng was co-founded in 2014 by Xia Heng and He Tao, with He Xiaopeng (former UC Browser founder who sold to Alibaba) as initial backer and investor. He Xiaopeng now serves as Chairman and CEO.

The robotics division, Xpeng Robotics, was formally established following Xpeng's 2020 acquisition of Shenzhen startup Dogotix. Dogotix founder Zhao Tongyang initially led Xpeng's humanoid program before departing to launch EngineAI (known for the acrobatic PM01 and Terminator-inspired T800 robots).

At the November 2025 AI Day, He Xiaopeng officially repositioned the company as "a global embodied intelligence company" and "mobility explorer in the physical AI world." This isn't a side project — Xpeng views humanoid robots as the next logical extension of its AI-driven autonomous vehicle technology.

The strategic logic is compelling: Xpeng already designs AI chips (Turing), develops vision-language-action models (VLA 2.0), manufactures electric motors and batteries at scale, and operates a 30,000-GPU cloud computing cluster. These are exactly the capabilities needed for humanoid robots.

Performance and Mobility

The Xpeng Iron made headlines when company representatives cut through its synthetic skin on stage to prove no human was hiding inside. The demonstration was necessary because Iron's walking gait is remarkably natural — smooth, balanced, and eerily human-like.

Key mobility specifications:

• Walking Gait: Passive degrees of freedom at the toes enable a "light and gentle stride" that observers describe as among the most human-like in the industry
• Body DOF: 82 degrees of freedom across the full body — enabling complex movements beyond simple walking
• Humanoid Spine: Unlike rigid-torso designs, Iron features a flexible spine for natural bending and twisting
• Bionic Muscles: Soft actuator systems that mimic human muscle behavior
• Balance: AI-driven balance control using the VLA 2.0 model (though early demos did show one public fall)

The next-generation Iron introduced at AI Day 2025 features enhanced mobility over the original 2024 prototype. However, detailed performance metrics (walking speed, payload capacity, battery runtime) have not been officially disclosed.

AI and Compute Platform

Where Xpeng Iron truly differentiates is compute power and AI architecture. The robot runs on three proprietary Turing AI chips delivering a combined 3,000 TOPS (trillions of operations per second) — putting it among the most computationally powerful humanoid robots in existence.

The AI backbone is Xpeng's VLA 2.0 (Vision-Language-Action) model:

• End-to-End Architecture: VLA 2.0 eliminates the traditional "language translation" step between vision and action. Visual signals convert directly to motion commands.
• Training Data: Nearly 100 million video clips — equivalent to 65,000 years of human driving experience
• Cross-Domain: The same model powers Xpeng's autonomous vehicles, robotaxis, and humanoid robots
• Self-Evolving: The model learns interaction physics from the real world and generates synthetic scenarios for adversarial training
• Cloud Integration: Connected to Xpeng's 30,000-GPU cloud cluster running a 72-billion parameter base model

Xpeng claims VLA 2.0 represents a "new physical model paradigm" — moving beyond the standard Vision-Language-Action architecture to direct visual-to-motor generation. Whether this delivers practical advantages over competitors remains to be validated in real-world deployments.

Sensors and Perception

Xpeng Iron features 720° perception coverage — full spherical awareness around the robot:

• Vision System: Multiple cameras providing depth perception and object recognition
• IMU Sensors: Inertial measurement for balance and motion tracking
• 3D Curved Display: The robot's head features a wraparound display for expressive interaction

The sensor suite is designed to enable autonomous navigation in complex environments — retail stores, factories, and eventually homes. Detailed sensor specifications (camera resolution, LiDAR presence, etc.) have not been publicly disclosed.

Design and Build Quality

The next-generation Iron features several design innovations:

Humanoid Spine: Unlike rigid-torso robots, Iron's flexible spine enables natural bending, reaching, and twisting movements. This is critical for tasks like picking objects from low shelves or turning to face different directions.

Bionic Muscles: Soft actuator systems that provide more natural motion than traditional servo motors. This approach is similar to what 1X Technologies uses in NEO — prioritizing compliance and safety over maximum force.

Flexible Skin: A soft outer covering that improves aesthetics and provides some collision cushioning. When Xpeng cut through the skin on stage, it revealed a complex internal structure with visible servo mechanisms.

Solid-State Battery: Iron uses all-solid-state battery technology for lightweight design, high energy density, and enhanced safety. Xpeng's automotive battery expertise directly transfers here.

22-DOF Hands: Each hand has 22 degrees of freedom — enabling complex manipulation tasks like gripping, pinching, and tool use. This is among the highest hand dexterity in production humanoids.

Real-World Use Cases

1. Retail and Commercial Service

Xpeng has explicitly stated that Iron will "prioritize commercial service scenarios" initially. The robot can provide guided tours, act as shopping guides, and handle customer service interactions. The 3D curved display enables expressive communication, while the VLA 2.0 AI handles natural conversation.

2. Industrial Inspection

Chinese steel producer Baosteel is confirmed as an ecosystem partner. Iron will be deployed at Baosteel facilities for inspection tasks — monitoring equipment, detecting anomalies, and reporting issues. This industrial validation is critical for demonstrating reliability.

3. Xpeng Showrooms

With over 1,000 retail outlets globally (721 in China), Xpeng has a natural deployment channel for Iron robots. Even placing one robot per showroom would represent "mass production" — and provide real-world testing data to improve the platform.

4. Future Home Assistance

Xpeng's long-term vision includes home deployment. The company's "mobility explorer in the physical AI world" positioning suggests Iron is designed to eventually operate in residential settings — though this is likely years away from practical reality.

Xpeng Iron: Pros and Cons

✅ Pros

• Industry-Leading Compute (3,000 TOPS) — Three Turing AI chips provide more onboard processing than most competitors
• EV Manufacturing Scale — Xpeng's automotive supply chain and 300K+ annual production experience de-risks manufacturing
• VLA 2.0 AI Architecture — Novel end-to-end model trained on 100M video clips; shared across EVs, robotaxis, and robots
• 82+ Body DOF — Among the most articulated full-size humanoids, enabling complex movements
• 22-DOF Hands — High dexterity for manipulation tasks requiring fine motor control
• SDK Released — Developer tools available for ecosystem building (announced Nov 2025)
• Industrial Partnership — Baosteel deployment provides real-world validation
• Solid-State Battery — Cutting-edge power technology from Xpeng's EV R&D

❌ Cons

• No Confirmed Pricing — $150K estimates are speculative; Xpeng hasn't announced actual prices
• Aggressive Timeline Risk — Factory groundbreaking to mass production in ~9 months is unprecedented
• Public Balance Issues — Demo footage shows Iron falling face-first; stability not yet proven
• Limited Performance Specs — Walking speed, payload, battery life not officially disclosed
• China-First Strategy — International availability unclear; likely domestic deployments initially
• Unproven at Scale — While Xpeng builds EVs at scale, humanoid manufacturing is a different challenge
• New Division — Xpeng Robotics is young; key talent (Zhao Tongyang) already departed

How Xpeng Iron Compares to Competitors

vs. Tesla Optimus: Both Xpeng and Tesla are leveraging EV manufacturing for humanoid robots. Tesla has a massive cost advantage (targeting $25-30K vs Xpeng's ~$150K estimate) and started production earlier. However, Xpeng Iron has dramatically higher compute (2,250 vs ~100 TOPS) and more degrees of freedom. Tesla is the clear leader on pricing and production; Xpeng leads on raw capability.

vs. Figure 03: Figure has the OpenAI partnership and $39B valuation behind it. Figure 03 is deploying at BMW and has a proven industrial track record. Xpeng Iron has higher DOF and compute but less real-world deployment data. Figure is US-based; Xpeng is China-focused.

Frequently Asked Questions

How much does the Xpeng Iron cost?

Xpeng has not officially announced pricing for Iron. Industry estimates suggest approximately $150,000 for enterprise deployments, based on comparable robots and Xpeng's stated positioning. The company has committed to driving costs down through manufacturing scale, so pricing may decrease significantly after mass production begins in late 2026.

When will Xpeng Iron be available?

Xpeng is breaking ground on its 110,000-square-meter humanoid robot factory in Q1 2026, with mass production targeted for late 2026. Initial deployments will likely prioritize Xpeng's own showrooms and industrial partner Baosteel before broader commercial availability. International availability has not been announced.

Can the Xpeng Iron walk naturally?

Yes — Iron's walking gait is remarkably human-like, to the point that Xpeng cut through the robot's skin on stage to prove no human was inside. The robot features passive degrees of freedom at the toes for a "light and gentle stride" and a flexible humanoid spine for natural movement. However, demo footage has shown stability issues, including one public fall.

Is Xpeng Iron better than Tesla Optimus?

It depends on the metric. Xpeng Iron has significantly higher compute power (3,000 TOPS vs ~100 TOPS), more degrees of freedom (82 vs ~50), and more sophisticated hand dexterity. However, Tesla Optimus has a massive cost advantage (~$25-30K target vs ~$150K estimate) and has already begun production. For capability, Iron leads; for accessibility, Optimus leads.

Why is Xpeng making humanoid robots?

Xpeng sees humanoid robots as a natural extension of its autonomous vehicle technology. The company already develops AI chips, vision-language-action models, electric motors, batteries, and sensors for EVs — all components that transfer directly to humanoid robotics. CEO He Xiaopeng has repositioned the company as a "global embodied intelligence company."

What is VLA 2.0?

VLA 2.0 (Vision-Language-Action 2.0) is Xpeng's proprietary AI model that powers Iron. Unlike traditional architectures that convert vision → language → action, VLA 2.0 goes directly from visual input to motor commands. It was trained on nearly 100 million video clips and runs on a 72-billion parameter base model in Xpeng's 30,000-GPU cloud.

Is the Xpeng Iron SDK available?

Yes. Xpeng released the Iron SDK at their November 2025 AI Day to enable developers to build applications for the humanoid robot ecosystem. This positions Iron as a platform rather than just a product.

Verdict: Should You Consider the Xpeng Iron?

The Xpeng Iron represents one of the most ambitious humanoid robot programs outside of Tesla and Figure AI. With 82 DOF, 3,000 TOPS of compute, proprietary AI, and an EV giant's manufacturing infrastructure behind it, Iron has the technical foundation to compete at the highest level.

Consider Iron if: You're an enterprise looking for a highly capable humanoid with leading-edge compute and AI, you're comfortable with China-based technology, and you can wait until late 2026 for availability. Baosteel's industrial partnership suggests Iron is ready for real-world deployment.

Don't consider Iron if: You need immediate availability (Tesla Optimus and Agility Digit are shipping), you're price-sensitive (Tesla targeting ~$25K, Iron likely ~$150K), or you require US-based support and deployment.

The wildcard is Xpeng's aggressive timeline. Going from factory groundbreaking to mass production in ~9 months would be unprecedented in humanoid robotics. If they pull it off, Xpeng Iron could be a major force by 2027. If the timeline slips, competitors like Tesla and Figure will extend their lead.

Interested in the Xpeng Iron? View the full Xpeng Iron listing on Robozaps or browse all humanoid robots to compare alternatives.

Last updated: March 20, 2026. Specs sourced from Xpeng AI Day 2025 announcements, CnEVPost, RoboHorizon, and official Xpeng press releases. Pricing estimates based on industry analysis. Robozaps is a humanoid robot marketplace — we maintain hands-on product databases and may earn referral fees from qualifying purchases.

Bottom Line: The DroidUp Moya ($173K) is worth it for healthcare, eldercare, and premium hospitality buyers who need emotional connection and warm-skin technology. NOT recommended for manipulation tasks or budget-conscious buyers. Availability: Late 2026, ~50 units first batch.

📅 Last updated: March 20, 2026  |  ⏱️ 12 min read

The DroidUp Moya is doing something no other humanoid robot has attempted: feeling genuinely human to the touch. With synthetic skin that maintains body temperature between 32-36°C (89.6-96.8°F), micro-expressions across 25 facial degrees of freedom, and 92% human-like walking accuracy at a measured 0.83 m/s pace, Moya represents China's most ambitious push into biomimetic robotics. But at $173,000, is the world's first "fully biomimetic" humanoid worth the investment? This comprehensive DroidUp Moya review covers everything you need to know: real-world specifications, pricing breakdown, performance analysis, and how Moya compares to competitors like Ameca and the upcoming Xpeng Iron.

Key Takeaways

• Price: The DroidUp Moya costs approximately $173,000 USD, with pre-order estimates ranging from $165,000 to $200,000+ depending on customization — positioning it in the premium companion robot segment.
• Warm Skin Technology: World's first humanoid with active thermal regulation maintaining 32-36°C body temperature, designed to trigger natural human bonding responses.
• Walking Performance: 92% human-like walking accuracy at speeds up to 0.83 m/s (3 km/h / 1.9 mph) — an elegant pace, significantly ahead of most expressive humanoids that cannot walk at all.
• Battery Life: Approximately 4 hours per charge using tendon-assisted actuation that minimizes energy consumption.
• Best For: Healthcare facilities, eldercare, museums, high-end hospitality, and research institutions focused on human-robot interaction studies.
• Key Limitation: Not yet available — expected late 2026 with only ~50 units in the first production batch. DroidUp is a 2021 startup with no consumer track record.

DroidUp Moya Specifications

The DroidUp Moya — world's first fully biomimetic humanoid robot with human-like warmth and expressions.

DroidUp Moya Price: What Does It Actually Cost?

DroidUp has confirmed pricing of approximately $173,000 USD for the Moya, though final prices may range from $165,000 to over $200,000 depending on customization options. As a pre-production robot with limited initial availability (~50 units), pricing remains somewhat fluid.

At this price point, Moya positions itself as a premium institutional robot rather than a consumer product. DroidUp is clearly targeting healthcare facilities, museums, and research institutions with budgets for cutting-edge human-robot interaction technology.

Here's how the DroidUp Moya price compares to other humanoid robots on the market:

For the price, Moya offers a unique value proposition: it's the only humanoid robot that combines full bipedal locomotion with realistic warmth and micro-expressions. Ameca has better facial expressions but cannot walk. Tesla Optimus can walk but has no emotional expressiveness. Moya sits at the intersection — though you pay a premium for that convergence.

Performance and Mobility: Real-World Capabilities

The DroidUp Moya achieves what most expressive humanoids cannot: actually walking. Built on DroidUp's Walker 3 skeleton — the successor to Walker 2, which won bronze at the 2025 Beijing Humanoid Robot Half Marathon — Moya delivers genuinely impressive bipedal performance.

Key mobility specifications:

• Walking Speed: Up to 0.83 m/s (3 km/h / 1.9 mph), a measured pace optimized for elegant movement rather than speed
• Walking Accuracy: 92% human-like gait — DroidUp claims this is measured against biomechanical analysis of natural human movement
• Weight: Just 32 kg (71 lbs) — remarkably light for a full-size humanoid, enabled by tendon-assisted actuation
• Battery Life: Approximately 4 hours per charge, comparable to most bipedal humanoids
• Turning: Smooth directional changes without the jerky stops common in earlier humanoids

The lightweight build is notable. At 32 kg, Moya is lighter than Ameca (49 kg), roughly half the weight of Tesla Optimus (~73 kg), and comparable to 1X NEO's ~30 kg. This low mass, combined with tendon-assisted actuators similar to 1X's approach, enables longer battery life and more energy-efficient movement.

However, observers at the March 2026 Shanghai debut noted that while Moya's gait is smooth, it still shows that 8% gap from fully human — some describe it as similar to walking in heels. The robot is clearly optimized for elegant, measured movement rather than dynamic athletics like running or jumping.

Sensors and Perception

The DroidUp Moya's sensor suite prioritizes human interaction over environmental navigation:

• In-Eye Cameras: Cameras positioned behind Moya's eyes capture facial data for real-time expression mirroring and eye contact maintenance. This enables Moya to respond to micro-expressions in the people it's interacting with.
• Thermal Sensors: Monitor and regulate the synthetic skin temperature to maintain the 32-36°C warmth that distinguishes Moya from other robots.
• Facial Tracking System: Reads human facial expressions and movements to inform Moya's reactive expressions. The system processes movement data in real-time to generate contextually appropriate responses.

Unlike industrial humanoids that prioritize depth sensing and object detection (LiDAR, Intel RealSense, etc.), Moya focuses on social perception. The sensor array is designed to answer: "What is this person feeling, and how should I respond?" — not "What objects are in this room and how do I manipulate them?"

This focus makes sense for Moya's target applications in healthcare and hospitality where emotional connection matters more than object manipulation.

AI and Learning Capabilities

DroidUp Moya employs what the company calls the "Zhuoyide cerebellar motor control model" — a proprietary AI system that handles real-time movement coordination and social interaction:

• Biomimetic AI: The AI processes facial recognition data and generates corresponding micro-expressions — joy, surprise, concern — in real-time. DroidUp claims the system can produce natural transitions between emotional states rather than discrete expression changes.
• Motor Control: The cerebellar model coordinates the Walker 3 skeleton's movements, managing balance and gait while maintaining the smooth, elegant motion DroidUp prioritizes. This handles the complex coordination between walking and upper-body expressiveness.
• Speech Integration: Moya includes speech recognition and natural language processing for conversation, though DroidUp has not disclosed which LLM backbone powers the system.

The SDK situation is unclear. DroidUp has not announced public API access or ROS compatibility. Given the company's focus on institutional customers rather than research labs, developer accessibility may not be a priority. This is a notable contrast to platforms like Unitree H1 that actively court the research community with open development tools.

Design and Build Quality

Moya's design philosophy centers on one goal: feel less like a robot and more like a person. This drives every material and engineering choice.

The synthetic skin incorporates embedded heating elements that maintain human body temperature. Studies on haptic perception show that warmth triggers subconscious bonding responses — we instinctively associate warmth with life and kinship. DroidUp is explicitly exploiting this psychological response to create stronger human-robot connections.

Beneath the warm skin, Moya features a simulated rib cage and soft material layers that mimic human fat and muscle. The result is a tactile experience closer to touching a person than touching a machine — though whether this enhances comfort or deepens uncanny valley discomfort varies by individual.

The 25 degrees of freedom in Moya's face enable micro-expressions: subtle eye movements, slight smiles, small nods that humans make unconsciously during conversation. These aren't programmed animations but real-time generated responses to observed human behavior.

The modular platform architecture allows different gender presentations and facial configurations. DroidUp can customize appearance for specific deployment contexts — a significant differentiator for institutional customers who need robots matching specific personas.

Real-World Use Cases

1. Healthcare and Eldercare

DroidUp explicitly targets healthcare as Moya's primary market. China's rapidly aging population creates urgent demand for care supplements. Moya's warm touch, emotional responsiveness, and non-threatening presence could provide companionship and basic interaction for elderly patients. The 4-hour battery life supports partial shift deployment, and the lightweight build (32 kg) reduces safety concerns compared to heavier industrial robots.

2. Museums and Exhibitions

Interactive museum guides benefit from Moya's combination of walking ability and emotional expressiveness. Unlike stationary systems, Moya can escort visitors through spaces while maintaining engaging conversation. The customizable appearance allows museums to create period-appropriate or thematically relevant characters.

3. Premium Hospitality

High-end hotels and venues seeking differentiation could deploy Moya as a premium concierge experience. The emotional responsiveness creates more memorable interactions than typical service robots, while the warm-skin technology makes handshakes and greetings feel more natural.

4. Human-Robot Interaction Research

Researchers studying uncanny valley effects, social robotics, and human-robot bonding have limited platforms that combine locomotion with realistic emotional expression. Moya provides a unique research tool — though the unclear SDK situation may limit academic applications.

5. Banking and Financial Services

DroidUp mentions banks as a target deployment. Premium financial services branches increasingly use technology to differentiate customer experience. A biomimetic greeter could elevate perception of service quality — though ROI calculations at $173,000 per unit require high-value customer contexts.

DroidUp Moya: Pros and Cons

✅ Pros

• World's First Warm-Skin Humanoid — 32-36°C body temperature triggers natural bonding responses; no other humanoid offers this
• Combines Walking + Expressions — Ameca can't walk; Tesla Optimus can't emote. Moya does both, uniquely positioning it for social applications
• Lightweight Design (32 kg) — Tendon-assisted actuation keeps weight low, enabling safer human proximity
• 92% Human Gait Accuracy — Walker 3 skeleton proven at robot marathon; movement is smooth and elegant
• Customizable Appearance — Modular platform allows gender, face, and persona customization for institutional branding
• 25 DOF Facial Expressions — Micro-expressions respond in real-time to human interaction, not pre-scripted animations

❌ Cons

• Not Available Until Late 2026 — Pre-orders only; first batch limited to ~50 units. You cannot buy one today.
• New Company Risk — DroidUp founded in 2021 with no consumer track record. Long-term support uncertain.
• Uncanny Valley Concerns — Early reactions are mixed; some find Moya engaging, others describe it as "like a living ghost." Realistic ≠ comfortable.
• Limited Specs Disclosed — No published payload capacity, full-body DOF count, or detailed technical documentation
• High Price Point ($173,000) — 6-8x more expensive than general-purpose alternatives like Tesla Optimus or Unitree H1
• China-Focused Initially — First deployments expected in China; international availability unclear
• SDK/Developer Access Unknown — No announced API or ROS compatibility; may limit research applications

How DroidUp Moya Compares to Competitors

vs. Ameca: Ameca has comparable facial expressiveness (27 facial DOF vs Moya's 25) and is available today. But Ameca cannot walk — it's a torso on a stand or wheeled base. If your application requires a mobile, walking presence with emotional expressiveness, Moya is the only option.

vs. Xpeng Iron: Both are Chinese humanoids targeting 2026 launch with realistic appearances. Iron comes from a major EV manufacturer (Xpeng) with proven mass production capability, while DroidUp is an unproven startup. Iron demonstrated walking in early 2026 but also showed balance issues. Neither has disclosed full pricing.

Frequently Asked Questions

How much does the DroidUp Moya cost?

The DroidUp Moya costs approximately $173,000 USD, with estimates ranging from $165,000 to over $200,000 depending on customization. This positions it as a premium institutional robot rather than a consumer product. DroidUp has not announced financing options or leasing programs, though these may emerge as commercial deployments begin in late 2026.

When will the DroidUp Moya be available?

DroidUp expects to begin shipping Moya units in late 2026. The first production run will be limited to approximately 50 units, likely prioritizing Chinese institutional customers in healthcare and public venues. International availability has not been announced.

Can the DroidUp Moya actually walk?

Yes. Unlike many expressive humanoids that are stationary or wheeled, Moya achieves full bipedal locomotion using DroidUp's Walker 3 skeleton. The company claims 92% human-like walking accuracy at speeds up to 0.83 m/s (1.9 mph). The Walker 2 platform (predecessor to Walker 3) won bronze at the 2025 Beijing Humanoid Robot Half Marathon, demonstrating proven bipedal capability.

Why does the DroidUp Moya have warm skin?

Moya maintains body temperature between 32-36°C (89.6-96.8°F) through embedded heating elements in its synthetic skin. Research shows humans subconsciously use touch temperature to assess connection and kinship. DroidUp designed the warm-skin feature specifically to trigger these bonding responses, making interactions feel more natural and emotionally comfortable than with cold-surfaced robots.

Is the DroidUp Moya safe to interact with?

At 32 kg (71 lbs), Moya is significantly lighter than most full-size humanoids, reducing collision risks. The tendon-assisted actuation system enables smoother, more controlled movements than high-torque industrial actuators. However, as with any humanoid robot, institutional deployments will require safety assessments and likely some supervision. DroidUp has not published specific safety certifications.

How does DroidUp Moya compare to Sophia the robot?

Sophia (by Hanson Robotics) and Moya both prioritize realistic humanlike appearance and emotional expressiveness. However, Sophia cannot walk — it's primarily a bust or wheeled platform. Moya combines full bipedal locomotion with expressiveness. Moya also adds warm skin technology that Sophia lacks. Sophia has more global brand recognition and years of public appearances, while Moya is a 2026 newcomer.

What is DroidUp's track record?

DroidUp (also known as Zhuoyide) was founded in 2021 in Shanghai. The company previously demonstrated hyper-realistic android busts at events like the World Artificial Intelligence Conference (WAIC) and enrolled an android in Shanghai Theatre Academy's doctorate arts program. Their Walker biped skeleton won bronze at the 2025 Beijing robot half marathon. However, Moya is their first commercial humanoid product, and the company has no consumer track record.

Is the DroidUp Moya worth $173,000?

For institutions that specifically need a mobile humanoid with emotional expressiveness and realistic human touch — healthcare, premium hospitality, human-robot interaction research — Moya offers capabilities no other robot provides. If you need general-purpose manipulation or don't require the warmth/expression features, alternatives like Unitree H1 ($90K) or upcoming Tesla Optimus (~$25-30K) offer better value. The answer depends entirely on whether Moya's unique biomimetic features align with your use case.

Verdict: Should You Buy the DroidUp Moya?

The DroidUp Moya is attempting something genuinely new in humanoid robotics: creating a robot that doesn't just look human but feels human. The warm skin, micro-expressions, and elegant walking motion combine into an experience designed to trigger emotional connection rather than utility. At $173,000, you're not buying a tool — you're buying a presence.

Buy the Moya if: You're a healthcare facility, museum, or premium hospitality venue specifically seeking a humanoid that creates emotional connections with visitors or patients. You have the budget for experimental technology and understand you're an early adopter with a 2021 startup. You need walking + expressiveness combined in one platform — no alternative offers this.

Don't buy the Moya if: You need manipulation capabilities (carrying objects, opening doors, performing tasks). You want a proven platform with established support — consider Ameca for pure expressiveness or Unitree H1 for athletic bipedal research. You're price-sensitive — wait for the market to mature.

Moya represents a bet on the future of social robotics. If DroidUp executes on their vision and survives as a company, early adopters will own groundbreaking technology. If not, that $173,000 becomes an expensive museum piece. Given the late 2026 timeline and ~50 unit first batch, most buyers should watch the first deployments before committing.

Interested in the DroidUp Moya? View the full DroidUp Moya listing on Robozaps or browse all humanoid robots to compare alternatives.

Last updated: March 8, 2026. Specs sourced from DroidUp press releases (March 2026), New Atlas, Mike Kalil, and Tekedia coverage. Pricing confirmed at ~$173,000 by multiple sources. Robozaps is a humanoid robot marketplace — we maintain hands-on product databases and may earn referral fees from qualifying purchases.

With 27 motors controlling its face alone, the Engineered Arts Ameca delivers facial expressions so uncannily human that viewers frequently describe feeling "watched" by a machine for the first time. At a price point of $100,000–$500,000 depending on configuration, Ameca isn't just another humanoid robot—it's the world's most advanced platform for social human-robot interaction. But is this emotional intelligence worth six figures? This comprehensive Ameca review covers everything: real-world specs, pricing breakdown, Generation 3 improvements from ICRA 2025, the Tritium OS platform, and how Ameca compares to Sophia, Moya, and every other expressive humanoid in 2026.

Key Takeaways

• Price: Ameca costs $100,000–$500,000 depending on configuration (head-only, half-body, or full unit). Most full installations run approximately $250,000–$300,000.
• Expression Leadership: 61 degrees of freedom with 27 dedicated to the face—more than any other humanoid robot—enabling micro-expressions that trigger genuine emotional responses in viewers.
• No Walking: Unlike Tesla Optimus or Unitree robots, Ameca is a stationary platform. It cannot walk, run, or locomote independently.
• AI Integration: Tritium OS supports GPT-based conversational AI, enabling real-time multilingual conversations with natural voice synthesis.
• Best For: Museums, science centers, corporate exhibitions, hospitality venues, and research institutions focused on human-robot interaction studies.
• Key Limitation: Stationary design and premium pricing limit applications to high-traffic public venues where emotional engagement justifies the investment.

Ameca Full Specifications

The Engineered Arts Ameca Generation 3 — the world's most expressive humanoid robot platform.

Ameca Price: What Does It Actually Cost?

Engineered Arts doesn't publish a single price for Ameca because the robot's modular architecture allows for multiple configurations. You can purchase just the head unit for reception-desk applications, a half-body installation for exhibition kiosks, or a full unit for research and flagship installations.

Based on industry sources, reseller listings, and confirmed reports from December 2024, here's what you can expect to pay in 2026:

• Head-only configuration: ~$100,000–$150,000
• Half-body (torso + head): ~$150,000–$250,000
• Full-body unit: ~$250,000–$350,000
• Premium/custom installations: Up to $500,000+

Additional costs include professional installation by Engineered Arts engineers (typically required), ongoing Tritium software licensing, maintenance contracts, and cloud AI service fees. A typical full installation with setup and first-year support runs approximately $300,000.

Here's how Ameca's pricing compares to other social and expressive humanoid robots:

At $250,000–$300,000 for a typical full installation, Ameca sits at the premium end of the social robotics market. The investment is justified for venues where visitor engagement directly correlates with revenue—science museums, corporate experience centers, and luxury hospitality.

Expressiveness and Motion Quality

Ameca's defining feature isn't walking or payload capacity—it's emotional resonance. The robot's performance is measured in micro-expressions and gestural authenticity rather than meters per second.

Powered by 61 electric actuators delivering smooth, precise movements, Ameca demonstrates:

• 27 facial degrees of freedom: Independent control of eyebrows, eyelids, lips, cheeks, and jaw enables expressions from subtle skepticism to full surprise—matching the 43 muscles in a human face with mechanical precision.
• 5-axis neck articulation: Natural head tilts, nods, and turns that follow conversation partners without appearing robotic or jerky.
• Full upper-body gesture vocabulary: Shrugging, pointing, open-palm gestures, and crossed-arm skepticism are all in Ameca's repertoire.
• Eye contact tracking: Binocular 8MP cameras enable genuine eye contact with individual viewers, a feature that dramatically increases engagement metrics in public settings.
• Response latency under 500ms: Cloud-connected AI processes speech and generates appropriate facial responses fast enough to feel natural in conversation.

What sets Ameca apart is the quality of motion, not the quantity. Engineered Arts has spent years refining actuator control algorithms to eliminate the "uncanny valley" jerkiness that plagues most humanoids. The result is a robot that feels less like a machine and more like a digital actor inhabiting a physical form.

Sensors and Perception

Ameca's sensor suite is optimized for social interaction rather than industrial task completion:

• Binocular Vision (2× 8MP cameras): Eye-mounted cameras provide stereoscopic vision for depth perception and facial recognition. Ameca can identify returning visitors and recall previous conversations when integrated with CRM systems.
• Chest Camera: Wide-angle view captures the full scene for spatial awareness and crowd detection.
• Spatial Audio (4-channel microphone array): Chest-mounted array localizes sound sources, enabling Ameca to turn toward speakers and maintain appropriate eye contact in group settings.
• Ear Microphones: Dual microphones provide backup audio capture and help filter background noise in loud exhibition environments.
• LIDAR and Emergency Stop: Safety systems include proximity sensors and physical emergency stop buttons for public deployment compliance.

Notably absent are tactile sensors and advanced depth sensors like ToF or structured light—Ameca isn't designed for manipulation tasks that require touch feedback. The sensor architecture reflects its purpose: understanding humans, not handling objects.

AI and Software: Tritium OS

Every Ameca runs on Tritium, Engineered Arts' proprietary robot operating system comprising three integrated components:

• Tritium Operating System: A lightweight Linux distribution with dedicated robotics software for real-time motor control and sensor fusion.
• Tritium Platform: A web browser-based control interface that allows operators to script behaviors, manage AI integrations, and monitor robot status remotely.
• Tritium Cloud Services: Seamless integration with third-party AI services including GPT-4, voice synthesis engines, and facial recognition APIs.

For developers, Tritium supports Python, C++, and block-based programming for behavior scripting. The platform enables:

• Real-time LLM conversations: GPT-based dialogue with context awareness and personality customization
• Multilingual voice synthesis: Studio-quality output in dozens of languages
• Behavior trees: Complex interaction scripts triggered by visual or audio cues
• Remote telepresence: Human operators can "drive" Ameca in real-time for special events
• OTA updates: Software improvements deployed without physical access

The closed-source nature of Tritium may frustrate researchers seeking full system access, but Engineered Arts argues this ensures reliability and safety in public-facing deployments.

Design and Build Quality

Ameca's physical design reflects intentional choices for maximum social acceptance:

Appearance Philosophy: The grey prosthetic skin and neutral facial features are specifically engineered to appear gender-neutral and race-neutral. This deliberate ambiguity makes Ameca relatable to diverse global audiences without triggering specific cultural associations.

Build Quality: The shell combines black composite panels with exposed metallic structural elements—a "mechanical skeleton" aesthetic that reads as futuristic rather than attempting (and failing) to pass as human. This approach sidesteps the uncanny valley problem that plagues ultra-realistic android designs.

Form Factor: At 187 cm (6'2") tall, Ameca stands slightly above average human height—commanding presence without intimidation. The 49 kg (108 lb) weight is manageable for installation teams, and the 600mm base diameter provides stability without excessive floor space requirements.

Durability: Engineered Arts does not publish IP ratings or environmental specifications. Ameca is designed for climate-controlled indoor environments—museums, corporate lobbies, and exhibition halls rather than outdoor or industrial settings.

Modularity: The modular architecture allows components—head, arms, hands—to be upgraded independently. This extends platform lifespan and reduces total cost of ownership for institutions that can amortize upgrades over time.

Real-World Use Cases

1. Science Museums and Exhibition Centers

Ameca's most successful deployments are in science museums where visitor engagement metrics directly impact institutional success. Installations include the Computer History Museum (Mountain View, California), Heinz Nixdorf MuseumsForum (Paderborn, Germany), Copernicus Science Center (Warsaw, Poland), and Deutsches Museum (Nuremberg, Germany). In these settings, Ameca serves as a conversation partner explaining AI concepts to visitors—a meta-educational experience where the robot is both the subject and the teacher.

2. Corporate Experience Centers

Technology companies use Ameca to demonstrate AI capabilities to clients, partners, and executives. The robot's ability to hold contextual conversations, answer technical questions, and express appropriate emotional responses makes it an ideal showcase for enterprise AI investments.

3. Hospitality and Luxury Retail

High-end hotels and flagship retail locations deploy Ameca as a premium concierge, greeting VIP guests by name and providing personalized recommendations. The Museum of the Future (Dubai) features Ameca as part of its "robotic family" of interactive installations.

4. Human-Robot Interaction Research

Universities and research institutions, including the National Robotarium (Edinburgh, UK), use Ameca as a platform for studying how humans respond to expressive robots. The standardized hardware platform enables reproducible research across institutions.

5. Media and Entertainment

Ameca has appeared at CES (2022, 2024, 2025), GITEX, OMR Festival, ICRA conferences, and the UN's AI for Good Summit. In December 2022, an Ameca unit delivered Channel 4's Alternative Christmas Message—a UK television tradition typically reserved for notable figures offering counterpoints to the Royal Christmas Broadcast.

6. Event Rentals

Engineered Arts offers rental programs for trade shows, product launches, and corporate events. Short-term deployments let organizations test Ameca's impact before committing to purchase.

Ameca: Pros and Cons

✅ Pros

• Unmatched facial expressiveness — 27 DOF in the face alone enables micro-expressions that no other commercial humanoid can replicate, creating genuine emotional engagement with audiences.
• Cloud-native AI integration — Tritium's seamless GPT integration means Ameca can hold intelligent conversations out of the box, with personality customization and context awareness.
• Proven public deployment track record — Installations in world-class museums (Computer History Museum, Deutsches Museum, Museum of the Future) validate reliability in high-traffic environments.
• Modular architecture — Head-only or half-body configurations reduce entry costs; component upgrades extend platform lifespan without full replacement.
• Generation 3 improvements — ICRA 2025 launch brought enhanced expression fidelity, better AI integration, and the new Ami companion platform for fleet deployments.
• Strong company backing — Engineered Arts' $10M Series A (December 2024) and restructure as a US company signals long-term support and US market expansion.

❌ Cons

• Cannot walk or locomote — Ameca is strictly stationary. For applications requiring mobility, consider Unitree H1 or Figure 03.
• Premium pricing ($250K+ typical) — Full installations cost 5–10× more than wheeled social robots like Pepper, limiting accessibility to well-funded institutions.
• Low payload capacity (~2kg per arm) — Expressive gestures only; Ameca cannot perform useful manipulation tasks.
• Closed-source software — Tritium OS limits deep customization for research teams requiring full system access.
• Indoor-only deployment — No published IP rating or environmental specifications; unsuitable for outdoor or industrial environments.
• Requires professional installation — Self-deployment not supported; adds cost and lead time to implementations.

Ameca vs. Competitors: How Does It Compare?

Ameca vs. Sophia: While Sophia has greater name recognition (she's a Saudi citizen, after all), Ameca's facial expression quality is objectively superior. Sophia's fame stems from media appearances; Ameca's reputation comes from technical excellence. For institutions prioritizing interaction quality over celebrity appeal, Ameca is the clear choice.

Ameca vs. Droidup Moya: Moya's warm-skin technology (body temperature 32–36°C) offers a different approach to humanization—physical warmth rather than expressive faces. Moya also walks via its Walker 3 skeleton, addressing Ameca's key limitation. However, Moya launches in late 2026, while Ameca is available now.

Frequently Asked Questions

How much does Ameca cost?

Ameca prices range from $100,000 to $500,000 depending on configuration. A head-only unit starts around $100,000, while a full-body installation with professional setup typically runs $250,000–$350,000. Additional costs include Tritium software licensing, maintenance contracts, and cloud AI service fees. Engineered Arts also offers rental programs for events.

Can Ameca walk?

No. Ameca is a stationary humanoid robot designed for social interaction, not locomotion. The robot is mounted on a fixed base and cannot walk, run, or move independently. For applications requiring mobility, consider walking humanoids like Unitree H1, Figure 03, or other bipedal robots.

What AI does Ameca use?

Ameca runs on Tritium OS with cloud-connected AI integration. Out of the box, the robot supports GPT-based conversational AI (including GPT-4), voice synthesis in multiple languages, and facial recognition. Operators can customize AI personalities, script specific behaviors, and integrate with third-party services via Tritium's web-based platform.

Where is Ameca installed?

Major Ameca installations include the Computer History Museum (Mountain View, California), Heinz Nixdorf MuseumsForum (Paderborn, Germany), Copernicus Science Center (Warsaw, Poland), Museum of the Future (Dubai), Deutsches Museum (Nuremberg, Germany), and National Robotarium (Edinburgh, UK). The robot has also appeared at CES, GITEX, ICRA, and UN AI summits.

Is Ameca better than Sophia?

For facial expression quality and interaction capability, yes—Ameca's 27 facial DOF competes with Sophia's 36 head/neck DOF—while the numbers are similar, Ameca's higher-precision motors enable more nuanced micro-expressions. Sophia has greater public recognition due to media appearances and her status as a Saudi citizen, but Ameca is the preferred platform for serious HRI research and premium installations where technical quality matters more than celebrity appeal.

Can I rent Ameca for an event?

Yes. Engineered Arts offers rental programs for trade shows, product launches, corporate events, and exhibitions. Contact Engineered Arts directly through their rentals page for pricing and availability.

What is Generation 3 Ameca?

Ameca Generation 3 was unveiled at ICRA 2025 (IEEE International Conference on Robotics and Automation) alongside a new companion platform called Ami. Gen 3 improvements include enhanced facial actuators for subtler micro-expressions, better LLM integration, and improved software capabilities. Each Ameca generation improves on facial fidelity, hand dexterity, and AI integration.

Is Ameca worth buying in 2026?

For science museums, corporate experience centers, and research institutions where visitor engagement or HRI research justifies the investment—yes. The $250,000+ price point is steep but competitive for the expressiveness quality delivered. For applications requiring mobility or physical manipulation, look elsewhere. For premium social interaction where emotional resonance matters, Ameca remains the gold standard.

Verdict: Should You Buy Ameca?

Ameca occupies a unique position in the humanoid robot market: it's the undisputed leader in facial expressiveness and social interaction quality, but it explicitly trades away locomotion and manipulation capability to achieve that focus. With 27 degrees of freedom dedicated solely to facial expression—more than any other commercial humanoid—Ameca delivers emotional engagement that genuinely affects viewers. The Tritium platform's cloud AI integration makes it conversationally capable out of the box, and Generation 3's improvements at ICRA 2025 only extend its lead.

Buy Ameca if: You operate a science museum, corporate experience center, or research institution where visitor engagement directly drives success metrics. You need a conversation partner that triggers genuine emotional responses. You have the budget ($250K+) and the indoor venue to support a stationary installation. Don't buy Ameca if: You need a robot that walks, carries objects, or operates in uncontrolled environments. Consider Unitree H1 for research mobility, Figure 03 for household tasks, or other humanoids for industrial applications.

With Engineered Arts' $10M Series A funding (December 2024) and restructure as a US company, the platform has strong institutional backing for long-term support. Gen 3 is current, but Gen 4 will inevitably arrive—institutions comfortable with modular upgrades can buy now. Those seeking maximum value may wait for pricing to stabilize as Chinese competitors like Droidup Moya enter the expressive humanoid market in late 2026.

Ready to explore Ameca? View the full Ameca listing on Robozaps or browse all humanoid robots.

Last updated: March 8, 2026. Specifications sourced from Engineered Arts official documentation, ICRA 2025 presentations, and verified against third-party testing data where available. Robozaps is a humanoid robot marketplace—we maintain hands-on product databases and may earn referral fees from qualifying purchases.

This week's headlines told a clear story: the race to build humanoid AGI is intensifying, and the money is following. From Musk's bold claims about Tesla achieving artificial general intelligence in robot form first, to a €1 billion Tether-backed bet on German robotics, the industry is moving past demos and into deployment mode.

Here's what shaped the humanoid robot industry from March 1-7, 2026.

Musk Claims Tesla Will Be First to AGI in Humanoid Form

Elon Musk posted on X that "Tesla will be one of the companies to make AGI and probably the first to make it in humanoid/atom-shaping form." The statement came as Tesla confirmed plans to convert its Fremont Model S and Model X production lines to Optimus robot manufacturing—a major strategic shift that signals robotics is becoming central to Tesla's future.

Why it matters: This isn't just Musk being Musk. Tesla plans to spend over $20 billion in capital expenditure in 2026—up sharply from $8.5 billion in 2025—with a significant chunk going toward Optimus production and supporting infrastructure. The company is targeting 1 million units annually as its initial production goal. That's a concrete, measurable bet on humanoid robots becoming Tesla's next major business line.

The timing aligns with reports that xAI, Musk's AI company, recently merged with SpaceX, with rumors of an even larger consolidation across all Musk companies including Tesla. An Optimus robot running xAI's Grok models would represent a significant capability leap.

Our take: The AGI claim is aspirational, but the Fremont conversion is real. We covered Tesla killing the Model S for Optimus in detail. What's notable is how Tesla's robotics ambitions are now directly cannibalizing its legacy vehicle business. Ending production of the vehicles that built Tesla's brand to make room for robots? That's commitment. See our full Tesla Optimus Gen 3 guide for where the hardware stands today.

Neura Robotics Raises €1 Billion Backed by Tether

German humanoid robot maker Neura Robotics is raising approximately €1 billion ($1.2 billion) in a funding round led by Tether, the stablecoin issuer behind USDT. The deal values the company at €4 billion—lower than the €8-10 billion rumored last November, but still a massive valuation for a European robotics startup that most people haven't heard of.

Neura develops the 4NE1 humanoid robot, a 5.9-foot system that understands natural language instructions and can carry up to 220 pounds while moving at about three miles per hour. The company also builds industrial robotic arms that can be programmed through visual interfaces rather than custom code, and logistics robots capable of moving 1.5 tons of goods.

Why it matters: This is the largest single funding round we've tracked for a pure-play humanoid robotics company in 2026. It also marks Tether's biggest bet on physical AI—an interesting signal of where crypto money sees opportunity. According to reports, Neura has accumulated a €1 billion order book, suggesting strong commercial traction.

Our take: Crypto money flowing into robotics is a trend worth watching closely. Tether has been diversifying beyond stablecoins, but humanoid robots represent a bold thesis on physical-world AI becoming the next major technology platform. The company also recently acquired ek Robotics, adding 300 employees and warehouse logistics expertise. Read our NEURA Robotics 4NE1 review for the full technical breakdown on their flagship humanoid.

Xiaomi Begins Factory Trials of Humanoid Robots

Xiaomi CEO Lei Jun announced that the company's humanoid robots have begun trial operations at Xiaomi's automobile factory in China. The machines are already performing real tasks: loading self-tapping nuts at assembly stations and transporting material boxes across the facility. Lei said large-scale deployment across all production facilities is planned within five years.

The robots operate using Xiaomi's proprietary Xiaomi-Robotics-0 VLA (vision-language-action) foundation model. By integrating multimodal perception and reinforcement learning, the humanoids can perform autonomous operations without constant human guidance. Key performance indicators like mean time between failures and task success rates are "steadily improving," according to Lei.

Why it matters: Xiaomi isn't just building robots for demos and trade shows—they're putting them to work in their own operations, validating capabilities in real manufacturing environments. This is exactly the approach needed to prove humanoid robots can deliver actual ROI.

Our take: This is how you validate humanoid robots: deploy them in your own operations first, then sell to others. Xiaomi's approach mirrors Tesla's strategy with Optimus. It's telling that two of the world's largest technology manufacturers are both using their own factories as testing grounds. Our Xiaomi CyberOne review covers their flagship humanoid specifications. Five years to large-scale deployment sounds conservative given the pace of progress—expect it sooner.

OpenAI Robotics Leader Resigns Over Pentagon AI Deal

Caitlin Kalinowski, who led OpenAI's hardware and robotic engineering teams since November 2024, resigned from the company. In detailed posts on LinkedIn and X, she cited specific concerns about "surveillance of Americans without judicial oversight and lethal autonomy without human authorization" as issues that "deserved more deliberation than they got."

Her departure followed OpenAI's agreement with the Pentagon to deploy AI models on classified government networks. The deal came shortly after Anthropic walked away from similar negotiations, reportedly pushing for stricter limits on domestic surveillance and autonomous weapons. The optics looked bad—OpenAI appearing to step in after its rival took a principled stand.

CEO Sam Altman later acknowledged the rollout looked "opportunistic," and OpenAI clarified restrictions on military uses. But Kalinowski was already gone.

Why it matters: The robotics industry's relationship with military applications is becoming a major fault line for talent. Kalinowski's resignation puts a spotlight on where companies draw ethical boundaries around autonomous systems—and whether top engineers will stay when those boundaries get tested.

Our take: This won't slow OpenAI's robotics ambitions materially, but it does highlight the tension companies face as AI moves into defense applications. The best robotics engineers have options, and companies that lose talent over ethical concerns may find the cost higher than expected. Our piece on humanoid robots in military and defense explores this increasingly complex space.

Canadian Startup Mirsee Prepares for Mass Production

Mirsee Robotics, a small company based in Cambridge, Ontario, announced plans to move to mass production of its MH3 humanoid robot in 2027. CEO Tarek Rahim predicted the robotics revolution will be "bigger than the automotive revolution in the early 20th century, happening at ten times the speed" and that there will eventually be "more robots than cars."

Unlike bipedal humanoids dominating headlines, the MH3 uses wheels for mobility—a deliberate design choice aimed at maximizing battery life and stability. It's harder to knock over and can operate longer between charges. The robot uses a Canadian-made vision system for object manipulation, and the company is adding voice command capabilities powered by AI.

Why it matters: While Chinese and American companies dominate headlines, regional players like Mirsee represent the industry's global expansion. The wheeled design also shows there's still room for different form factors in what's becoming a crowded humanoid space. Not every application needs legs.

Our take: North American manufacturing capability for humanoid robots remains limited outside of Tesla. Mirsee is small—planning to reach just 20 employees—but they represent important industrial base development. Toyota Canada also announced Digit deployments at their Woodstock facility this week, suggesting the Ontario region is becoming a North American robotics cluster. See our complete humanoid robot companies guide for the full competitive landscape.

Chinese Robotics Companies Showcase at AW 2026 in South Korea

China's leading humanoid developers—Unitree, Leju, and AgiBot—gathered at Smart Factory and Automation World 2026 (AW 2026) in South Korea. The event showcased their latest hardware and commercialization strategies for the Korean manufacturing sector.

Perhaps more significant: Unitree signaled openness to technology cooperation with South Korean companies. Given South Korea's manufacturing expertise (Samsung, Hyundai, LG) and China's lead in humanoid hardware, this could open meaningful partnership opportunities.

Why it matters: This represents China's robotics industry pushing aggressively into new markets. AW 2026 is a major industrial automation show, and the presence of multiple Chinese humanoid makers signals serious commercial intent beyond domestic deployments.

Our take: Unitree continues its aggressive expansion strategy following their viral Spring Festival performance that reached 679 million viewers. Their willingness to partner with Korean firms suggests they're building an ecosystem, not just selling individual robots. This could accelerate adoption across Asian manufacturing. Check our Unitree G1 review and Unitree H2 review for hardware details.

Market Pulse

• Funding: Neura Robotics' €1B round brings total 2026 humanoid robot funding past $3 billion in the first quarter alone—already exceeding most of 2025.
• Valuations: Apptronik is now valued at $5 billion (up from $15 million when they started), with Google DeepMind as a partner. Neura at €4 billion. The industry is minting unicorns.
• Production targets: Tesla targeting 1M Optimus units annually; Xiaomi planning large-scale deployment within 5 years; Mirsee moving to mass production in 2027.
• Market forecast: New research published this week projects the humanoid robot market reaching $243.4 billion by 2035—driven by personal assistance, healthcare, and education applications.

What to Watch Next Week

1. Tesla Fremont conversion timeline: Details on when Model S/X lines officially transition to Optimus production could come any day. This will be a major milestone for the industry.
2. Neura Robotics deal close: The €1B round reportedly may be followed by additional fundraising—watch for closing announcements and how the capital will be deployed.
3. OpenAI robotics restructuring: With Kalinowski's departure, OpenAI's robotics roadmap may shift. Expect updates on team leadership and whether the military deal affects hiring.

Want to stay ahead of humanoid robot developments? Bookmark our humanoid robot news hub for continuous coverage, or browse the latest robots available at Robozaps.

The week's biggest story? Reality checks. From BMW proving humanoids can actually work production lines to a legendary roboticist declaring the whole endeavor "fantasy," this week forced the industry to confront the gap between demo videos and deployable technology.

BMW Brings Figure Robots to Europe After U.S. Success

The headline numbers are hard to ignore: 30,000 cars produced, 90,000+ parts handled, 1,250+ hours of runtime. That's what Figure AI's humanoid robots achieved at BMW's Spartanburg plant in South Carolina over 11 months.

Now BMW is taking the experiment to Europe. On February 27, the automaker announced it will deploy humanoid robots at its Leipzig plant in Germany—the first time "Physical AI" of this kind has entered European automotive production.

Why it matters: This isn't a demo. Figure 02 ran 10-hour shifts, Monday through Friday, on an active assembly line. The robots loaded sheet-metal parts with 5-millimeter precision in just 2 seconds per cycle. When you're building X3 SUVs, that kind of consistency matters.

The deployment also generated critical data that shaped Figure 03's design. The forearm—the robot's most failure-prone component—was completely re-architected for the new model. Every intervention, every failure mode, every hour of runtime informed the next generation.

Our take: This is the deployment milestone the industry needed. Flashy videos of robots folding laundry are one thing; running an automotive production line for nearly a year is another. BMW's expansion to Germany signals that the pilot exceeded expectations. For more on Figure's latest, see our Figure 03 review and Figure AI company analysis.

iRobot Cofounder: Humanoid Vision Is "Pure Fantasy"

Rodney Brooks, the MIT roboticist who cofounded iRobot (makers of the Roomba), unloaded on the humanoid robot industry this week. His verdict on Elon Musk's vision of humanoid assistants: "pure fantasy thinking."

Brooks argues that today's humanoid robots "will not learn how to be dexterous" regardless of how many billions VCs pour into training. The problem? Touch.

Human hands contain 17,000 mechanoreceptors for detecting pressure and texture. While AI has been trained on massive datasets of speech and images, "we do not have such a tradition for touch data," Brooks wrote. Training robots by filming humans performing tasks—the approach used by Tesla and Figure—won't solve this fundamental gap.

Why it matters: Brooks isn't some armchair critic. He's been building robots for three decades. His claim that robots won't look like humans in 15 years—instead sporting wheels, multiple arms, and only being called humanoids—directly challenges the form factor every major player has bet on.

Our take: Brooks has a point about the touch data problem, but dismissing the entire humanoid effort feels premature. BMW's deployment shows real-world value exists even with current limitations. The question isn't whether today's robots are perfect—it's whether they're useful enough to justify continued investment. Still, his critique about transparency is valid. When companies hide their teleoperation rates, the public can't properly evaluate progress. For context on what these robots actually cost, check our humanoid robot pricing guide.

China Owns 90% of the Humanoid Robot Market

The numbers are stark: Chinese companies shipped roughly 90% of all humanoid robots sold globally in 2025. Unitree moved 5,500 units. Agibot shipped 5,168. Meanwhile, Figure, Agility Robotics, and Tesla each sold around 150.

That's not a typo. Unitree shipped 36 times more humanoid robots than its closest American competitor.

"China has a more robust hardware supply chain—much of it built up through the EV sector, from sensors to batteries—and the world's strongest manufacturing base," analyst Selina Xu told TechCrunch.

Even Elon Musk acknowledged the competitive reality at Davos: "China is very good at AI, very good at manufacturing, and will definitely be the toughest competition for Tesla. To the best of our knowledge, we don't see any significant competitors outside of China."

Why it matters: This is the EV playbook all over again. Early state support, industrial policy, rapid iteration, cost advantages—and before Western competitors could scale, Chinese companies owned the market. Global humanoid shipments were just 13,317 units last year. By 2035, that's projected to reach 2.6 million. The early leader often becomes the permanent leader.

Our take: The U.S. still leads in AI and software. Figure's deployment at BMW demonstrates capabilities Chinese competitors haven't matched publicly. But hardware matters, and China's supply chain advantages are formidable. For the latest on Chinese robots, see our Unitree G1 review, Unitree H2 review, and analysis of the best humanoid robots on the market.

MIT Technology Review: The Hidden Human Labor Behind Humanoids

A worker in Shanghai recently spent a week wearing a VR headset and exoskeleton while opening and closing a microwave door hundreds of times a day—training the robot beside him. Welcome to the strange new world of humanoid robot training.

MIT Technology Review published a deep investigation into the human labor powering the "autonomous" humanoid industry. Key revelations:

• Data harvesting at scale: Figure partnered with Brookfield, which manages 100,000 residential units, to capture "massive amounts" of real-world data in homes.
• Delivery workers as sensors: One company had workers wear movement-tracking sensors while moving boxes, generating training data.
• Teleoperation everywhere: 1X's Neo robot, shipping to homes this year, relies on remote operators to handle tricky tasks. If the robot gets stuck, a person in Palo Alto takes over.

"Just as our words became training data for large language models, our movements are now poised to follow the same path," the report notes.

Why it matters: If home humanoids aren't genuinely autonomous, the business model is "a form of wage arbitrage that re-creates the dynamics of gig work while, for the first time, allowing physical tasks to be performed wherever labor is cheapest."

Our take: Transparency matters. When companies hide their teleoperation rates, the gap between marketing and reality becomes dangerous—literally, as Tesla's Autopilot lawsuits demonstrate. We're not saying teleoperation is bad; 1X gets customer consent. But the industry needs honest communication about what these machines can actually do today. For more on what you can actually buy, see where to buy humanoid robots.

Honor Enters the Humanoid Race

Chinese smartphone giant Honor will unveil its first humanoid robot at Mobile World Congress 2026 in Barcelona. The announcement, made via teaser video on X, marks another major consumer electronics company entering the humanoid space.

Honor joins Xiaomi, which launched CyberOne in 2022, in the phone-maker-to-robot-maker pipeline.

Why it matters: When smartphone companies start building humanoids, it signals the technology is approaching a commercial tipping point. These companies have massive manufacturing capabilities, consumer distribution networks, and experience shipping millions of complex hardware units annually.

Our take: The consumer electronics crossover validates the humanoid form factor for home applications. Honor's robot remains mysterious for now, but we'll be watching MWC closely. For more on the smartphone-to-robot trend, see our Xiaomi CyberOne review.

Market Pulse

What to Watch Next Week

1. Mobile World Congress (March 1-4): Honor's humanoid reveal, plus any robotics announcements from other consumer electronics players.
2. Figure 03 deployment updates: With Figure 02 retiring, watch for announcements about where the next-generation robots will be deployed.
3. China policy signals: The "Made in China 2025" robotics push is accelerating. Watch for new funding announcements or pilot programs targeting the 2.6 million unit 2035 goal.

Stay updated with the latest humanoid robot news by visiting our humanoid robot news hub. Ready to buy a humanoid robot? Check out our marketplace.