X Square

Human-scale architecture and dexterous hands

X SquareI’s humanoid design emphasizes human-scale reach, force-controlled arms, and bimanual manipulation. Public product pages for Quanta X2 list a 172 cm body, 756 mm arm reach, and 6 kg end-effector payload, indicating a platform sized for household furniture, hotel workspaces, and light industrial stations. 

End-to-end embodied AI

The firm promotes an end-to-end pipeline that couples perception, action generation, and control within the WALL model family. Marketing materials state a goal of fine-grained manipulation, with autonomous perception and decision-making packaged for “multi-scenario” execution across domestic and commercial settings. 

Open-source elements (WALL-OSS)

X SquareI maintains an open repository for parts of its stack—WALL-OSS / wall-x—covering data preparation, model configuration, and evaluation utilities for embodied policies. This code base underpins reproducible research and integrator experimentation while the company advances proprietary models for product deployments. 

Technology and Specifications

The WALL embodied foundation model

• Model family: WALL-A (closed) and WALL-OSS (open components) represent a foundation-model approach tailored to robots, with claims of “world-leading” manipulation performance across multiple metrics (as per official materials).

• Learning pipeline: The open repository describes end-to-end pipelines for dataset tooling (e.g., LeRobot), flow-matching action branches, and FAST branches for high-frequency control, suggesting a design that merges vision-language-action inference with low-latency execution.

Mechatronics and control

• Arms and hands: Force-controlled arms and anthropomorphic grippers aim at delicate grasping of soft and deformable objects (textiles, packaging, dishware). The 6 kg end-payload spec is aligned with common service/assistance tasks rather than heavy industrial payloads.

• Perception: Company communications emphasize autonomous perception (RGB-D cameras and proprioceptive sensors are typical for the class), feeding the WALL policy for long-horizon routines like clearing tables, loading dishwashers, or resetting rooms. (Inference based on class norms, corroborated by product language about autonomous perception and fine manipulation)

• Latency and safety: WALL-OSS utilities and the general embodied-AI framing indicate on-device, closed-loop control to handle contact-rich interactions with rate-limited motion and compliance—common requirements for safe operation near people.

Representative specs (Quanta X2)

• Height: ~172 cm

• Arm reach: ~756 mm

• End-payload: ~6 kg

• Parameter scale: “10+ billion” parameters referenced for the learning stack (contextual to the model family rather than just one device).

Applications and Use Cases

Hospitality and front/back-of-house service

Targeted tasks include surface cleaning, tableware handling, linen management, room reset, and simple concierge logistics. The bimanual form factor and compliance are aimed at operating in human-dense spaces with frequent contact and clutter. (Use cases consistent with company positioning and public demos common in the sector) 

Domestic assistance

Home-oriented sequences—loading dishwashers, folding laundry, organizing countertops—require reliable grasping and fine manipulation in varied lighting and layouts; the WALL approach is pitched at this generalization problem. 

Light industrial and logistics support

With 6 kg end-payload and human-scale reach, Quanta-class systems fit machine tending, bin picking of consumer items, and kitting in structured stations, often complementing conveyors or autonomous mobile bases. (Industry fit inferred from typical specs and the company’s multi-scenario messaging) 

Advantages / Benefits

• End-to-end policy learning: A single model that maps perception to action can simplify task hand-engineering, accelerating coverage across routines and objects. 

• Open-source on-ramp: WALL-OSS lowers experimentation barriers for labs and integrators, encouraging replication and benchmarking.

• Human-scale suitability: A 172 cm platform with compliant manipulation is dimensioned for existing human environments, reducing the need for custom fixtures.

• Capital and momentum: Multiple 2025 reports cite $100M–$140M in new financing led by Alibaba-related entities and others, supporting rapid hiring and industrialization. 

Comparisons (if relevant)

In the general-purpose humanoid landscape, X SquareI’s positioning blends foundation-model AI with service-first manipulation:

• Versus research-centric peers: Some competitors prioritize bipedal agility and locomotion metrics; X SquareI’s materials highlight fine-grained manipulation and routine completion in service contexts. (High-level comparison based on public focus areas) 

• Versus closed stacks: The release of WALL-OSS provides transparent components for data, training, and evaluation, contrasting with fully closed ecosystems. 

Pricing and Availability

X SquareI does not list a universal MSRP on its public pages. Availability and price depend on configuration, support package, and region (education/research kits vs. commercial pilots). 2025 coverage frames the company as scaling production and partnering with investors to accelerate deployments, but buyers should request formal quotations that confirm hardware specs, service levels, software features, and compliance. 

FAQ Section

What is X SquareI?
X SquareI (also known as X Square Robot / X2Robot) is a Chinese robotics company developing embodied-AI humanoids and the WALL foundation model series for general-purpose manipulation. 

How does X SquareI’s technology work?
The company trains end-to-end policies that map perception to action for contact-rich tasks. Open materials (WALL-OSS) detail data prep and model branches for high-frequency control, while closed models (e.g., WALL-A) target production performance. 

Why is X SquareI important?
It combines foundation-model learning with human-scale hardware, aiming to bridge the gap between lab demos and practical routines in hospitality, home, and light industry. Reported 2025 funding indicates strong momentum.

What are the benefits of X SquareI’s approach?
Key benefits include fine-grained manipulation, an open-source on-ramp for integrators (WALL-OSS), and human-scale dimensions that fit existing environments. 

Does X SquareI open-source everything?
No. The company publishes selected components (WALL-OSS) while maintaining proprietary models and tooling for commercial deployments. 

What are headline specs for Quanta X2?
Public pages cite ~172 cm height, 756 mm arm reach, and 6 kg end-payload, with a learning stack reportedly at 10B+ parameters.

Summary

X SquareI represents a foundation-model approach to humanoid robotics, pairing the WALL learning stack with human-scale, force-controlled hardware designed for real-world chores and services. With open-source components (WALL-OSS) for researchers and integrators, plus continuing investment and reported pilot activity, the company is positioned as one of the notable Chinese entrants pushing humanoids from demo-ready to deployment-ready—especially where fine manipulation, safety, and routine completion matter most.