Astribot Raises New 1 Billion Yuan Round, Accelerating Mass Production of Cable-Driven Robots | Yunqi Capital

Astribot chose the less-traveled path of cable-driven robotics, becoming the world's first company to mass-produce cable-driven AI robots. Around this, the company built a trinity full-stack architecture of "Lumo Embodied Foundation Model + Astribot OS + Cable-Driven Body," forming a closed technical loop of "body + data + model."

Recently, the company completed a new funding round of RMB 1 billion. In this edition of Yunqi Partners, we share the latest updates on Astribot, a project that Yunqi Capital led in the angel round.

The following content is adapted from Astribot

Recently, Astribot, the world's first embodied intelligence company to mass-produce cable-driven AI robots, completed a new funding round of RMB 1 billion. The company's valuation reached RMB 15 billion, firmly placing it in the first tier of China's embodied intelligence companies.

This round was led by Shenzhen Investment Holdings Donghai (Shenzhen Investment Holdings), with participation from key state-owned and financial investors including Hongtai APlus, Ninghai Xingtaihe Fund managed by Guanghua Yangguang, and Shanheng Capital. Existing shareholder Xianmu also made additional investments.

"Under the Design For AI philosophy, Astribot remains committed to using AI to define robots and using intelligence to unleash human creativity, advancing long-term value through globally leading talent, product strength, and commercialization — building a future world where humans and machines coexist, co-create, and win together," said Jie Lai, founder and CEO of Astribot. "Our investors are important 'fellow travelers.' This round reflects recognition and market consensus for Astribot's core philosophy and its trinity full-stack architecture route of 'AI model - embodied OS - cable-driven body,' and will serve as solid support for accelerating product breakthroughs and commercial landing."

According to sources, the funds from this round will be used for model iteration, talent expansion, commercialization, and production capacity building, with continued investment in the trinity full-stack architecture, persistently tackling world-class challenges and driving the conversion of technical capabilities into deliverable products and scalable deployment.

Four Funding Rounds in Six Months

Valuation Rose from RMB 10 Billion to RMB 15 Billion in Three Months

This marks Astribot's fourth funding round in six months. Public information shows that since March 2026, Astribot has completed A+, A++, B, and this current round in succession. In June 2026, the company completed its B round series with total financing exceeding RMB 1 billion and valuation first surpassing RMB 10 billion; just three months later, it announced another round with valuation reaching RMB 15 billion. A case of "four rounds in six months, valuation up RMB 5 billion in three months" is rare even in the booming embodied intelligence track.

Previously, Astribot received investments from Yunqi Capital, Matrix Partners China, Jinqiu Fund, Ant Group, and CASID, among others. Following this round, Astribot's shareholder structure has formed a diverse, high-quality base of "state-owned funds + industrial groups + leading financial institutions."

"From underlying AI large models and embodied OS to unique cable-driven body design, the Astribot team has demonstrated extremely strong closed-loop technical capabilities and a clear commercialization path," said Mu Chunwei, CEO of Shenzhen Investment Holdings Pengcheng Strategic Emerging Industries Fund. The investment in Astribot is based on confidence in the scarcity of the company's "full-stack self-research" in the embodied intelligence track — not only aligning with the fund's direction of deploying frontier hard tech, but also serving as an important step in helping Shenzhen build an embodied intelligence industry highland and cultivate new quality productive forces. "We look forward to Astribot accelerating the movement of AI robots from laboratories to thousands of industries, truly achieving scaled commercial application."

Market analysis suggests that diverse strategic shareholders are expected to provide support in policy guidance, industrial scenarios, and funding, with deep resource sharing. Astribot is anticipated to receive systematic resource empowerment, further strengthening its moat and corporate value.

"Investors are important 'fellow travelers' and 'enablers,'" said Lai. With the empowerment of various strategic shareholders, Astribot will continue building higher technical and product barriers, accelerating production capacity construction, scaled application, and global layout.

"AI Model + Embodied OS + Cable-Driven Body" Full-Stack Architecture

World's First Company to Mass-Produce Cable-Driven AI Robots

As a physical AI company oriented toward the real world, Astribot uses general-purpose humanoid robots as its product vehicle, with core capabilities covering embodied intelligence models, robot operating systems, and cable-driven bodies. Its core value and scarcity lie in: continuous iteration capability of full-stack AI systems, differentiation and mass-production barriers of cable-driven technology, and commercial growth potential validated by real-world scenarios.

"We need to design robots that can explore and learn in the physical world around AI's capabilities, rather than designing robots first and then adding AI on top," said Lai.

With a "Design for AI" hardware-software integrated design philosophy, Astribot has built a trinity full-stack architecture through full-stack self-research of the Lumo embodied foundation model, Astribot OS, and cable-driven body — forming a closed technical loop of "body + data + model." This architecture enables collaborative improvement of understanding, decision-making, and execution around real tasks, giving the company mechanisms for continuously accumulating data, improving models, and expanding applications, while also providing a foundation for the development of software, application, and model services.

"Cable-driven technology provides robots with flexible, precise execution capabilities more suitable for human-robot interaction, and provides high-quality data for model training. The Lumo model handles environment understanding, reasoning, and action generation. Astribot OS organizes these capabilities into applications that users and developers can call upon. The synergy of the three enables robots to both complete real tasks and continuously improve using data accumulated during deployment," Lai explained.

According to sources, Lumo-2, the foundation model released by Astribot in July this year, is the industry's first implicit world action model oriented toward home scenarios. In more than 22 complex evaluations conducted using real machines, it ranked first among evaluated models in the combined dimensions of temporal reasoning, physical understanding, and operational complexity.

At the body level, the accumulated cable-driven technology in product performance, model training, and engineering replication barriers, along with commercial mass-production capabilities, has built differentiated advantages and first-mover mass-production barriers in hardware for Astribot — making it the "world's first company to achieve mass production of cable-driven AI robots," and at least one generation ahead of 1X, the cable-driven route company invested in by OpenAI.

Accelerating Global Application and Commercial Layout

S Series Opens Thousand-Unit Orders, Combined S and T Production Capacity Reaches Ten Thousand Units

Leveraging its full-stack accumulation in cable-driven technology, Astribot has fully run through the development path from prototype validation to scaled mass production, demonstrating rare product iteration and cost control capabilities: In September 2025, the company's flagship S series opened thousand-unit orders; in June 2026, its new-generation humanoid robot T1 began shipping, becoming the first to bring high-precision manipulation robots to the RMB 90,000 price range. Combined with the company's synchronized ten-thousand-unit production capacity, this further widens the generational gap in technical differentiation, product普及 speed, and cost barriers.

In commercial scenario deployment, Astribot's applications cover retail, cultural tourism, industry, research, and education. Clients include Jinma Entertainment, ThunderSoft, Boden Intelligence, SEER Robotics, CCTV, and JD.com, among others. Partnerships with Jinma Entertainment and SEER Robotics will deploy thousands of AI robots in commercial and industrial scenarios, while collaboration with ThunderSoft will drive thousand-unit scale deployment, leveraging the latter's global sales network covering North America, Europe, and Asia to accelerate overseas expansion and global layout.

The self-developed robot retail service store "Robot MART" has been batch-deployed in six cities including Beijing, Shanghai, and Guangzhou since November 2025, pushing robots from single-point demonstrations to continuous operations.

In AI developer ecosystem collaboration, multiple AI companies including Ant Lingbo Technology, Deep Leap, Shengshu Technology, and Juna Brain are using Astribot's platform for model deployment and scenario development, driving iteration of their own product capabilities and expansion of application ecosystems.

New Product Development Accelerates Alongside Technical Talent Expansion

"Commercial Services" and "Developer Ecosystem" Drive Commercial Growth

Astribot stated it will continue focusing on building capabilities in talent acquisition, technical advancement, model iteration, production capacity, and commercialization.

Public information shows that leading domestic AI talent Sun Peng officially joined Astribot in September this year, with primary responsibility for technical R&D and application exploration in robot reinforcement learning. The company's open positions have grown substantially year-over-year in 2026, continuously attracting scarce talent in high-barrier directions including VLA, world models, and dexterous manipulation.

In R&D direction, Astribot's newly released online reinforcement learning framework SmoothRL enables robots to continuously optimize their behavior during task execution, reducing dependence on offline training and better handling uncertainty in the real world.

In commercialization, Astribot stated it will continue advancing along dual paths of "commercial services" and "developer ecosystem" — for the commercial services market, integrating robot bodies, AI models, and embodied operating system AOS full-stack capabilities into scenario solutions for retail, guided tours, and artistic performances, validating product capabilities through real operations, accumulating user feedback, and沉淀 reusable interaction, manipulation, and safety control capabilities to lay groundwork for expansion into home service scenarios. For research institutions, AI, and industry application developers, Astribot provides standardized robot platforms, skill modules, and development tools to lower barriers for robot application development, leveraging ecosystem partners' industry knowledge to expand product capabilities and application scenarios.

As an important direction in long-term product layout, Astribot's product series for home and individual users are already in development. As installed base and usage frequency grow, Astribot plans to expand software and application services for its embodied operating system AOS, continuously exploring commercialization models for models.

"We hope that through robots + AI, we can unleash human creativity," said Lai. Astribot will adhere to its Physical AI full-stack route, turning every real deployment into fuel for continuous model and product evolution, bringing truly usable and useful embodied intelligence products to global users.