"Boundless Dynamics" Closes 300 Million Yuan Series A, Co-Led by HSG and Linear Capital | Linear Portfolio

General-purpose embodied intelligence robotics company Anyverse announced today that it has raised 300 million yuan in its first angel round, co-led by HSG and Linear Capital, with Hillhouse, Horizon Robotics, and others participating.**

**

Its founder and CEO, Yufeng Zhang, has a complete track record from core technology R&D to large-scale commercial deployment**. He previously served as Vice President at Horizon Robotics, President of the Intelligent Vehicle Business Unit, board member, and member of the executive management committee, leading his team through the full cycle from technology development to mass delivery of tens of millions of units.**

** **

Harry Wang, Founder and CEO of Linear Capital, said: "Yufeng is one of the rare builders who has deeply experienced the entire journey from technology to product, product to market, and market to scale. We look forward to advancing together with Anyverse and accelerating the deep integration of technology across industries."**

On November 10, general-purpose embodied intelligence robotics company Anyverse completed its first angel round of 300 million yuan, co-led by HSG and Linear Capital, with Hillhouse, Horizon Robotics, Hua Capital, Eastern Bell Capital, BV Baidu Ventures, and Tongge Ventures participating, garnering recognition from top-tier financial investors and diverse strategic capital in its very first funding round. At the same time, an angel+ round is nearing completion, bringing total funding to over 500 million yuan.

Founded in 2025 in Beijing, Anyverse focuses on building a "general brain" and "operational intelligence" for robots, breaking through critical bottlenecks in hand-eye-brain coordination to transform embodied intelligence into broadly deployable, continuously evolving infrastructure. The company pursues a dual-track approach of general foundational model R&D and domain-specific expert model deployment, committed to providing global customers with integrated hardware-software embodied intelligence solutions with high reliability.

Yufeng Zhang, Founder and CEO of Anyverse

Founder and CEO Yufeng Zhang has a complete track record from core technology R&D to large-scale commercial deployment, making him a cross-disciplinary entrepreneur with both technical depth and engineering management capabilities. He previously held R&D management roles at headquarters of global technology leaders including Sony and ARM, building deep foundational technical expertise. He joined Horizon Robotics in 2017, serving as Vice President, President of the Intelligent Vehicle Business Unit, board member, and member of the executive management committee, where he led a thousand-person team to achieve breakthroughs in intelligent driving software and algorithm products, rapid engineering deployment, and scaled delivery — capturing the top market share for assisted driving in Chinese domestic-brand passenger vehicles. He also spearheaded multiple strategic partnerships with leading domestic and international enterprises including BYD, Changan Auto, Volkswagen Group, and Continental, demonstrating exceptional international expansion and industrialization capabilities.

Anyverse has already assembled a core team with cutting-edge innovation and engineering execution capabilities. Co-founder and CTO Wenda Xu holds a PhD in Robotics from Carnegie Mellon University, with extensive experience in autonomous driving technology development and a track record of consecutive successful entrepreneurship ventures. During his doctoral studies, he co-founded autonomous driving algorithm company Ottomatika in the United States, leading the achievement of the world's first commercial autonomous ride-hailing operation. He later served as Vice President of Autonomous Driving and Chief Algorithm Architect at Kargobot. The Anyverse team also brings together top scientists and technical leads with milestone contributions in multimodal large models, reinforcement learning, and world models, complemented by R&D and engineering teams with experience in mass delivery of tens of millions of hardware-software systems, forming full-chain capabilities from algorithmic breakthroughs to product deployment.

Embodied intelligence has entered a period of accelerated development, but industry consensus holds that achieving full general-purpose generalization will still require more than a decade. Rather than pursuing an all-at-once "omni-capable model," a more viable path is to first achieve "industry-level generalization" in select vertical domains before gradually expanding capabilities. The embodied intelligence industry currently faces a critical inflection point from laboratory demonstrations to real-scenario value generation, yet existing general models still fall far short of commercial deployment thresholds for generalization capability and task success rates in real-world environments.

Confronting this challenge head-on, Anyverse will drive embodied intelligence advancement through multi-dimensional technological innovation and systematic engineering capabilities.

On foundational models, Anyverse is systematically advancing algorithmic architecture innovation and model iteration. A truly general-purpose embodied model capable of "understanding" the physical world and driving robots to complete complex tasks across scenarios requires not only precise perception of multimodal 3D spatial information and spatiotemporal dynamics, and deep comprehension of real-world physical laws, but also real-time reasoning and decision-making in complex dynamic environments, ultimately acquiring the core capability of "learning to learn." The series of frontier explorations at the architecture level led by multiple top scientists on the team provides solid foundations for Anyverse's algorithmic innovation at the底层 of general embodied models.

Addressing the data challenges that plague embodied models, Anyverse pursues data breadth through simulation, internet video, and other methods while prioritizing systematic practice with "real machines, real people, real environments" — obtaining precise physical interaction data through real machine operation on factory production lines and in commercial spaces, combined with lightweight, flexible collection through human demonstration. Real machine data significantly improves model task success rates, while the combined application of human demonstration and diverse data types effectively enhances model generalization. This data strategy can effectively bridge the simulation-to-reality gap while enabling robots to confront real physical constraints including environmental changes and spatial limitations, continuously discovering and resolving issues in actual deployment to form a virtuous cycle of "scaled application, emergent intelligence."

To support efficient, high-quality data acquisition and foundational model innovation, iteration, and validation, Anyverse is establishing a closed-loop evolutionary system driven by dual tracks of "general foundational model R&D" and "general expert model deployment" — achieving scaled deployment through general expert models with commercial-grade task success rates, and using the real-world data and closed-loop feedback they generate to continuously drive evolution of the general foundational model in actual combat. In specific operation, drawing on the shadow mode mechanism from the intelligent driving field, the general foundational model serves as a "behind-the-scenes brain," running in parallel and silently alongside expert models in real scenarios, precisely collecting error cases and validating capability boundaries without disrupting the stability of active systems, providing feedback and drive for efficient, continuous model iteration from the real world.

Learning in application, evolving through learning, Anyverse is systematically building hardware-software products with commercial-grade reliability and task success rates. This system deeply couples continuously evolving embodied models with computing platforms, precision actuators, and real-time operating systems to construct a stable, reliable integrated robotics platform. Drawing on automotive-grade engineering experience and systematic safety assurance capabilities accumulated in the intelligent driving field, the Anyverse team ensures that every robot maintains consistently high reliability across diverse real-world scenarios, truly achieving the leap from laboratory algorithms to industrial-grade solutions.

Leveraging its efficient evolution loop and large-scale complex hardware-software system R&D and engineering capabilities, Anyverse is rapidly completing technology productization and deployment validation, driving standardized humanoid robot platforms with a unified general-purpose operational brain to achieve efficient capability migration across different scenarios. Currently, its first-generation dual-arm wheeled robot platform has achieved substantive breakthroughs in both industrial manufacturing and commercial service directions, and is advancing deep scenario deployment exploration with internationally renowned partners, with factory entry scheduled for this month.

Harry Wang, Founder and CEO of Linear Capital, said: "Yufeng is not only from a technology R&D background — his entrepreneurial experience at Horizon Robotics enabled him to transform his role through the full chain from technology to product business ownership. In a difficult and important technology product domain, he is one of the rare builders who has deeply experienced the entire journey from technology to product, product to market, and market to scale. We believe the learning ability, combat effectiveness, and organizational forging capability he demonstrated in this process will be well displayed in the Anyverse project. What the Anyverse team excels at is anchoring top-tier embodied intelligence technology in real scenarios, transforming the 'reach for the stars' dream of generalization into tangible productivity every minute. We look forward to advancing together with Anyverse and accelerating the deep integration of technology across industries."

Kai Yu, Founder and CEO of Horizon Robotics, said: "Yufeng was a core comrade-in-arms during my years at Horizon Robotics. He led his team through the full cycle from technology development to mass delivery of tens of millions of units, with deep understanding of both frontier technology exploration and industrialization. Now leading Anyverse, he is driving systematic innovation from embodied intelligence底层 algorithms and architecture to engineering deployment, which aligns highly with Horizon Robotics' philosophy of 'driving the intelligent era through hardware-software synergy.' I believe this team, with both technical depth and system-level deployment experience under his leadership, will inject new vitality into the embodied intelligence industry. I and Horizon Robotics will continue to support Anyverse, jointly accelerating the arrival of the general embodied intelligence era."

Yufeng Zhang, Founder and CEO of Anyverse, said: "This funding will significantly accelerate our dual-track path of 'foundational model R&D and expert model deployment,' enabling the team to create immediate value for customers while steadfastly advancing the long-term evolution of our general-purpose operational brain." Going forward, Anyverse will continue focusing on foundational model R&D and iteration driven by high-value scenarios, and through integrated hardware-software product innovation, collaborate with global industry partners to drive scaled application and capability leaps of embodied intelligence robots, jointly realizing our long-term vision of "making robots a perpetual force for human development."

Contact Anyverse:

Recruitment: HR@anyverse.tech

Investment Partnership: IR@anyverse.tech

Business Partnership: Business@anyverse.tech

Media Inquiries: PR@anyverse.tech