IROS 2025 Keynotes - Humanoid Robot Systems: Xingxing Wang

Why It Matters

The announcement signals a move toward mass‑market humanoid robots, combining lower cost hardware with advanced AI, which could reshape automation across consumer, industrial, and entertainment sectors.

Key Takeaways

• Company launched G1, a 1.3‑meter low‑cost humanoid robot.
• New R1 model is lighter, pre‑orders open, shipments later this year.
• Demonstrated coordinated performance of 16 robots in large‑scale show.
• Introduced Dector hand and boxing capabilities for complex tasks.
• Emphasized data collection, teleoperation, and video‑to‑action AI research.

Summary

Xingxing Wang’s IROS 2025 keynote highlighted his firm’s rapid evolution from early Kodrader platforms to a diversified humanoid portfolio. Since its 2016 founding, the company unveiled the full‑size H1, then the compact 1.3‑meter G1 in 2023, and most recently the ultra‑light R1, now accepting pre‑orders with shipments slated for later this year. Wang emphasized that the G1 has become the world’s most shipped humanoid, while the R1 aims to broaden accessibility for non‑specialist users.

The presentation showcased several technical milestones: a synchronized performance of sixteen robots executing complex choreography, a new Dector hand enabling fine manipulation, and a boxing‑style demonstration where the robot can combine any of twenty distinct actions. Speed records were also cited, with the H1 achieving five meters per second and winning four championships in recent sports‑robot competitions. Algorithmic upgrades now allow the robots to recover from falls and learn new motions within one to two days.

Wang underscored the importance of large‑scale data collection and whole‑body teleoperation, noting collaborations with NVIDIA’s ASIC Dream platform and open‑source reinforcement‑learning tools. He pointed to video‑generation models as a promising but compute‑intensive avenue for translating visual cues into robot actions, urging larger firms to invest in embodied AI research.

The broader implication is a shift toward affordable, high‑performance humanoids that can operate in everyday environments, from entertainment to logistics. By lowering hardware costs and advancing AI‑driven control, the company positions itself to accelerate market adoption and stimulate cross‑industry partnerships, while also highlighting the current bottleneck of computational resources for embodied AI.

Original Description

"Keynote Title: ""A New Era of Global Collaboration in Intelligent Robotics"" Speaker Biography

Wang Xingxing, Founder, CEO, and CTO of Unitree, is a technology purist. While studying for his master’s degree, he independently pioneered the development of the XDog – a high-performance quadruped robot dog powered by low-cost external rotor brushless motors. The creation of XDog set a technological precedent for cost-effective and high-performance legged robots, leading to widespread coverage from domestic and international media in 2016 and generating significant attention and impact within the global robotics community. In August 2016, Wang founded Unitree, which became the first company in the world to publicly retail high-performance quadruped robots. Unitree has consistently led global sales over the years and has significantly accelerated the commercialization of high-performance quadruped and humanoid robots worldwide. Wang has filed over 200 domestic and international patent applications, and has secured multiple rounds of investment from major investors such as Meituan, HongShan, Shunwei Capital, and Matrix Partners China. He also led the team in showcasing the company's product on premier global stages such as the CCTV Spring Festival Galas (Year of the Ox and Year of the Snake), the Beijing Winter Olympics Opening Ceremony, and the Super Bowl. Abstract

This presentation will briefly cover recent technological developments and achievements of Unitree and our global customers. Furthermore, it will outline a vision for international collaboration to accelerate the growth of the intelligent robotics industry in the next few years.

"