RoboParty Launches RP1: The Open-Source Humanoid Challenging Tech Giants

SHANGHAI – October 15, 2026 — As artificial intelligence rapidly transitions from digital screens to the physical world, a silent war is brewing over the hardware that will carry it. While venture-backed tech giants construct walled gardens around their proprietary humanoid robots, a Shanghai-based startup is attempting to tear the gates down. Today, RoboParty officially launched RP1 (ROBOTO 01), a high-performance, full-stack open-source humanoid robot designed to democratize advanced robotics research and embodied AI development.

Unveiled globally at the IROS 2026 conference late last month with a dynamic "Kick Me" demonstration proving its stability and disturbance rejection, the RP1 is now officially hitting the market. The platform is offered in two configurations: a 1.3-meter-tall Standard version with 24 degrees of freedom (DoF) for gait and balance research, and a Pro version boasting 30 DoF—including a two-DoF neck and dual seven-DoF arms—tailored for complex human-robot interaction and end-to-end policy training.

Yet, the most significant feature of the RP1 is not its physical agility, but its intellectual transparency. In an industry dominated by closed ecosystems, RoboParty is releasing the mechanical computer-aided design (CAD) files, bill of materials (BOM), printed circuit board (PCB) schematics, and low-level firmware that make the robot tick.

Busting Open the Embodied AI Black Box

For years, academic researchers and independent developers have faced a severe bottleneck. Commercial humanoids from industry leaders like Figure, Tesla, and Unitree offer impressive capabilities but operate as technological black boxes. When a robot fails to execute a complex task, researchers are often left guessing whether the flaw lies in their AI learning model, the low-level control policy, the actuator response, or the physical mechanical structure.

RoboParty aims to eliminate this guesswork by providing a transparent ecosystem. Alongside the hardware specifications, the company is opening access to its proprietary software stack, which includes the PartyOS embodied AI system, the UFO unsupervised reinforcement learning framework, and MimicLite motion imitation technology.

"We believe the future of embodied AI should not be defined by just a handful of closed systems," said Yi Huang, founder of RoboParty, in the company's press release. "With RP1, we want to open up more than code. We want to give researchers the opportunity to truly understand, modify and continuously advance robotic capabilities. When more teams can innovate on a shared technical foundation, the entire industry can move further forward."

By granting access to the low-level drivers and the in-house-developed ROMOMO joint actuator modules, RoboParty is enabling a new paradigm of research. Developers can now trace problems across all layers of the system, optimizing hardware and software in tandem rather than treating the physical robot as an unchangeable variable.

The Hidden Economics of Open Hardware

While the promise of an open-source humanoid is alluring, the practical realities of hardware manufacturing present a formidable challenge. The software industry experienced its "Linux moment" decades ago because code is infinitely replicable at zero marginal cost. Hardware, however, is bound by the laws of physics, supply chain logistics, and manufacturing tolerances.

RoboParty's previous open-source initiative, ROBOTO ORIGIN (RPO), launched in January 2026 and quickly garnered over 2,600 GitHub stars. RPO served as a proof of concept, boasting a BOM cost of approximately $6,800. However, it was explicitly a prototype-grade platform. Replicating it required access to CNC machining for custom aluminum parts, creating a barrier to entry for smaller labs or individual makers.

RP1 represents a significant leap in complexity. The robot relies on RoboParty's proprietary ROMOMO actuators, which deliver a peak joint torque of up to 160 N·m. Manufacturing high-performance actuators requires specialized equipment, tight tolerances, and expertise in mechatronics that most software-focused AI labs simply do not possess.

Industry analysts note that while releasing the CAD files for these actuators is a monumental step for transparency, it does not automatically equate to manufacturability. To address this, RoboParty is adopting a "progressive openness" strategy. The complete RP1 stack is not being dumped onto GitHub overnight; rather, the company plans a phased rollout of resources throughout the fourth quarter of 2026, culminating in a mass-production program.

This strategy reveals the dual nature of RoboParty's business model. By fostering a massive open-source community—already numbering over 5,000 developers—the company is building an "Embodied Infrastructure." Ultimately, while the designs are free, many research institutions will likely opt to purchase fully assembled ROMOMO actuators or complete RP1 units directly from RoboParty, generating revenue through core component supply and Joint Design Manufacturing (JDM) services.

China’s Open-Source Robotics Ambition

The launch of RP1 cannot be viewed in a vacuum; it is deeply intertwined with China's broader push to establish global dominance in the robotics supply chain and AI standards. RoboParty's ascent has been nothing short of meteoric. Founded in early 2025 by Huang—a 21-year-old prodigy who previously built an open-source bipedal robot in his university dorm—the company has rapidly secured formidable financial backing.

By July 2026, RoboParty had raised nearly $70 million across multiple funding rounds, drawing investments from heavyweights such as Matrix Partners China, Shunwei Capital, and Xiaomi Strategic Investment. Xiaomi's involvement is particularly notable, given the tech giant's extensive experience in hardware manufacturing and ecosystem building. This capital influx has allowed RoboParty to attract top-tier engineering talent from institutions like Tsinghua University, Stanford, and Carnegie Mellon, accelerating the development of the RP1 platform.

The strategic implications of a Chinese startup providing the foundational open-source hardware for global AI research are profound. Open-source platforms like Stanford's Mobile ALOHA have made significant strides in democratization, but they primarily focus on wheeled manipulation rather than complex bipedal locomotion. RP1 fills a massive void in the market, offering a high-performance bipedal baseline that could become the default standard for academic and commercial research worldwide.

If RoboParty successfully navigates the transition from progressive open-source releases to a reliable, mass-produced hardware ecosystem, it could fundamentally alter the competitive dynamics of the robotics industry. By lowering the barrier to entry, RP1 enables smaller teams to focus on high-level intelligent policies and complex task execution, rather than wasting years engineering basic motor skills and joint mechanics.

The true test for RP1 will unfold over the coming months as independent researchers begin downloading the CAD files, compiling the firmware, and attempting to replicate the robot's impressive physical feats in their own labs. If the global community can successfully bridge the gap between open-source theory and hardware reality, the black box of embodied AI may finally be broken wide open.