American company creates operating system for Chinese humanoid robots

“A crowd came, kids ran away, cars stopped, then the police showed up, because there were humanoid robots in the park,” Jan Liphardt, founder and CEO of OpenMind in California, told SCMP about the “riot” scene last year, when the company filmed a commercial for its Unitree G1 humanoid robot in a park in San Francisco. “If G1 passes by a street in Shanghai, people probably won’t stop to look, because they’ve seen this before.”

According to Liphardt, the contrast highlights the opportunity for OpenMind. “China is clearly ahead of the US in many aspects when it comes to robots,” Liphardt commented. “Hardware is the key factor. They have a fully integrated supply chain and fierce domestic competition.”

The founder of OpenMind believes that the world needs a common platform for humanoid robots instead of being fragmented like today. According to him, to prepare well for the future of living with machines, people should understand how they think and should help them think better.

“Robots today are stuck in the single ecosystem of the supplier that makes them, limiting collaboration and unable to adapt to the complexities of the real world,” Liphardt told Interesting Engineering.

OM1 was born based on this goal. As an AI-native operating system for robots developed from scratch by OpenMind, the platform can run on any robotic hardware from quadrupeds, humanoid robots, wheeled robots to drones, acting as a universal brain for machines.

Unlike traditional robotic platforms, OM1 is specifically designed for intelligent and adaptive behavior. The combination of context awareness with AI and memory models allows robots to “interpret” their surroundings, “understand” natural language, and make decisions in real time.

Late last year, OpenMind introduced FABRIC, a blockchain application protocol that allows robots to verify identities and share context with each other decentralized. FABRIC enables robots to securely identify themselves, verify location and share knowledge with “unfamiliar machines”, allowing them to collaborate seamlessly in real time – essential for robots deployed in open environments.

Establish an open ecosystem for machines to learn from each other

On its website, OpenMind says the focus is not just on a single robotics platform. Instead, the company wants to establish an open ecosystem where machines can collaborate and learn from each other. By collecting data from real-life usage, OpenMind aims to identify the most potential applications, where robot capabilities best match human needs.

OpenMind currently cooperates with a series of Chinese humanoid robot manufacturers, including names that are ambitious to reach the global market, such as Unitree Robotics, UBTech Robotics, AgiBot, LimX Dynamics, Booster Robotics or Engine AI. Last year, the company raised $20 million to build an open, decentralized system connecting smart machines, including Pantera Capital and Pi Network Ventures (the fund behind the controversial cryptocurrency project Pi Network).

Liphardt assessed that Chinese robot products have a “significant presence” in Silicon Valley and top US universities. Collaborations between companies like OpenMind and Chinese robot manufacturers help them reduce barriers to entry in the US market.

“Using software developed and hosted in the US, where data and cloud computing operations are local, will help Chinese hardware manufacturers reduce regulatory barriers,” he said.

However, Liphardt said the OpenMind platforms developed will only focus on situations where “robots interact directly with humans”, mainly in the retail, hospital and school sectors. He also affirmed that he would not join the industrial sector because “that opportunity has passed”.

The head of OpenMind has high hopes for a world where humans and robots coexist, but admits there is still a long way to go. “At least in the US, we’re still in the phase where most people are encountering humanoid robots for the first time. Just the presence of a humanoid robot walking on the street is enough to cause awe,” he added.

Theo Rude Baguettethe advances achieved by OpenMind represent a significant change in the way robotic systems are conceived and developed with the goal that “robots must learn together”. By prioritizing software solutions and open collaboration, the company is setting new standards for this nascent industry.

However, goals also come with challenges. As OM1 and similar platforms gain popularity, the company may have to address a series of questions, such as how this smart machine “integrates” into daily life, or what ethical issues need to be addressed as robots become more and more autonomous.