Making the Case for Crypto to the World’s Top Robotics Minds

At the 2025 Conference on Robot Learning (CoRL) in Seoul, BitRobot co-founder Michael Cho took the stage before an audience of the world’s leading robotics researchers.

He opened with a line that drew a few gasps in surprise: “Crypto will drastically accelerate embodied AI research.”

The skepticism was expected, and welcome. The presentation was not about persuading a thousand roboticist in the audience to buy a hype crypto token. It was about presenting a case for how crypto’s coordination mechanisms could address some of robotics’ most persistent bottlenecks: fragmented data, limited compute, and the high cost of deployment.

We did a recap of all his key takeaways from the speech below or you can watch the full keynote.

1. The coordination problem in robotics

Embodied AI progress depends on three core inputs: data from real-world environments, compute to train and evaluate models, and deployment to validate results in physical settings.

Today, those resources are largely concentrated in a few private labs. Independent researchers, startups, and universities often struggle to access them, which slows collective progress.

Michael outlined how crypto, not as a speculative token, but as a coordination layer, could help unlock those silos. By using onchain incentives and transparent verification, contributors around the world could pool resources and prove their participation. Collaboration in robotics could become measurable, sustainable, and scalable far beyond any single lab.

2. BitRobot’s mission: a decentralized robotics lab

BitRobot’s mission is to crowdsource "the world’s open robotics lab.”

Through a network of subnets, each focused on a specific research mission, contributors such as researchers, teleoperators, and hardware owners can coordinate resources and generate measurable outputs: datasets, models, and real-world evaluations.

This structure transforms robotics R&D into an open, distributed research network supported by shared infrastructure and aligned incentives.

3. Proof through practice

Several active projects already demonstrate this model in practice. FrodoBots, the development lab building on BitRobot, operates a global fleet of affordable sidewalk robots remotely controlled by hundreds of contributors. Together, they’ve generated more than 7,000 hours of navigation data, now used by academic teams—including UC Berkeley.

The EarthRover Challenge brought together students and researchers from National University in Seoul (NUS), Georgia Tech, George Mason University and Google DeepMind to pit navigation models against human drivers in real urban environments.

Frodobots also launched RoboCap, an open-source wearable for egocentric video data collection capable of over ten hours of continuous operation. BitRobot has also provided compute resources through io.net for world-model training and evaluation projects.

Each initiative reflects a central principle: by coordinating contributions through a shared network, BitRobot helps researchers focus on advancing embodied intelligence itself instead of building infrastructure from scratch.

4. Introducing the $5M BitRobot Grand Challenges Fund

Inspired by DARPA’s self-driving challenge, BitRobot announced a $5 million global prize fund to accelerate embodied AI research through public benchmarks.

Each challenge carries a $1 million prize for the first team whose robots convincingly outperform human experts in:

• Urban Navigation – Long-distance, real-world missions such as Berkeley to Stanford

• Origami Dexterity – Fine-motor manipulation in collaboration with the Nippon Origami Association

• IKEA Assembly – Multi-part assembly tasks testing planning and adaptability

These benchmarks are designed to measure real progress, not in simulation, but in the physical world. Check out the full announcement of the Grand Challenges here.

5. A collective future for embodied AI

Michael closed the keynote with an invitation to the robotics community: “Maybe there are thousands of you [roboticists] here, but there are millions of us out there willing to help.”

Crypto communities, he argued, shouldn’t be seen as speculators — but as potential collaborators, eager to contribute resources, time, and talent to accelerate progress in embodied AI. The next great breakthrough in robotics won’t come from a single lab.

A call to collaborate

Labs, researchers, and contributors are invited to get involved:

• Participate in or propose a Grand Challenge here.

• Launch a Subnet to coordinate contributors and receive network-level support for your research mission.

The next major leap in embodied AI won’t come from a single lab. It will come from the collective effort of many, connected through open infrastructure, shared incentives, and transparent coordination.