BitRobot

BitRobot is a distributed robotics research network built on Solana, founded in early 2025 with the stated mission of becoming the world's open robotics lab. The platform's core argument is simple: the biggest bottleneck in embodied AI — the branch of artificial intelligence that controls physical robots moving through real environments — is not compute or model architecture but data. Robots learn to navigate and manipulate the physical world through experience, and real-world experience remains scarce, expensive, and concentrated inside a small number of corporate and academic labs. BitRobot is designed to break that concentration by coordinating contributions from anyone with a robot, a research idea, or relevant compute.

The Subnet Architecture

BitRobot organizes work through modular clusters called subnets. Each subnet is a self-contained mission with a defined goal — collecting navigation data, training manipulation models, running autonomous benchmarks — and its own rules for what counts as valid work, how that work is measured, and how rewards are split among contributors. Subnets can be public, making them eligible for network-wide emission rewards, or private, operating under independent commercial arrangements.

Three roles structure every subnet. Subnet Owners define the mission parameters, establish measurement standards for what counts as Verifiable Robotic Work (VRW), and set the reward splits. Subnet Validators evaluate whether contributor outputs meet those standards. Subnet Contributors provide the actual resources: researchers, teleoperators, robot manufacturers, and hardware operators.

VRW is the network's unit of measurable robotic labor. Different subnet types produce different flavors of VRW — a navigation subnet might measure Qualified Miles driven, while a manipulation subnet might count successful task completions or hours of teleoperation video. The VRW framework allows the network to compare and reward heterogeneous contributions across subnets doing fundamentally different things.

Individual robots register on the network through Embodied Node Tokens (ENTs), NFT-based digital identities that represent specific physical machines. ENTs require collateral to become active on the network, creating an accountability mechanism: operators who violate subnet rules can lose their stake. ENTs serve as the conduit through which robot owners receive payments for their hardware's participation.

Governance: Foundation, Senate, and Gandalf AI

BitRobot distributes governance authority across three layers. The BitRobot Foundation handles infrastructure development, administrative operations, ecosystem grants, and the commercialization of network-generated assets — datasets and AI models created by public subnets are open-sourced for non-commercial use, while the Foundation licenses them commercially to support the broader ecosystem.

The BitRobot Senate is a nominated body drawing representatives from cryptocurrency and AI research communities. It meets during Voting Epochs to decide which subnets are admitted to the network's public emission pool and to assign weights that determine how emissions are distributed across active subnets based on their performance.

The third actor is Gandalf AI, an open-source AI agent model that generates its own alternative emission weight proposals. Network participants can delegate their voting power to either the Senate or to Gandalf AI, and the actual emission distribution is computed from the weighted sum of those delegated choices. This creates a structural counterbalance to Senate influence — a mechanism intended to prevent any single nominated group from fully controlling resource allocation as the network scales.

The economic loop the network calls Measurement-Evaluation-Rewards (MER) ties these layers together. Subnets generate measurements, the Senate and Gandalf AI assign evaluation weights, and rewards flow to contributors proportionally. Subnet Owners pay per-epoch registration fees to maintain active status; robot owners pay ENT registration fees; commercial licensing of the network's data and models feeds back into ecosystem funding.

Active Missions

As of mid-2025, BitRobot's website reports eight active missions and cumulative metrics of more than 4.5 million tasks completed and over 1,400 hours of teleoperation data collected. Representative subnets include:

ET Fugi (SN/01): A sidewalk rover mission focused on urban navigation. Participants remotely operate delivery-style robots to accumulate Qualified Miles in real city environments, generating navigation datasets for embodied AI training.

Earth Rover Grand Challenge: An autonomous navigation benchmarking subnet designed to evaluate AI models against standardized real-world courses.

TeleArms (SN/07): A teleoperation subnet in which participants remotely control robotic arms, collecting the manipulation video data that is notoriously difficult to generate at scale without physical lab infrastructure.

Robotic Origami and IKEA Assembly: Challenge-based subnets that test AI model performance on fine manipulation tasks with everyday objects.

The network lists $5 million in pledged funding for grand challenge prizes across these and future missions.

Research Partnerships and Team

The BitRobot whitepaper, published March 2025, lists Michael Cho, Jonathan Victor, and Juan Benet as authors. Benet is the founder of Protocol Labs and creator of IPFS and Filecoin, bringing substantial prior experience building distributed resource coordination networks. The platform cites published research partnerships with UC Berkeley, MIT, UCLA, and other institutions. The company has raised $6 million according to information on its website, though investor names and round details have not been publicly disclosed.

Roadmap and Solana Integration

BitRobot's near-term technical priorities center on establishing reliable VRW tracking and teleoperation networking infrastructure before layering more sophisticated cryptographic guarantees on top. The whitepaper describes a deliberate phasing strategy: initial versions rely on public inputs and outputs with minimal cryptographic requirements, with zero-knowledge proofs planned as the cost of ZK computation continues to fall.

Longer-term roadmap items include enabling AI agents — not just human teleoperators — to directly control physical resources through the network, and progressively shifting governance authority away from the Foundation and Senate toward network participants and Gandalf AI as the system matures.

BitRobot's Solana integration situates the project within the broader DePIN (Decentralized Physical Infrastructure Networks) ecosystem, where similar coordination models have been applied to wireless coverage, distributed mapping, and shared compute. Embodied AI represents a harder surface: the combination of real-time teleoperation requirements, physical hardware accountability enforced through collateral, research-grade data quality standards, and the fundamental sim-to-real gap in robot training make robotics coordination more technically demanding than most DePIN categories. BitRobot's subnet architecture and VRW measurement framework are its primary answers to those challenges.