Skip to content
AI.info

companies

Gensyn

London protocol building decentralized infrastructure to train, verify, and pay for machine learning work across untrusted devices.

Gensyn

Gensyn is a London startup, founded in 2020 by CEO Ben Fielding and CTO Harry Grieve out of the Entrepreneur First program, building blockchain-coordinated infrastructure for distributed machine learning. Rather than renting GPUs from a single provider, Gensyn's protocol lets anyone contribute compute to training or reinforcement-learning jobs, with a custom OP Stack-based layer 2 handling identity, task assignment, payment, and dispute resolution between participants who don't know or trust each other. The company has raised roughly $50-70 million across seed, Series A, and later rounds, including a $43 million Series A in 2023 led by a16z crypto, and its public testnet went live in March 2025. Its most visible working product so far is RL Swarm, an open-source framework that lets multiple language models train together over the internet by critiquing and learning from each other's reasoning, which researchers can run on anything from a laptop to a rented GPU. The harder problem Gensyn is chasing, verifying that a stranger's GPU actually performed the training work it claims, cheaply enough to be worth doing, is still largely unproven at production scale. Most of what shipped in 2025 remains testnet and research-stage rather than something replacing a company's real training infrastructure.

Founded
2020
Headquarters
London, United Kingdom
Sector
infrastructure

Tools

  • Gensyn Testnet

    Gensyn Testnet is an open network for coordinating, verifying, and recording decentralized machine-learning activity.

  • RL Swarm

    RL Swarm is an open-source framework for distributed reinforcement learning, where models learn through peer-to-peer collaboration.

Official website