Gensyn operates a blockchain-based protocol that matches buyers and sellers of compute power and provides cryptographic verification to prove machine learning work was completed correctly. The platform is designed to harness large-scale, distributed compute networks in a way analogous to projects like SETI@home or Folding@home. The company argues that expensive access to cloud providers (AWS, Azure) is a bottleneck for AI development and aims to cut out intermediaries to sharply increase supply. By decentralising access to devices for training models, Gensyn expects a deflationary effect on the price of compute, enabling larger models and broader participation in model training. The team set out on this journey three years ago and is positioning the product as a means to expand computational access and “computational liberty.” Financially, Gensyn recently closed a $43 million Series A round, providing capital to further develop and expand the protocol. Gensyn is developing a decentralized compute network that applies a blockchain and token to verify and pay for deep-learning compute tasks performed on a distributed set of machines. The platform is designed to monetize unused compute capacity and to provide a lower-cost alternative to centralized cloud GPU/TPU providers. Gensyn plans to launch its network for training AI models, using on-chain verification to trigger payments via a token. The company says this approach offers environmental benefits by using otherwise idle compute and claims material cost savings versus traditional cloud services. Gensyn estimates its average price per standard NVIDIA GPU-hour can be up to 80% cheaper (about $0.45/hour versus roughly $2/hour on other cloud services). On the financial side, Gensyn has raised a $6.5M seed round and previously raised an unannounced $1.1M pre-seed in 2021.

Total raised

$50M

Funding rounds

2

Latest round

Series A

Latest activity

Jun 2023