GPUnet raises $5.25 million in series-A round

Decentralised cloud computation project GPU.NET has raised $5.25 million as part of their Series-A round with Momentum6, Spicy capital, Exnetwork, Blackdragon, Zephyrus Capital, Aza Ventures, F7 foundation, Halvings Capital, Bigger than Race ventures.

GPUnet aims to run affordable pre-training clusters for Enterprise clients to rent GPU compute, a very scarce resource in the current climate where everyone is experimenting with new use cases for Artificial Intelligence and Machine Learning algorithms. Globally, the 4 largest cloud providers – AWS, Google, Microsoft Azure and Oracle – own 80% of the global parallel computation resources leaving funded startups and researchers to either buy it from them at a heavy premium or get their own GPUs which needs skilled datacenter management plus long wait times for the chips to come from Nvidia.

GPUnet’s supply comes from independent datacenter operators who have been providing compute to Web3 projects for mining and have since gained the skills needed to run data centres to provide compute for accelerated cloud and HPC applications. These providers often have valuable computational resources available, but in small clusters. GPUNet’s team plans to use a distributed computing architecture to bring them on a single network, helping with providing a user friendly cloud environment for developers and researchers.

GPUnet has built a public blockchain where all providers, large or small can compete on a level playing field. While building an environment that is conducive for AI and Gaming applications, they are also working on cutting edge applications such as federated learning and Fully Homomorphic Encryption.

“I envision a future 2030 where a trillion dollar GDP will be created only out of energy & compute. This will be enabled by massive progress in the AI application space. If we see today the biggest AI companies are owned by those who also own the biggest data centers such as Microsoft and Meta, this will lead to massive monopoly or duopoly in global level LLMs that will be consumed. This is possible if we decentralise the ownership of the computational resources” said Suraj Chawla, Founder and CEO of GPU.NET.

Partner Resources

Popular Right Now

Edgecore Insight Podcast

Ep-1: Navigating the Waters of Sustainability

Others have also read ...