Lumino Launches Decentralized ML Training Protocol

Lumino Announces Devnet Launch of Decentralized Machine Learning Training Protocol

Lumino, an AI infrastructure startup, announced that it has launched its development network (devnet) for machine learning training.  Training AI models requires large amounts of computing power, which has led to a price surge for GPUs. Additionally, large enterprises have reserved GPUs and have made it impossible for others to rent out GPUs. Finally, large public cloud providers are able to pick and choose who has the ability to build AI, potentially leading to censorship or concentrating power in the few. 

Our devnet launch enables Compute Providers (CPs) such as data centers and individual compute enterprises to monetize excess or latent hardware, by deploying these compute resources to the protocol and earning money by completing training jobs. CPs are able to maximize GPU utilization rate due to the serverless nature of the protocol, thus increasing returns on the capital invested.

We are able to drop the cost of compute by up to 80% through compute procurement, a higher utilization rate as compared to other providers, and improvements in machine learning training at the hardware, compiler, and model levels. By building a decentralized physical infrastructure network (DePIN), we ensure that large public cloud providers are not able to pick the winners and losers in AI. Every company needs to be able to build models to generate revenue, and now companies are able to get access to compute when they need and at a price that does not overrun their budgets.

Our mission is to make it easy for everyone to build and use AI in a safe, affordable, and quick manner! If that mission resonates with you, and are interested in joining the team, you can find the jobs page here.

To learn more about Lumino, visit https://www.luminolabs.ai  

‍

Twitter: https://twitter.com/luminoai 

LinkedIn: https://www.linkedin.com/company/luminoai/  

Warpcast: https://warpcast.com/~/channel/lumino  

‍