Akash Network Launches 'Homenode' for Consumer GPUs
What happened
Akash has launched early access for its Homenode, a platform that allows everyday devices to contribute to a decentralized compute network. The first phase enables users with high-end consumer GPUs, like the RTX 4090 or 5090, to provide compute power for AI and developer workloads.
Why it matters
- The company behind Akash Network is Overclock Labs, founded by CEO Greg Osuri and CTO Adam Bozanich in 2015, with the network's mainnet officially launching in September 2020. - Akash is a prominent project in the Decentralized Physical Infrastructure Network (DePIN) sector, which uses blockchain to build open alternatives to traditional infrastructure. In February 2025, it was reported as the highest-earning DePIN project, with $4.2 million in annual recurring revenue. - The network operates as a peer-to-peer marketplace where users needing compute power are matched with providers who have spare capacity; providers are then paid in the native AKT token. - Founder Greg Osuri has stated the project plans to move away from its own Cosmos-based blockchain to a new network that offers stronger security and deeper liquidity, calling Solana a "strong contender". - Homenode is part of a broader initiative called StarCluster, which aims to build a globally distributed AI compute network by making it simple for non-technical users to contribute hardware. - The decentralized compute space is growing, with networks positioning themselves as a cost-effective layer for AI inference and data processing workloads, which are less demanding than training large-scale AI models.
Key numbers
- The first phase enables users with high-end consumer GPUs, like the RTX 4090 or 5090, to provide compute power for AI and developer workloads.
- - The company behind Akash Network is Overclock Labs, founded by CEO Greg Osuri and CTO Adam Bozanich in 2015, with the network's mainnet officially launching in September 2020.
- In February 2025, it was reported as the highest-earning DePIN project, with $4.2 million in annual recurring revenue.
What happens next
- Founder Greg Osuri has stated the project plans to move away from its own Cosmos-based blockchain to a new network that offers stronger security and deeper liquidity, calling Solana a "strong contender".
- Homenode is part of a broader initiative called StarCluster, which aims to build a globally distributed AI compute network by making it simple for non-technical users to contribute hardware.
Quick answers
What happened in Akash Network Launches 'Homenode' for Consumer GPUs?
Akash has launched early access for its Homenode, a platform that allows everyday devices to contribute to a decentralized compute network. The first phase enables users with high-end consumer GPUs, like the RTX 4090 or 5090, to provide compute power for AI and developer workloads.
Why does Akash Network Launches 'Homenode' for Consumer GPUs matter?
The company behind Akash Network is Overclock Labs, founded by CEO Greg Osuri and CTO Adam Bozanich in 2015, with the network's mainnet officially launching in September 2020. Akash is a prominent project in the Decentralized Physical Infrastructure Network (DePIN) sector, which uses blockchain to build open alternatives to traditional infrastructure. In February 2025, it was reported as the highest-earning DePIN project, with $4.2 million in annual recurring revenue. The network operates as a peer-to-peer marketplace where users needing compute power are matched with providers who have spare capacity; providers are then paid in the native AKT token. Founder Greg Osuri has stated the project plans to move away from its own Cosmos-based blockchain to a new network that offers stronger security and deeper liquidity, calling Solana a "strong contender". Homenode is part of a broader initiative called StarCluster, which aims to build a globally distributed AI compute network by making it simple for non-technical users to contribute hardware. The decentralized compute space is growing, with networks positioning themselves as a cost-effective layer for AI inference and data processing workloads, which are less demanding than training large-scale AI models.
