Akash Network Launches 'Homenode' for Decentralized GPU Access
Akash has launched the first phase of its Homenode early access program. The initiative allows individuals to contribute compute power from laptops, desktops, and edge devices equipped with high-end GPUs like the RTX 4090. The goal is to create a decentralized network to expand access to compute resources for AI development.
- The Homenode initiative is part of a broader "Akash at Home" plan intended to simplify the process for individuals to become compute providers, as many potential participants lack the technical expertise to run a full Kubernetes cluster. The goal is to allow users to install a client on their devices, which will then handle the necessary configurations to join the provider network. - Akash Network's broader strategy includes integrating more powerful, next-generation GPUs, such as NVIDIA's Blackwell-based B200 and B300, to support more intensive on-chain AI training and inference workloads. This complements its existing support for a range of GPUs including the H100, A100, and various RTX models. - Decentralized GPU networks like Akash are primarily positioned to handle AI inference and other workloads that can be easily partitioned, as large-scale AI model training requires tightly synchronized, low-latency hardware found in centralized data centers. The distributed nature of these networks can, however, reduce latency for end-users by processing tasks closer to their geographic location. - The project was co-founded by CEO Greg Osuri and CTO Adam Bozanich of Overclock Labs in 2018, with the mainnet launching in 2020. Osuri's background includes founding AngelHack, the world's largest hackathon organization, and designing cloud architecture for Kaiser Permanente. - The network utilizes the Cosmos SDK and a delegated Proof-of-Stake (DPoS) consensus mechanism, with its native AKT token used for security, governance, and transactions within the marketplace. There are plans to migrate from its current Cosmos-based chain to a new Layer 1 blockchain to improve security and capital efficiency. - Akash faces competition from other decentralized compute projects such as Render, Flux, and Aethir, each targeting different customer segments from 3D artists to cloud gaming. A key challenge for Akash has been a supply shortage of high-demand GPUs like the A100, with utilization rates previously reaching 91%.
