This content originally appeared in The Drop newsletter.
Decentralized Cloud Partnerships: io.net and Walrus Launch AI Compute Solution
Decentralized cloud computing provider io.net is collaborating with the Walrus team to offer startups access to custom AI model development. The integrated service allows developers to train, run, and store custom AI models within the decentralized ecosystem.
IO provides a network of GPUs for AI training and fine-tuning, while Walrus offers decentralized model storage. The solution operates on a pay-per-use basis, charging users solely for the computing power and storage consumed.
This “Bring Your Own Model” (BYOM) approach enables AI developers and application builders to create and manage AI models without necessitating their own data centers or specialized hardware.
According to io.net, its network currently comprises over 10,000 GPUs and CPUs deployed globally. Walrus and io.net anticipate competition from established players like Bittensor, Lambda, Spheron, Akash, Gensyn, Vast AI, and Google’s Vertex AI services.
Rebecca Simmonds, Managing Executive at Walrus Foundation, stated, “Traditional centralized cloud models are not only expensive — they come with significant privacy risks and limited composability options that are challenging for developers who prioritize decentralization. By leveraging our decentralized data storage solution, io.net will be able to provide necessary compute power for AI and ML development without traditional drawbacks.”
Decentralization Imperative
Concerns about the centralization of major cloud providers, as evidenced by recent outages impacting services like Google Cloud and Cloudflare, underscore the appeal of decentralized alternatives for sensitive applications like AI development.
Market Context
Walrus, whose Sui-based mainnet launched in March following a $140 million funding round, focuses on programmable decentralized storage. The partnership could accelerate adoption of its storage protocol among AI developers.
The escalating demand for AI compute power is substantial. McKinsey analysts predict global data center investment will reach $6.7 trillion between now and 2030 to meet overall demand, with AI compute specifically requiring $5.2 trillion in investments over the next five years, significantly outstripping investment needed for non-AI IT applications ($1.5 trillion).
The joint offering from Walrus and io.net is positioned to potentially gain traction among AI professionals within and beyond the cryptocurrency sphere.
