This is a section from the drop newsletter. To read full versions, subscribe.
Decentralized Cloud Computing Provider IO.NET, who has its own IO tokens, is working with the walras team, to train startups to train, run and store your own custom AI model.
The IO AI provides a network of GPU for training and fine-tuning, while the AI model in the walraus deal is enabling storage. Integration will be available in the form of a pay-per-use offer, which means that the builders are charged only for the amount of storage and computing power they use.
Bring it your own model (Bime) Offering AI agent Dev or AI app allows builders to develop and operate the AI model without the need to install their own data center or hardware.
IO says this is More than 10,000 GPU And globally CPU.
The offering of walus and Io’s byom will have to compete with other AI developers cloud services, however, such as Bittensor, Lambda, Spheron, Akash, Gensyn, Vast AI and Google’s vertex products.
“Traditional centralized cloud models are not only expensive – they come up with significant privacy risks and limited composibility options that are challenging for developers who prefer decentralization,” Rebecca Simmonds said, executive in the Walras Foundation.
“By taking advantage of our decentralized data storage solutions, IO.NET will be able to provide the calculation power required for advanced AI and ML development without any shortcomings of traditional models, making it a clear victory for developers, users and the entire web 3 industry,” EXEC continued.
When the Internet can feel very centralized about the experience of giant cloud services Happened for google Last week (and affected Cloudflare Also other apps and a series of sites). This is a clear reason that centralized AI calculation may not be ideal.
Walrus’ mennet was launched in March, with its main pitch programable, decentralized storage. Walrus Foundation announced a growth $ 140 million That month.
More broadly, the requirement of computing power for AI is expected to increase annually. Macinse researcher predict That data centers require $ 6.7 trillion globally to stay with demand by 2030 (although there is a series of potential scenarios given below).
The AI Compute requires $ 5.2 trillion in investment to maintain the current demand over the next five years, while non-AI IT apps require much less than $ 1.5 trillion.
Walus’s AI storage offering with Io can help adopt more than AI builders and app gods, even outside the crypto space.
Walrus Mennet currently has 4,167 TB in total storage capacity, about 26% of it is currently being used. Protocol is 103 total Operator in 121 storage nodes.
Get news in your inbox. Explore blockwork newsletters:
