IO's computing power pool

CN
1 year ago

The computing power pool of IO is invincible in the DePin section. Among the 805,000 GPUs in the computing power pool, there are over 50,000 A100s worth $10,000 to $20,000, and over 3,600 H100s worth $29,000 to $36,000, which truly puts them in a leading position.

In addition to the AI computing power cloud services provided by the computing power pool, IO's basic stack also has a very important Exclusion/ML Task (data cleaning preprocessing and machine learning) layer.

This layer supports the deployment of declarative deep learning open source framework Ludwig (GitHub 10.8 stars), large language model service solution RAY LLM, distributed large model training tool Pytorch FSDP, AI model inference terminal HuggingFace, and so on.

With this kind of architecture, IO is a bit like a decentralized version of Google's Colab, providing AI developers with a ready-to-use machine learning tool (computing power + model training inference SDK).

IO recently reached a cooperation with AI coprocessor Ritual, providing computing power support for AI Dapp/De Agent through Ritual's AI oracle infernet.

Overall, despite the fact that many people in the market deny and doubt the combination of AI and Crypto (including myself in 23 years), seeing the current computing power scale of IO and its future SDK support capabilities, as well as the maturity of AI+Crypto stacks such as Ritual, MyShell, Sahara, etc., I believe that there will be a small outbreak of landing applications in the AI+Crypto field in the next 6 to 12 months.

免责声明:本文章仅代表作者个人观点,不代表本平台的立场和观点。本文章仅供信息分享,不构成对任何人的任何投资建议。用户与作者之间的任何争议,与本平台无关。如网页中刊载的文章或图片涉及侵权,请提供相关的权利证明和身份证明发送邮件到support@aicoin.com,本平台相关工作人员将会进行核查。

Share To
APP

X

Telegram

Facebook

Reddit

CopyLink