IO is finally going online.
This type of computing power mining project mainly serves inference computing and edge computing scenarios, which do not have such high performance requirements.
Judging from the current computing power consumption of ChatGPT, about 70% is used to run large models, which can only rely on centralized super data centers. 30% of the computing power consumption is used to run this kind of inference and edge computing, which are the main markets for IO in the future.
I hope that computing power depin can become the next large-scale application scenario.
If this is the case, it is good for $SOL $APT
