A company named Akash is changing how we use our computing power — They’re challenging the big cloud providers and introducing a new way to access computing resources, through a decentralized marketplace where anyone with powerful hardware can share their computing power and get paid with tokens (as of July, one token ($AKT) was worth $0.56).
Their Supercloud for AI platform features high-end GPUs like Nvidia H100 and Nvidia A100, which are essential for running advanced AI models and they also support top AMD GPUs and consumer-grade GPUs; Known as Mainnet 6, this Supercould for AI is creating an open-source marketplace specifically for high-performance GPUs.
Existing network providers (including both small operators and big players like hyperscalers), can now offer their GPUs alongside their existing services.
Akash’s console page.
A Growing Network
As this network grows, it becomes more robust, hosting a larger number of GPUs from providers worldwide. It’s becoming a crucial part of the infrastructure for AI development. Akash believes that having an open-source network is vital to prevent a handful of big corporations from controlling these resources.
Akash points out that a big issue with the current cloud market is that it’s dominated by just four providers — Amazon, Microsoft, Google, and Alibaba — and that these top cloud providers charge way too much, not to mention that, according to the company, many of their data centers are underused.
The most powerful GPUs often end up with hyperscalers and big corporations (solidifying their grip on the public cloud market) so, with Akash’s Supercloud for AI GPU network, the future of AI model training could see some significant changes. Let’s wait and see what will happen in the following months.
Filed in . Read more about AI (Artificial Intelligence).