Akash, a cloud startup, is disrupting the AI industry by introducing its Supercloud for AI platform. This innovative solution allows businesses to access a network of GPUs remotely, overcoming the current shortage and democratizing AI. Unlike other cloud providers, Akash supports GPUs of any kind, providing a more accessible option for companies looking to leverage AI technologies.
The traditional barriers to running large language models (LLMs) or training generative AI systems, such as technical requirements and financial limitations, have hindered many organizations. Companies often need the best GPUs to train and manage these models effectively. Nvidia currently leads the industry with high-end GPUs like the A100, powering ChatGPT and other advanced AI workloads. However, there is a shortage of these GPUs due to rising demand.
Akash’s Supercloud for AI platform addresses this shortage by allowing businesses to tap into a network of GPUs remotely. The platform also enables others to contribute their GPUs, creating a decentralized approach to AI computing. The CEO and founder of Akash, Greg Osuri, believes that this could lead to a democratization of AI if the concept takes off.
Akash has built a technically complex decentralized cloud computing platform called AkashML. Their software acts as a dashboard, managing a heterogeneous network of computational resources, including GPUs. This approach has helped overcome challenges associated with aggregating lower-end GPUs for heavy AI workloads. Akash has prioritized high-powered GPUs like Nvidia’s A100 to ensure optimal performance.
The Supercloud for AI platform has been successfully tested, with over 1,300 participants and 3,100 deployments. Akash aims to offer developers more control over their applications by augmenting existing IT stacks, rather than forcing a complete infrastructure swap.
In addition to AI training, Akash is also exploring other applications on its network. It plans to support video encoding, visual asset rendering, and modeling, making it a general-purpose computational network.
Currently, Akash supports all Nvidia GPU models and is working towards adding support for AMD chips in the future. While onboard GPUs on smartphones and home computers are not currently optimized for AI/ML training, Akash is looking to explore this area in the long run.
With its decentralized GPU network, Akash is paving the way for a new era of AI cloud computing. By democratizing access to GPUs and offering more control to developers, Akash is revolutionizing the AI industry and making advanced AI capabilities more widely accessible.
The source of the article is from the blog coletivometranca.com.br