Back to all
    Back to all
    Back to all
      Back to all
        [ABOUT US]

        Australia’s sovereign, sustainable, and secure AI cloud has been 5 years in the making

        We’ve reinvented the way the cloud is built, using new technology to host powerful compute in a more sustainable, powerful, and cost-effective way.

        [OUR STORY]

        Building the world’s most sustainable and powerful cloud

        • The R&D journey

          Searching for a paradigm shift in efficiency, Firmus looked for a solution. Immersion cooling was it, but no solutions were efficient enough or scaled to industry standards for reliability and compatibility so we decided to build our own. 5 years of R&D, testing, and commercialisation followed, including partnering with the infrastructure and accelerated compute industry.

          During that process, Firmus devised a hyperscale test load – 20MW – at its continuously operating, renewable powered R&D facility in Tasmania. 

          Out the other end came the first truly scaled immersion system – infrastructure to compute – built to industry standards for reliability and performance. Able to go anywhere; in DCs or stand-alone.

        • Power in the cloud

          We believe in the power of technology to change the world. With infrastructure able to host more powerful compute than ever before, we press that advantage – building the AI factory of the future. 

          AI factories are needed – right now – to solve humanity’s greatest challenges. Based on reference supercomputing architecture from NVIDIA, Supercloud’s infrastructure is set from the ground up to change the power in the cloud. 

          This is why, from day one, we sought to include GPU infrastructure and the highest-performing networked environment in our offering.

        • Supercloud today

          Our parent company Firmus mastered the infrastructure and we’ve championed its use.

          As an AI GPU cloud, our software and control stack features deployable tools for AI, ML, and visual computing applications, as well as flexible APIs in a familiar UX.

          Supercloud left beta in early 2022, initially based in our Australia (South 1) cloud region. We’ve just announced our second region in Australia (Canberra 1), due to be online in late 2022.

          Hardware availability is growing now, with additional A100 and A40 capacity coming online during 2H22. Working with our partner NVIDIA, Supercloud will host early POCs for Hopper H100 and Grace ARM CPUs as they become available.

          Although Supercloud has grown, our core principles have not changed. We still deliver compute with some of the lowest carbon impact in the world. We still provide full service, lower cost infrastructure, and we’re still sovereign and secure, and always will be.

        [OUR PARTNERS]
        • firmus
        • nvidia
        • canonical
        • supermicro