Our Modular Supercomputer is purpose-built to optimize the training of DNNs, LLMs and multimodal LLMs, and hybrid neural-symbolic architectures such as OpenCog Hyperon, tailored for scaling up dynamic AI workloads and AGI applications.

Some of the recent hardware purchases include advanced GPU compute systems, such as a Modular Data Center equipped with Nvidia L40S GPUs, AMD Instinct and Genoa, Tenstorrent Wormhole server racks, and servers featuring H200 GPUs, as well as Nvidia GB200 (Nvidia Blackwell) systems.

Updates about the project's developments will be shared with the community as we progress through our established milestones.