SingularityNET, an artificial intelligence platform developer, has revealed a $53 million investment in a modular supercomputer dedicated to decentralized artificial general intelligence (AGI).

According to a July 23 announcement, the investment’s first phase will allocate $26.5 million to the supercomputer, incorporating modular data center solutions from Ecoblox, graphics processing units (GPUs) and processors from companies such as Nvidia, AMD and Tenstorrent, as well as AI servers from Asus and Gigabyte.

A modular supercomputer is a high-performance computing system built with a flexible and scalable architecture that can be easily expanded and upgraded by adding or replacing modules. In other words, developers can add more power as needed without replacing an entire system. According to SingularityNET, the modular supercomputer will be the world’s first dedicated to decentralized AGI and artificial superintelligence research.

SingularityNET’s supercomputer will optimize the training of deep neural networks (DNNs), large language models (LLMs) — including multimodal variations — and hybrid neural-symbolic computing architectures, such as OpenCog Hyperon.

Some of the platform’s recent hardware purchases include a modular data center equipped with Nvidia L40S GPUs, AMD Instinct and Genoa, Tenstorrent Wormhole server racks, servers featuring H200 GPUs and Nvidia GB200 systems.

“The dramatic progress the AI field has seen recently is the result of convergence of multiple aspects, including sophisticated learning algorithms and cognitive architectures, and massive amounts of data, processing infrastructure and energy,” said SingularityNET CEO Ben Goertzel.

According to the announcement, the infrastructure will enable a major shift toward AGI continuous learning and self-improvement in high-load scenarios involving large-scale knowledge distillation, pattern matching and multi-step machine reasoning.

AI race heats up

SingularityNET’s investment follows a growing global dispute over AI development. On July 22, the Monetary Authority of Singapore (MAS) allocated 100 million Singapore dollars ($74.36 million) to develop quantum computing and AI-based solutions for its finance sector.

In March, Fetch.ai, a co-founder of the Artificial Superintelligence Alliance alongside SingularityNET and Ocean Protocol, announced a $100 million investment in its infrastructure program Fetch Compute. This investment will deploy Nvidia H200, H100 and A100 GPUs to create a more powerful platform for developers.

AI development has fueled a growing demand for high-performance hardware. On July 20, Taiwan Semiconductor Manufacturing Company became the first Asian firm to surpass a trillion-dollar market capitalization driven by the increasing demand from tech and manufacturing giants that rely on its semiconductors, including major clients like Apple, AMD, Intel, Nvidia and Qualcomm.

Magazine: ‘Sic AIs on each other’ to prevent AI apocalypse says David Brin, sci-fi author