2CRSi launches the ATLAS 1.8GG, the densest server in the world designed exclusively for dual-phase immersion cooling

Strasbourg, October 10, 2024 – 2CRSi, a leader in the design and manufacturing of high-performance, energy-efficient servers, announces the launch of a major innovation: the ATLAS 1.8GG server, the world’s densest 1U GPU server, designed exclusively for dual-phase immersion cooling and capable of accommodating up to 8 high-performance GPUs such as NVIDIA H100, marking a significant technological advancement in the server industry.

Dual-phase immersion cooling is a cutting-edge technology that uses a dielectric liquid to dissipate heat from electronic components. This liquid, in direct contact with the electronics, absorbs heat and transitions to a gaseous state before condensing and returning to liquid form, enabling efficient and silent thermal management. This method drastically reduces the energy consumption dedicated to cooling, increasing overall system efficiency while significantly extending the lifespan of the components.

World First in Server Innovation

The launch of the ATLAS 1.8GG positions 2CRSi at the forefront of global innovation in density, being the first manufacturer in the world to integrate 8 NVIDIA H100 GPUs in a single U, while utilizing dual-phase immersion cooling technology. This unique ultra-dense design not only optimizes space utilization but also reduces energy consumption associated with cooling high-performance computing infrastructures, addressing today’s environmental challenges while maximizing performance. This server is particularly suited for HPC and Artificial Intelligence applications close to users, in sensitive sovereign sites, and especially in telecommunications infrastructures.

Availability

The first ATLAS 1.8GG servers have already been sold, with new series available starting in November 2024. A version equipped with 8 NVIDIA H200 GPUs, in 2U, is expected to be unveiled in early 2025. These 2CRSi server solutions dedicated to AI target data centers and companies looking to optimize their AI and high-performance computing infrastructures while adopting an eco-friendly approach.