Sesterce launches high-performance green SuperComputer for AI training and inference at Digital Realty facility

Sesterce launches high-performance green SuperComputer for AI training and inference at Digital Realty facility

Sesterce has launched a new SuperComputer featuring NVIDIA H100 Tensor Core GPUs hosted in a Digital Realty data centre located in Marseille.

The SuperComputer helps to tackle the growing demand for powerful cloud GPU solutions and offers a decarbonised AI-as-a-Service to Sesterce’s customers.


Located in Digital Realty’s Marseille data centre, the cluster is equipped with NVIDIA H100 Tensor Core GPUs interconnected through InfiniBand technology. These cards provide excellent scalability and performance, contributing to optimal energy efficiency for Artificial Intelligence. 

With this SuperComputer, Sesterce empowers clients to develop highly scalable AI projects with optimal performance and sustainability, providing unmatched flexibility for AI projects of any size. 

The use of AI has been democratised, resulting in demand for storage capacity and computing power continuing to rise. In 2023, France had over 600 startups dedicated to AI in the country, a 24% increase since 2021. This trend is expected to continue in the coming years. However, this exceptional growth parallels a significant rise in energy consumption, raising environmental concerns. 

The cluster will benefit directly from Digital Realty’s on-going programme dedicated to minimising the environmental impact of its data centres. These actions include enhancing energy efficiency and procuring renewable energy with a strong focus on locally sourced renewables, ensuring that its entire European portfolio continues to be matched with 100% renewable energy.

Additionally, Digital Realty is committed to achieving carbon neutrality across its European data centre portfolio by 2030, alongside a 68% reduction in Scope 1 and 2 emissions globally (against a 2018 baseline) and a 24% reduction in Scope 3 emissions globally within the same timeframe.

Browse our latest issue

Intelligent Data Centres

View Magazine Archive