
In development
Bosquel
Somme, France
A 1 GW campus, developed with SoftBank Group.

An elite team building the best supercomputers in the world, in campuses designed with the towns that host them.
Bosquel, Hauts-de-France1 GW of AI compute, developed with SoftBank Group.
Campuses
Every Sesterce Campus follows the same six principles, whatever the country.

In development
Somme, France
A 1 GW campus, developed with SoftBank Group.

In development
Drôme, France
A Sesterce Campus in the Drôme.

In development
Central Finland
A Sesterce Campus in the Finnish Lakeland.
News
Company ·
A planned investment of over €10 billion would bring AI infrastructure to the former Kaipola paper mill, starting with a 200 MW first phase.

Company ·
A confirmed 40 MW cap, six measurable commitments and a proposed dialogue committee with the local community.

Partnerships ·
Joint venture aims to advance France’s AI leadership, create jobs, and support AI adoption across Hauts-de-France.

The highest bar ever set for a supercomputer. The same in every country.

Architecture
A campus you are proud of, not one you try to hide. We give it the care usually reserved for a public building.

Living
Every campus pairs the supercomputer with a living building: hotel rooms, restaurants, training and education, open to residents.

Energy
For every megawatt a campus consumes, we build one megawatt of new renewable generation on the same grid, within ten years. New plants, not certificates.

Water
Direct liquid cooling in a closed loop. Every liter that tops up the loop comes from on-site rainwater and used water we filter and treat.

Jobs
Three local jobs per MW while the campus is being built, and at least one long-term job for every 2 MW consumed.

Community
Paid at commissioning. The municipality spends it locally, on what improves life for the people who live there.
“We want the places that host our supercomputers to be better off for it. That is why every campus is built on these six principles.”
Customers
1:23, watch the interview (1 minute 23)“Sesterce builds infrastructure ready for long-term AI evolution.”
Damien PatriarcheArchitect, President of Patriarche
1:48, watch the interview (1 minute 48)“Sesterce gives us European capacity for critical AI workloads.”
Sébastien UriosEMEA Sales Director, Flex AI
2:55, watch the interview (2 minutes 55)“Sesterce connects the infrastructure layers enterprise AI needs.”
Dan ChesterEMEA Sales Director, VAST Data
Supercomputers
Dedicated bare-metal clusters, built with Sesterce Modular Design and run by Sesterce OS.

Sesterce Modular Design
Power distribution, backup power and cooling are standard modules. We assemble them in the factory, ship them to the campus and connect them to the building. We own every hand-off, from watt to token.
Module by module. New capacity is added as demand grows.

Bare metal, run by Sesterce OS
Dedicated hardware, network and storage, with root access and out-of-band management on every node. Sesterce OS runs the campus around it, from the grid to the chip. It watches every hall, the power chain, the cooling loop and every node, and recommends what to do next.
Campus
Operations
Row A2 · CDU 1 · busways A2 and B2
Campus
Operations
Campus
Operations
| CDU | Supply | Return | L/min | Filter |
|---|---|---|---|---|
| 1 | 30.0 | 39.6 | 1,240 | 21 kPa |
| 2 | 30.0 | 39.9 | 1,236 | 24 kPa |
| 3 | 30.1 | 40.2 | 1,251 | 19 kPa |
| 4 | 30.0 | 39.7 | 1,188 | 46 kPa |
| 5 | 29.9 | 39.8 | 1,262 | 22 kPa |
| 6 | 30.0 | 39.5 | 1,243 | 20 kPa |
All dry. LS-A14 tripped for 40 s at 13:58, then cleared. A technician checks the manifold at 15:00.
Campus
Operations
| Node | Health | Hottest GPU |
|---|---|---|
| b4-r07-n01 | Healthy | 68 °C |
| b4-r07-n02 | Healthy | 70 °C |
| b4-r07-n03GPU reset at 10:17 | Recovered | 71 °C |
| b4-r07-n04ECC errors, GPU 5 | Warning | 79 °C |
| b4-r08-n01 | Healthy | 69 °C |
| b4-r08-n02PSU 2 failed | Critical | Off |
| b4-r08-n03NVLink errors, GPU 6 | Warning | 72 °C |
| b4-r08-n04 | Healthy | 71 °C |
| b4-r09-n01 | Healthy | 67 °C |
| b4-r09-n02 | Healthy | 69 °C |
| b4-r09-n03 | Healthy | 70 °C |
Showing 11 of 64 nodes · sorted by rack
BMC reachable · Redfish · 8 GPUs
GPU temperatures
Event log
Illustrative interface. The values shown are examples.

Construction, grid, commissioning and permits

Modular design, cooling, equipment and energy

Facilities, operations center and on-site teams

Clusters, networks, hardware and software
Contact
Customers, partners, towns and landowners, press and candidates: tell us who you are and what you have in mind. Your message goes straight to the right team.
ComputeDedicated bare-metal clusters, from a few racks to a full campus.
Host a Sesterce CampusFor towns, regions and landowners.
Partners and investorsEnergy, construction, technology and finance.
PressInterviews, images and the media kit.
CareersOur open positions are listed above.