Sesterce

Energy to Token, end-to-end AI Infrastructures.

[Anne Le Hénanff @ALehenanff]

Choose France ! 🚨

L’investissement de @SoftBank dans le projet de data center d’IA de 1 GW dans la Somme est une excellente nouvelle pour la France.

Bravo à @sestercegroup qui démontre qu’une entreprise française peut être au cœur des plus grands projets mondiaux d’infrastructures IA.

Nous construisons les capacités de calcul nécessaires pour faire de la France une grande nation de l’intelligence artificielle. 🇫🇷

[Sesterce @sestercegroup]

Sesterce is pleased to announce a strategic partnership with @SoftBank to develop a 1 GW AI Factory in Bosquel, France.

More Informations: https://t.co/YM4cITtdWI

[Sesterce @sestercegroup]

no gradient descent, only ascent

tomorrow @neon_noir_paris, Paris, France

by @PyTorch @GPU_MODE @PrimeIntellect @verdacloud @SemiAnalysis_ @sestercegroup

https://t.co/SJspUMGAnQ

[Neon Noir @neon_noir_paris]

We are excited to partner with @GPU_MODE and @PyTorch , alongside friends from @SemiAnalysis_ , @verdacloud , @PrimeIntellect and @sestercegroup , for a side-event of the PyTorch Paris Conference.

➡️Register here: https://t.co/3DXqhZi8vO

#AI #hackaton #LLM #conference

[PyTorch @PyTorch]

GPU MODE (@GPU_MODE) & PyTorch Foundation are organizing an ML systems hackathon in Paris on April 9, immediately following PyTorch Conference Europe 2026.

Researchers and engineers will compete across two tracks:

-Distributed training (LLM speedrun) and inference optimization (leaderboard)
- Access to a B300 cluster from Verda and H200 instances from Sesterce
- Cloud credits as prizes, including 48-hour access to a GB300 NVL72 rack
- Talks from PyTorch (Helion), vLLM, Prime Intellect, and more
- Food and refreshments

Doors open at 9:30, with the closing ceremony at 20:00. Attendees can join with a pre-formed team or match on-site. Location details are shared upon registration.

Spots are limited. Register: https://t.co/1BYgfGWJIr

#OpenSourceAI #PyTorchCon #PyTorch #vLLM #Helion

[Sesterce @sestercegroup]

Sesterce, Nvidia and GPU MODE are glad to host one of the main side events during GTC 2026.

The night will spotlight the winners of the Blackwell NVFP4 Kernel Hackathon, a global challenge where developers pushed GPU kernels to the edge on NVIDIA Blackwell hardware.

#GTC #AI

[Sesterce @sestercegroup]

Sesterce is touching down in San Jose for NVIDIA GTC 2026!

Special coverage plan: Our Investment, Sales, & Executive teams are heading out this year to connect with the innovators shaping the future of AI.

More info on special curated events later this week!

#GTC #Nvidia #AI

[Sesterce @sestercegroup]

LUMIÈRE : faire du calcul une infrastructure nationale, comme la France l’a fait pour l’énergie et les télécoms.

Le consortium LUMIÈRE — Sesterce, @Core_Scientific et @AVAIO — annonce sa volonté d’investir 20 Md€ pour construire 2 GW d’infrastructures de calcul IA en France, réparties sur deux sites (2 campus de 1 GW).

Hier et ce matin, nous avons présenté ce plan au Gouvernement et à l’Élysée.

LUMIÈRE repose sur un modèle clair : une gouvernance majoritairement française et européenne, associée à des leaders américains du secteur pour exécuter en France avec des standards industriels de tout premier rang.

LUMIÈRE : faire du calcul une infrastructure nationale, comme la France l’a fait pour l’énergie et les télécoms. C’est ce choix d’infrastructure — durable, mutualisé, structurant — qui a permis à la France de bâtir sa puissance industrielle. Et c’est par la même logique, appliquée au calcul, que la France éclairera encore le monde : en transformant l’innovation en capacité nationale, au service des usages critiques, de l’industrie et de la souveraineté.

Nous nous réjouissons du dialogue engagé avec les équipes de l’État, afin de faciliter la coordination nécessaire à la sécurisation des conditions d’exécution (énergie, raccordements, foncier, autorisations), dans le respect des procédures.

Nous remercions M. Emmanuel Moulin, Secrétaire général de la Présidence de la République, M. @RolandLescure , ministre de l’Économie, des Finances et de la Souveraineté industrielle, énergétique et numérique, Mme @ALehenanff, ministre déléguée chargée de l’Intelligence artificielle et du Numérique, ainsi que M. @matthieulandon1, conseiller du Président de la République (économie), et Mme @VVandeville, conseillère du Président de la République (attractivité & export), pour la qualité de leur écoute et l’opportunité de présenter ce plan pour faire rayonner la France sur l’IA.

----------------------------------------

LUMIÈRE: making compute a national infrastructure — just as France did for energy and telecommunications.

The LUMIÈRE consortium — Sesterce, Core Scientific and AVAIO Capital — has stated its intent to invest €20bn to deliver 2 GW of AI compute infrastructure in France, across two sites (two 1 GW campuses).

Yesterday and this morning, we presented the plan to the French Government and the Élysée.

LUMIÈRE is built on a clear model: French & European majority governance, combined with US industry leaders to execute in France at world-class industrial standards.

LUMIÈRE is about turning compute into national infrastructure — just as France did for energy and telecom networks. This is how France has historically transformed innovation into lasting national capability. Applying the same logic to compute is how France will keep lighting the way: by building durable, shared, strategic capacity for critical use cases, industry, and sovereignty.

We welcome the constructive dialogue with public teams to help coordinate the conditions required for execution (power, grid connections, land and permitting), with speed, clarity, and full respect of due process.

We thank Mr Emmanuel Moulin , Secretary General of the Presidency of the Republic, Mr @RolandLescure , Minister of the Economy, Finance and Industrial, Energy and Digital Sovereignty, Ms @ALehenanff, Minister Delegate for AI and Digital, as well as Mr @matthieulandon1, Economic Adviser to the President of the Republic, and Ms @VVandeville, Adviser to the President of the Republic (attractiveness & exports), for the opportunity to present the LUMIÈRE plan and strengthen France’s AI leadership.

@_Adam_Sullivan1, @russellcann, Ryan dewar,
@MMcC_gynonc, John Malone, Marie Jousset, @adrien_trl, Domitille Legrand, Julien Talagrand

#Lumiere #Sesterce #CoreScientific #AVAIOCapital #AICompute #HPC #CriticalInfrastructure #DigitalSovereignty #France #Europe #IndustrialStrategy

[Sesterce @sestercegroup]

From Energy Exporter to AI Powerhouse: Why France Holds the Cards. 🌍

In his latest op-ed for La Tribune, our chairman Anthony Tchakerian argues that France is at a strategic inflection point. With abundant, low-carbon, dispatchable nuclear power, the country faces a clear choice: keep exporting largely commoditized electricity—or convert that advantage at home into 21st-century value: compute, and the AI capabilities it enables.

At Sesterce, we believe the highest value is created not by moving electrons across borders, but by turning them into processing power. That’s why we are building the critical infrastructure—AI factories—that transforms competitive energy into intelligence, jobs, and technological sovereignty, here in France.

This requires a pragmatic partnership between the public sector and the tech industry: long-term visibility, stable and competitive energy pricing, and a clear commitment to high-performance computing as critical national infrastructure—on par with grids and telecom networks.

🔗 Read the full opinion piece: https://t.co/PK5jlcsAJ5

-------------------------------------
De l’export d’énergie à la puissance IA : pourquoi la France a les cartes en main. 🌍

Dans son nouvel article publié dans La Tribune, notre président Anthony Tchakerian défend une conviction simple : la France est à un moment charnière. Grâce à un parc nucléaire décarboné et pilotable, nous avons un choix stratégique : continuer à exporter une énergie peu valorisée, ou la transformer sur notre sol en valeur économique et industrielle — la puissance de calcul, au cœur de l’IA.

Chez Sesterce, nous partons d’un principe : la valeur ne se situe plus uniquement dans les électrons, mais dans ce qu’ils permettent de produire. C’est pourquoi nous développons des infrastructures de calcul — de véritables “usines à IA” — capables de convertir une énergie compétitive en capacité d’innovation, en emplois, et en souveraineté technologique.

Cette ambition suppose un alignement clair entre État, énergie et industrie : visibilité de long terme, prix stables et compétitifs, et reconnaissance du calcul haute performance comme une infrastructure critique, au même titre que les réseaux ou les télécoms.

🔗 Lire la tribune complète : https://t.co/PK5jlcsAJ5

#ArtificialIntelligence #EnergyInfrastructure #HPC #Sesterce #TechSovereignty

[La Tribune Dimanche @TribuneDimanche]

OPINION. « Colonie numérique ou Empire de l'IA ? Pourquoi la France joue sa survie sur le prix de l'électricité », par Anthony Tchakerian https://t.co/PPq7CqX4d3

[Sesterce @sestercegroup]

Productive couple of days at @SlushHQ 2025!

Our CEO, @yoemsri , was on the ground in Helsinki connecting with partners and founders, discussing the infrastructure and high-performance compute power required to scale the startups we saw this week. The demand for reliable cloud solutions is growing, and it was good to align directly with the people building the next wave of tech.

We're leaving Helsinki with reinforced commitments and exciting new partnerships that will help us continue to provide high-performance, resilient cloud solutions for the world's most ambitious startups.

See you next year, Helsinki! 👋

#Slush2025 #Sesterce #CloudComputing #HPC #Infrastructure

[Sesterce @sestercegroup]

From 6MW to a vision of 500MW by 2030 🚀

Sesterce's CEO, @yoemsri, shares updates on Sesterce’s international roadmap and our massive infrastructure projects.

Check out the full interview by BFM business: https://t.co/Mn2Sl5q7hP

#Sesterce #DataCenters #HPC #Infrastructure #TechNews

[Sesterce @sestercegroup]

The infrastructure requirements for AI are fundamentally diverging from traditional HPC.

At SC25, Sesterce’s Head of Data, AI/ML Ops, Kacem Lounissi, will join a panel of experts to discuss this critical shift.

Session: “AI Factory Supercomputers Are Not HPC Supercomputers”

The discussion will cover why established paradigms like file-based I/O and centralized metadata cannot meet the latency and concurrency demands of modern AI workloads.

Kacem will be speaking alongside @JeffDenworth, CoFounder at VAST Data, @glennklockwood, Principal Technical Strategist at VAST Data, and John Blaas, HPC Operations at Lambda.

This is a key session for anyone navigating the challenges of building and scaling AI infrastructure.

Event Details:
Date: Thursday, 20 November 2025
Time: 3:30pm – 5:00pm CST
Location: Rooms 240-241-242

If you are attending SC25, we invite you to join the session!

#Sesterce #SC25 #AI #MLOps #AIInfrastructure #HPC #Supercomputing

[Sesterce @sestercegroup]

Present today at the #ChooseFrance Summit at the Maison de la Chimie, opened by the Prime Minister @SebLecornu, we are proud to choose our country, France, to invest, create jobs and contribute to French excellence in sovereign cloud computing.

Our commitment is concrete:

Sesterce is investing €1.5 billion in our cloud campus project at Alixan, near Valence. We are excited to announce we are doubling the power of this site, from 40 MW to 80 MW.

This significant increase in capacity reflects the surge in computing needs for artificial intelligence and high-precision simulation.

This project will deliver one of the nation's most powerful supercomputers by December 2027, providing critical capacity for France's AI and high-performance computing needs.

We are operating on a tighter timetable to respond as quickly as possible to this demand. The first phase of commissioning will begin in December 2026, with full 80 MW operation scheduled for December 2027.

At 80 MW, the campus will target the training and inference of large models, industrial and scientific simulation, and data-intensive processing. It will provide multi-client access for businesses, laboratories, public actors, and digital service operators.

This is a powerful step forward for French technology, and we are honored to build it here.

🚀 Business leaders, entrepreneurs, French investors, let's choose France! 🇫🇷

#Sesterce #ChooseFrance #AI #Cloud #Supercomputer #Drôme #Alixan

[Sesterce @sestercegroup]

It was a privilege to be part of the conversation at AIM 2025 today!

Sesterce's CEO & Co-founder, @yoemsri , joined an inspiring panel of "Les nouvelles stars de l'IA française" (The new stars of French AI).

This was a fantastic opportunity to showcase the innovation happening right here in Marseille and to share Sesterce's vision for the future.

Thank you to @LaTribune for organizing a world-class event, and to the fellow panelists for the insightful discussion. Onward!

[Sesterce @sestercegroup]

Great to see La Provence's "Générations engagées" series feature our CEO, @yoemsri.

The article highlights his journey and Sesterce's core mission: building the supercomputing infrastructure needed to "give AI the power to create."

A fantastic read on the future of technology and innovation happening right here in Provence.

Check out the full article to learn more: https://t.co/KxX8KsVqxI

#Sesterce #AI #Supercomputing #HPC #Innovation

[alex zhang @a1zhang]

The wait is over! We’re so excited to announce the @GPU_MODE x @NVIDIA kernel optimization competition for NVFP4 kernels on Blackwell B200s!

We will be awarding NVIDIA DGX Spark’s & RTX 50XX series GPUs for individual rankings on each problem, as well as a Dell Pro Max with NVIDIA GB300 (an absolutely insane prize btw) for the grand prize winner of the entire competition! Winners will also be invited to GTC 2026 for free.

The competition will feature 4 different single-device kernels over a 3 month period; we highly encourage competitors to try out and use CuTe DSL / CUTLASS 4.0 on these problems, and we provide reference submissions in CuTe DSL. We thank @sestercegroup, an HPC cloud platform, for providing DGX B200s for participants to use for free during the competition.

Registration through Luma is linked below.

[Sesterce @sestercegroup]

What an amazing few days we had in San Francisco!

Sesterce team was right in the thick of things at the PyTorch Summit (Oct 22-23), and let me tell you, the energy was absolutely buzzing. The big takeaway? The future of AI isn’t just about creating larger models; it’s about making them faster, more efficient, and ready to deploy everywhere.

Here are our three key takeaways from the summit:

• Performance is the New Frontier: The spotlight has shifted from merely achieving model accuracy to optimizing performance. We delved deep into torch.compile and explored new techniques to significantly reduce inference latency and training costs—something we’re really passionate about in FinOps for AI.

• GenAI in Production is Maturing: The discussions are evolving from "what if" scenarios to "how to" implement them. Sessions on scalable deployment, model sharding, and managing complex AI pipelines clearly show that the industry is committed to taking these powerful models from the lab to real-world applications.

• The Rise of On-Device & Edge: There’s a strong push for more efficient models that can operate on the edge. This is a game-changer for data privacy and real-time applications, but it also brings a new challenge: managing a distributed, hybrid AI infrastructure.

It’s evident that the next wave of AI innovation will hinge on the cloud infrastructure that supports it. Building, training, and deploying these models at scale demands a robust, high-performance, and cost-effective cloud strategy.

It was fantastic to connect with so many developers and leaders who are shaping the future of AI. We’re heading back with a wealth of new ideas!

#PyTorchCon #AI #PyTorch #SesterceGroup

[Sesterce @sestercegroup]

Training gets the glory.

But inference runs the show.
Every chat, every recommendation, every AI task = inference.
Live. Continuous. Revenue-critical.

Still scaling like it's 2021?
You’ll hit latency walls fast.

Inference is your factory.
Architect like it.

#AIInfra #InferenceAtScale #LLM #MLOps #LatencyMatters #SesterceThinking

[Sesterce @sestercegroup]

Infra should be fast. Not buried in tabs.

Sesterce Cloud is 100% CLI-native.

No dashboards. No dropdowns.
Just:
✅ GPU VMs & Bare Metal
✅ Region & type flags
✅ Script & CI/CD ready

🖥️ sesterce create --region=par4 --gpu=H200 --count=8

➡️ https://t.co/IiloW7H6f8

[Sesterce @sestercegroup]

Deployed in France. Live in seconds.

No fine print, just sovereign GPU clusters:
✅ H200, H100, L40, L4
✅ Regions: Paris & Marseille
✅ Ubuntu-based envs
✅ No-commit pricing

Born in 🇫🇷. Built for global AI.
➡️ https://t.co/8eChbqYo40

#GPUCloud #SovereignAI #HPC #PrivateAI