On the Mistral AI NOW Summit, VAST Information outlined its partnership with Mistral AI and Mistral Compute, detailing a joint effort to deploy NVIDIA-accelerated AI factories in Europe. The collaboration aligns three core layers required for manufacturing AI: NVIDIA accelerated compute, Mistral’s frontier fashions, and the VAST AI Working System because the unified knowledge platform.
On the middle of the deployment is a large-scale implementation of NVIDIA GB300 NVL72 techniques, representing one of many highest-density concentrations of this structure in Europe. Mistral Compute will function the infrastructure as a part of its AI cloud platform, whereas VAST offers the info layer that manages knowledge entry, motion, governance, and efficiency throughout the surroundings.
Information Platform for Manufacturing AI Workloads
VAST positions its AI Working System because the foundational knowledge layer for AI factories, designed to assist end-to-end workflows with out the fragmentation usually seen in multi-system pipelines. Inside the Mistral Compute surroundings, the platform offers a shared knowledge structure spanning coaching, inference, retrieval, and enterprise deployment. This eliminates the necessity for duplicated datasets and reduces operational overhead related to siloed storage and knowledge motion.
The platform is constructed to deal with the complete vary of knowledge sorts and runtime states required by trendy AI workloads. This consists of structured and unstructured knowledge similar to information and tables, in addition to vectors, key-value cache, occasion streams, and protracted agent reminiscence. These capabilities are delivered throughout distributed environments, enabling constant efficiency and entry no matter location.
Extending Throughout Analysis, Cloud, and Enterprise
The partnership extends past a single deployment and displays a broader integration throughout Mistral’s AI lifecycle. VAST is already deployed in AI cloud environments utilized by Mistral AI for mannequin coaching and analysis, supporting the event and operation of its personal fashions, together with Voxtral, Ministral, and Codestral. Via VAST DataSpace, Mistral groups can function throughout a number of cloud environments utilizing a unified namespace, avoiding the necessity to redesign knowledge pipelines when shifting workloads.
Mistral Compute, now an NVIDIA Cloud Accomplice, has adopted the VAST AI Working System as its core knowledge platform for its managed AI cloud companies. The system is at the moment in manufacturing, supporting each inside Mistral workloads and customer-facing deployments. The addition of GB300 NVL72 infrastructure extends this surroundings to assist higher-scale AI manufacturing facility operations.
Supporting Enterprise AI and Information Management
As Mistral brings its fashions into enterprise environments, the requirement shifts towards integrating fashions with enterprise knowledge whereas sustaining efficiency and governance. VAST offers a typical working layer that connects fashions to enterprise datasets, enabling constant knowledge entry and management throughout coaching and inference workflows.
The structure is designed to assist necessities round knowledge locality, governance, and isolation, that are more and more vital for European enterprises and public-sector organizations. By sustaining a unified knowledge basis throughout analysis, cloud, and enterprise deployments, the platform permits organizations to retain management over how knowledge is saved, accessed, and utilized in AI workflows.
Positioning for European AI Sovereignty
The deployment displays a broader push towards regional AI infrastructure and knowledge sovereignty. With compute, fashions, and knowledge platform built-in inside a European-operated surroundings, the answer offers a framework for constructing and working superior AI techniques with localized management.
On this mannequin, the info layer turns into the first management level for AI operations, governing how knowledge flows between coaching, inference, and manufacturing techniques. VAST’s function because the unified knowledge platform positions it as a key element in enabling Mistral Compute to ship each high-performance AI infrastructure and the governance capabilities required for regulated and enterprise use instances.
