Microsoft Expands AI Infrastructure with Deals for Mistral and AMD Chips

Microsoft announced two major partnerships on the same day to bolster its artificial intelligence infrastructure, signaling a dual strategy to expand cloud computing capacity in Europe and diversify its chip supply. The deals involve a "multimillion-dollar" agreement with French AI leader Mistral and a deployment of AMD's new Helios system on Azure, both aimed at meeting surging demand for AI services.

Mistral Deal Boosts European Compute

The agreement with Mistral includes Microsoft integrating Mistral's models (Mistral Medium 3.5 and OCR 4) into its platforms like Microsoft Foundry and Copilot Studio. Microsoft will also use Mistral's European GPU infrastructure based on NVIDIA chips to expand training and inference capacity. The partnership emphasizes digital sovereignty, allowing customers to deploy models on Azure or locally, even offline, targeting regulated sectors. Source: infobae

AMD Helios to Power Azure AI

Separately, Microsoft will deploy AMD's Helios, a new AI infrastructure system, to enhance inference capabilities for Microsoft and Azure AI customers. Azure will also add new virtual machines based on AMD EPYC processors and expand use of Pensando DPUs. This move chips away at NVIDIA's dominance, which holds over 95% of the data center GPU market. Source: portafolioco

Key Points

• Microsoft's Mistral partnership includes joint go-to-market plans, pilot funding, Azure credits, and training programs. Source: infobae
• AMD's Helios shipments to Microsoft begin in the second half of 2026. Source: portafolioco
• Microsoft CEO Satya Nadella said clients need optimized infrastructure for diverse workloads. Source: portafolioco
• Brad Smith of Microsoft stated the goal is European access to AI without compromising data control. Source: infobae