Microsoft
Microsoft co-engineers GB300 NVL72 with Azure Boost/HSM and embeds Nemotron in SQL Server 2025, boosting Azure AI performance (18pp of Azure's 40% YoY growth). As a $240B company, this co-engineering is a small fraction of Microsoft's total business.
Co-engineering Azure infrastructure with Nvidia GB300 NVL72 including Azure Boost and HSM integration; jointly integrating Nvidia Nemotron into SQL Server 2025 and delivering NVIDIA NIM on Azure Marketplace
Notes
Named co-engineering: GB300 NVL72 with custom Azure Boost, NVIDIA Nemotron in SQL Server 2025, Microsoft 365 Copilot integrations, Rubin co-design across interconnects and rack scale. Both Nvidia blog and Microsoft Azure blog confirm. Distinct from Azure GPU cloud procurement.
Milestones
- Microsoft and Nvidia announce deep Azure co-engineering for GB-series AI infrastructure
Microsoft named as a co-engineering partner for Nvidia GB-family GPUs at Microsoft Ignite 2023, beginning a named collaboration on custom Azure infrastructure for leading-edge AI training and inference.
→ azure.microsoft.com - Microsoft deploys world's first GB300 NVL72 supercomputing cluster with custom Azure Boost and HSM
Microsoft launches the NDv6 GB300 VM series with co-engineered Azure Boost I/O acceleration and hardware security module integration — the first supercomputing-scale production cluster of GB300 NVL72 for OpenAI workloads.
→ azure.microsoft.com - NVIDIA Nemotron integrated into Microsoft SQL Server 2025 for enterprise RAG
Nvidia Nemotron open models and NVIDIA NIM microservices integrated into SQL Server 2025, delivering GPU-accelerated retrieval-augmented generation directly within enterprise database infrastructure on-premises and in Azure.
→ developer.nvidia.com