MSFT Dev Blogs

Energy at Microsoft Ignite was electric. Over 20,000 attendees gathered in San Francisco and 200,000 joined digitally to explore the future of cloud and AI. Online reactions and technical community conversations show how quickly announcements become action—building solutions for billions.

and Azure Integrated HSM. Announcements are starting points—register for Microsoft AI Dev Days (Dec 10–11, 2025), attend AI Tour workshops,

and join the technical community to build what comes next. Find more information at https://azure.microsoft.com/en-us/blog/actioning-agentic-ai-5-ways-to-build-with-news-from-microsoft-ignite-2025/.

and troubleshooting. The migration agent scans environments and auto-generates infrastructure-as-code templates to accelerate modernization. Available in the Portal, CLI,

and Operations Center at no extra cost. Azure hardware: limitless power and security. NVIDIA Blackwell Ultra GPUs featured on the 'Frontier Street' activation, plus Azure Cobalt 200 CPU, Azure Boost DPU,

Postgres-compatible service built for cloud-native and AI workloads: up to 3x transactional throughput, auto-scaling to 128TB, scale-out to 3,072 vCores,

and sub-ms multi-zone commit latencies. Built-in DiskANN vector indexing enables semantic search where the data lives. Azure Copilot: agents change the game. New specialized agents span migration, deployment, observability, optimization, resiliency,

and performance. Azure now supports both Claude and GPT frontier models with enterprise-grade security. IQ Revolution: semantic understanding that works. Foundry IQ and Fabric IQ streamline knowledge retrieval across SharePoint, Fabric, and the web,

powered by Azure AI Search. Pre-configured knowledge bases and policy-aware retrieval make agentic AI practical—cutting schema mapping, data wrangling, and prompt-engineering headaches. Azure HorizonDB: PostgreSQL power. Previewed as a fully managed,

Claude comes to Microsoft Foundry: choice for builders. The community lit up about Claude Sonnet and Claude Opus arriving in Foundry alongside OpenAI GPT models. Developers wanted model choice to evaluate tone, safety,

The age of AI small talk is over. Enterprise apps demand more than clever chat — they need a reliable reasoning partner that solves ambiguous, high‑stakes problems, plans multi‑agent workflows, and delivers auditable code.

Built for real industries — finance, healthcare, manufacturing, customer support — with use cases in analytics and decision support, app modernization, ETL auditing, and context‑aware assistants. Start building enterprise agents in Microsoft Foundry today.

Find more information at https://azure.microsoft.com/en-us/blog/introducing-gpt-5-2-in-microsoft-foundry-the-new-standard-for-enterprise-ai/.

Two flavors: GPT-5.2 — the advanced reasoning model for harder problems, better information work, and polished spreadsheets/slides; GPT-5.2‑Chat — an efficient workhorse for info‑seeking, technical writing, learning, and career guidance.

Enterprise success requires structured outputs, reliable tool use, and guardrails. GPT‑5.2 in Foundry delivers multi‑step logical chains, context‑aware planning across large inputs, agentic end‑to‑end execution, and enterprise safety/governance.

OpenAI’s GPT-5.2 is now generally available in Microsoft Foundry — a frontier model series purpose-built for enterprise developers and technical leaders, setting a new standard for a new era of AI.

GPT-5.2 introduces deeper logical chains, richer context handling, and agentic execution that prompts shippable artifacts: design docs, runnable code, unit tests, and deployment scripts produced with far fewer iterations.

At Microsoft Ignite 2025, the era of Frontier transformation was spotlighted: embedding AI across every part of the business to improve decision-making, speed, and create new value. Leaders make AI foundational—rethinking processes and integrating new technologies from the start.

vector indexing, and deep AI/analytics integration for mission-critical apps. Resale enabled offers are available across most Marketplace regions, enabling channel-led sales and simpler listing management for publishers and partners. Partners: explore session highlights, practical guidance,

and the full Azure partner updates—stay connected via Microsoft Partner channels and the Azure News tag. Find more information at https://partner.microsoft.com/en-us/blog/article/azure-updates-december-2025.

The Agent Pre-Purchase Plan (P3) pools 32 services under one flexible plan, simplifying procurement and unlocking FDE support and role-based training. Agent 365 extends user management to govern agent fleets across M365, Azure,

and Foundry. Microsoft Foundry now offers Anthropic Claude models for advanced reasoning in research, coding, and agentic workflows—within unified governance and observability. Azure HorizonDB for PostgreSQL (private preview) brings auto-scaling storage, rapid compute scale-out,

Foundry Control Plane (public preview) centralizes policy, lifecycle governance, and observability for AI agents across the Microsoft Cloud—enterprise-grade governance for safe,

scalable agent management. Foundry IQ and Fabric IQ (public preview) unify agent knowledge and live enterprise data for context-rich RAG and real-time analytics. Microsoft Agent Factory is now available to help organizations move from experimentation to execution faster.

Azure Copilot (private preview) brings specialized agents into the Azure portal, PowerShell, and CLI. ARM-driven scenarios plus advanced Microsoft and partner models streamline migration, assessment, and modernization with data-driven insights, guided workflows, and tighter governance.

Microsoft recognized as a Leader in the 2025 Gartner® Magic Quadrant™ for AI Application Development Platforms, positioned furthest for Completeness of Vision. Leadership reflects a conviction: the next wave of applications is agentic, and real impact needs more than great demos.

and policy enforcement—because “trust but verify” doesn't scale without tooling. Build and run models anywhere with Foundry Models and Foundry Local for low-latency or regulated scenarios.

Find more information at https://azure.microsoft.com/en-us/blog/microsoft-named-a-leader-in-gartner-magic-quadrant-for-ai-application-development-platforms/.

Workflow integration, not just conversation. The shift from chatbot to agent means moving from Q&A to action. Foundry Agent Service supports multi-agent orchestration so agents can hand off tasks, coordinate decisions, and drive end-to-end business processes—deployable into Copilot or apps.

Observability, governance, and models from cloud to edge. Foundry Control Plane gives org-wide visibility, audit trails,

Microsoft Foundry is the unified platform for building, deploying, and governing AI applications. Focused on the elements customers say separate production AI from proof-of-concept: grounding agents in real data, integrating workflows, observability, and flexible model deployment.

Real data, real tools. Agents are only as useful as what they can access. Foundry IQ provides a single secure grounding API and Foundry Tools offers 1,400+ pre-built connectors for document processing, translation, speech, and business systems.

Gartner evaluates vendors on two dimensions: Completeness of Vision (where the platform is headed) and Ability to Execute (can it deliver today). Being furthest on vision isn’t about the boldest roadmap—it's about whether the vision maps to real capabilities customers need.

CES 2026 marks the arrival of NVIDIA Rubin—and Azure’s long-range datacenter strategy was engineered for moments like this. NVIDIA’s next-gen systems slot into infrastructure that anticipated their power, thermal, memory,

cooling abstraction layers for tight thermal windows, next-gen power design (liquid cooling loops, CDU scaling, high-amp busways), and regional AI superfactory modularity for predictable global rollouts. Co-design with NVIDIA enables faster deployment, scaling,

and impact. Find more information at https://azure.microsoft.com/en-us/blog/microsofts-strategic-ai-datacenter-planning-enables-seamless-large-scale-nvidia-rubin-deployments/.

and storage bottlenecks so GPU investments deliver maximum value. Ready for Rubin: NVIDIA Vera Rubin Superchips deliver ~50 PF NVFP4 inference per chip and ~3.6 EF NVFP4 per rack—about 5× the GB200 NVL72 rack.

Azure’s rack architecture already supports Rubin assumptions: gen‑6 NVLink (~260 TB/s), ConnectX‑9 1,600 Gb/s networking, HBM4 thermal planning, SOCAMM2 memory expansion, and multi‑die GPU scaling. Design principles that matter: pod-exchange trays for fast servicing,

large-scale Ampere and Hopper deployments on NVIDIA Quantum-2 InfiniBand helped enable models like GPT-3.5 and record-setting supercomputing clusters. Azure unveiled the first and largest GB200NVL72 and GB300NVL72 rack supercomputers. Systems approach: Azure treats compute, networking, storage,

software, and infrastructure as one integrated platform. High-throughput Blob storage, proximity placement, CycleCloud and AKS tuned for low-overhead scheduling, plus Azure Boost and offload engines, clear IO, network,

Vera Rubin NVL72 racks integrate across Fairwater sites—from Wisconsin and Atlanta to future superfactories. Rubin requires major power, cooling, and optimization upgrades;

Azure’s Fairwater experience demonstrates flexible, stepwise expansion. Proven at scale: Microsoft has years of market-proven design and deployment for scalable AI infrastructure. Early,

and networking needs years ahead. Long-term co-design with NVIDIA ensures Rubin fits Azure’s forward platform design. Building with purpose: Azure AI datacenters are engineered for accelerated computing.

The MongoDB Entity Framework (EF) Core provider has been generally available since May 2024, creating excitement in the .NET developer community. Its adoption continues to grow across various applications, showcasing how it empowers developers to integrate MongoDB seamlessly.

Once set up, querying encrypted data can be done simply, and vector search capabilities add powerful data handling options, allowing developers to leverage MongoDB’s capabilities fully. Dive into this exciting integration and elevate your .NET applications!

Find more information at https://devblogs.microsoft.com/dotnet/mongodb-efcore-provider-queryable-encryption-vector-search/.

Two major features that have been added are Queryable Encryption and Vector Search, allowing the use of key MongoDB capabilities through convenient APIs without needing the underlying .NET C# driver.

Queryable Encryption enables encryption of sensitive data while allowing queries, a crucial feature for industries with strict data governance like healthcare and finance. This translates to enhanced security and simplified compliance efforts.

With EF Core features like LINQ queries, change tracking, and optimistic concurrency, the provider harnesses MongoDB’s flexibility and scalability, ensuring developers can work efficiently in their projects.