Microsoft Foundry

Microsoft Foundry

Microsoft Foundry is Microsoft's enterprise platform for building, grounding and governing AI applications and agents. Microsoft Foundry bundles a model catalogue of more than 11,000 models spanning OpenAI, Anthropic, Meta, Google, xAI and Hugging Face with an agent service, pre-built tools for OCR, translation and speech, a grounding layer, a control plane for governance and monitoring, and an edge runtime for on-device work. It is the product formerly known as Azure AI Studio and then Azure AI Foundry, so older documentation and tutorials still use those names.

Production credibility: Microsoft publishes the platform's composition openly: Foundry Models with access to over 11,000 foundational, open, reasoning, multimodal and industry-specific models; Foundry Agent Service; Foundry Tools; Foundry IQ; Foundry Control Plane; and Foundry Local. Pricing is consumption-based and set per service rather than as a single subscription, with cost estimation tools provided. The naming history matters when evaluating: Azure AI Studio became Azure AI Foundry and is now Microsoft Foundry, which is why search results for all three names describe the same product.

Key Features

• Model catalogue spanning OpenAI, Anthropic, Meta, Google, xAI and Hugging Face
• Foundry Agent Service for building and running autonomous agents
• Foundry Tools with ready-made OCR, translation and speech capabilities
• Foundry IQ as the grounding layer over business data
• Foundry Control Plane for unified governance and monitoring
• Foundry Local for edge and on-device inference

Ideal Use Case

Enterprises standardised on Microsoft — Entra identity, Azure networking, Microsoft 365 data — that want AI applications governed by the same controls as everything else. The breadth of the model catalogue also suits teams that expect to switch models often and would rather not renegotiate a contract each time.

How Microsoft Foundry differentiates

Against Bedrock and Vertex AI the platform capabilities are broadly comparable, and the choice usually follows the cloud you are already in. Foundry's distinguishing pieces are the size of the model catalogue and Foundry Local, which extends the same platform to edge and on-device inference rather than keeping everything in the cloud. The cost of that breadth is naming confusion: three brand names in three years means the documentation trail is inconsistent, and buyers should confirm which name their tooling expects.

FAQ

Q: What is Microsoft Foundry? A: Microsoft Foundry is Microsoft's enterprise AI platform for building, grounding and governing AI applications and agents. It includes a model catalogue, an agent service, ready-made tools, a grounding layer, a governance control plane and an edge runtime.

Q: Was Microsoft Foundry called something else? A: Yes. It was Azure AI Studio, then Azure AI Foundry, and is now Microsoft Foundry. All three names refer to the same platform.

Q: How many models does Foundry offer? A: Microsoft states the catalogue holds more than 11,000 models, spanning OpenAI, Anthropic, Meta, Google, xAI, Hugging Face and industry-specific options.

Q: How is Microsoft Foundry priced? A: Pricing is consumption-based and set per service, so each component has its own billing model. Microsoft provides cost estimation tools rather than a single subscription price.

Q: How does Foundry differ from Azure Machine Learning? A: Azure Machine Learning is the platform for training and operating your own machine learning models. Foundry is aimed at building applications and agents on foundation models, with grounding and governance built in.

tl;dr

Microsoft Foundry is Microsoft's enterprise AI platform — 11,000-plus models, an agent service, grounding through Foundry IQ, a governance control plane and Foundry Local for the edge. Formerly Azure AI Studio and Azure AI Foundry; billed per service on consumption.