Key Details of the Deployment

• Announcement date: June 29, 2026 .
• Platform: Claude models are available through Microsoft Foundry, Azure's native model catalog, giving enterprises Azure-native controls, simplified procurement, and a production path for building AI agents .
• Infrastructure: Runs on NVIDIA GB300 NVL72 systems with NVIDIA Quantum-X800 InfiniBand networking for high-speed interconnects .
• Models initially offered: Claude Opus 4.8 and Claude Haiku 4.5 (some sources list Haiku 4) .
• Key features enabled: Prompt caching, extended thinking, and support for use cases ranging from coding and agentic workflows to complex reasoning .
• Enterprise agent focus: Customers can deploy autonomous, domain-specific AI agents and sub-agents using Nvidia's "Secure Agentic Workspace Reference Design" on Azure .
• This is noted as the first time Nvidia and Anthropic have entered a deep co-engineering partnership .

Strategic Partnership Timeline & Investment Commitments

November 18, 2025 — The Founding Alliance

On this date, Microsoft, Nvidia, and Anthropic announced a sweeping strategic partnership :

The combined direct investment into Anthropic was reported as up to $15 billion by WSJ and Bloomberg . Anthropic's valuation was reported in the range of $350 billion following the deal . The partnership notably signaled Microsoft diversifying beyond its longstanding OpenAI relationship .

June 29, 2026 — The GB300 Deployment Milestone

This deployment is the first tangible infrastructure deliverable of the November 2025 alliance — bringing Claude from a compute commitment to a live, enterprise-grade service on the latest Nvidia hardware .

Specific Azure Infrastructure Powering the Deployment

• NVIDIA GB300 NVL72 rack systems — the Blackwell Ultra architecture designed for massive AI inference workloads .
• NVIDIA Quantum-X800 InfiniBand networking for low-latency, high-bandwidth GPU-to-GPU communication .
• Hosted on Microsoft Azure infrastructure via the Microsoft Foundry platform .
• Nvidia also announced at Microsoft Build a unified agentic AI stack spanning Windows devices, cloud, and local deployment, with Claude models running natively on Nvidia GPUs within Azure Foundry Agent Service .

Claude's Expansion onto Other Platforms

Snowflake Cortex AI

• June 1, 2026 (Snowflake Summit '26): Snowflake and Anthropic announced a deepened partnership, integrating Claude models directly into Snowflake Cortex AI for governed, production-ready enterprise AI agents .
• December 2025 (initial agreement): The original multi-year partnership brought Claude 3.5 models to Snowflake Cortex AI .
• June 9, 2026: Claude Fable 5 became available on Snowflake Cortex AI as a private preview on launch day .
• Major enterprises including Block, eSentire, Indeed, and Carvana are already deploying Claude within Snowflake's governed data environment .
• The joint investment is reported at $200 million .

Anthropic's Claude Science Workbench

• Announced June 30, 2026 at Anthropic's "The Briefing: AI for Science" event .
• Claude Science is an AI workbench for scientific research — not a new model, but a specialized app that integrates literature search, data analysis, chart/graph creation, manuscript authoring, and compute resource management into a single environment .
• Designed to do for laboratory research what Claude Code did for software engineering .
• Represents Anthropic's expansion into healthcare and life sciences ahead of its anticipated IPO .
• Available as a beta release from Anthropic .

Significance

This deployment is significant because it:

1. Delivers on the largest AI infrastructure partnership in history — the $15B investment / $30B compute pact from November 2025 becomes a live service .
2. Makes Nvidia a direct strategic investor and co-engineer with Anthropic for the first time, beyond just being a GPU supplier .
3. Gives Azure enterprise customers a direct, managed path to Claude with Azure-native security, identity, and governance instead of going through Anthropic's direct API .
4. Runs on the newest Blackwell Ultra silicon — the GB300 NVL72 — purpose-built for next-generation agentic AI workloads .
5. Sits alongside parallel platform expansions (Snowflake Cortex AI, Claude Science) that broaden Claude's enterprise and scientific reach beyond the Microsoft ecosystem .

Key Details of the Deployment

• Announcement date: June 29, 2026 .
• Platform: Claude models are available through Microsoft Foundry, Azure's native model catalog, giving enterprises Azure-native controls, simplified procurement, and a production path for building AI agents .
• Infrastructure: Runs on NVIDIA GB300 NVL72 systems with NVIDIA Quantum-X800 InfiniBand networking for high-speed interconnects .
• Models initially offered: Claude Opus 4.8 and Claude Haiku 4.5 (some sources list Haiku 4) .
• Key features enabled: Prompt caching, extended thinking, and support for use cases ranging from coding and agentic workflows to complex reasoning .
• Enterprise agent focus: Customers can deploy autonomous, domain-specific AI agents and sub-agents using Nvidia's "Secure Agentic Workspace Reference Design" on Azure .
• This is noted as the first time Nvidia and Anthropic have entered a deep co-engineering partnership .

Strategic Partnership Timeline & Investment Commitments

November 18, 2025 — The Founding Alliance

On this date, Microsoft, Nvidia, and Anthropic announced a sweeping strategic partnership :

The combined direct investment into Anthropic was reported as up to $15 billion by WSJ and Bloomberg . Anthropic's valuation was reported in the range of $350 billion following the deal . The partnership notably signaled Microsoft diversifying beyond its longstanding OpenAI relationship .

June 29, 2026 — The GB300 Deployment Milestone

This deployment is the first tangible infrastructure deliverable of the November 2025 alliance — bringing Claude from a compute commitment to a live, enterprise-grade service on the latest Nvidia hardware .

Specific Azure Infrastructure Powering the Deployment

• NVIDIA GB300 NVL72 rack systems — the Blackwell Ultra architecture designed for massive AI inference workloads .
• NVIDIA Quantum-X800 InfiniBand networking for low-latency, high-bandwidth GPU-to-GPU communication .
• Hosted on Microsoft Azure infrastructure via the Microsoft Foundry platform .
• Nvidia also announced at Microsoft Build a unified agentic AI stack spanning Windows devices, cloud, and local deployment, with Claude models running natively on Nvidia GPUs within Azure Foundry Agent Service .

Claude's Expansion onto Other Platforms

Snowflake Cortex AI

• June 1, 2026 (Snowflake Summit '26): Snowflake and Anthropic announced a deepened partnership, integrating Claude models directly into Snowflake Cortex AI for governed, production-ready enterprise AI agents .
• December 2025 (initial agreement): The original multi-year partnership brought Claude 3.5 models to Snowflake Cortex AI .
• June 9, 2026: Claude Fable 5 became available on Snowflake Cortex AI as a private preview on launch day .
• Major enterprises including Block, eSentire, Indeed, and Carvana are already deploying Claude within Snowflake's governed data environment .
• The joint investment is reported at $200 million .

Anthropic's Claude Science Workbench

• Announced June 30, 2026 at Anthropic's "The Briefing: AI for Science" event .
• Claude Science is an AI workbench for scientific research — not a new model, but a specialized app that integrates literature search, data analysis, chart/graph creation, manuscript authoring, and compute resource management into a single environment .
• Designed to do for laboratory research what Claude Code did for software engineering .
• Represents Anthropic's expansion into healthcare and life sciences ahead of its anticipated IPO .
• Available as a beta release from Anthropic .

Significance

This deployment is significant because it:

1. Delivers on the largest AI infrastructure partnership in history — the $15B investment / $30B compute pact from November 2025 becomes a live service .
2. Makes Nvidia a direct strategic investor and co-engineer with Anthropic for the first time, beyond just being a GPU supplier .
3. Gives Azure enterprise customers a direct, managed path to Claude with Azure-native security, identity, and governance instead of going through Anthropic's direct API .
4. Runs on the newest Blackwell Ultra silicon — the GB300 NVL72 — purpose-built for next-generation agentic AI workloads .
5. Sits alongside parallel platform expansions (Snowflake Cortex AI, Claude Science) that broaden Claude's enterprise and scientific reach beyond the Microsoft ecosystem .