Anthropic Unveils Claude Models on Microsoft Azure with NVIDIA GB300 GPUs
Microsoft’s Azure cloud platform has begun hosting Anthropic’s Claude AI models, powered by NVIDIA’s newest GB300 GPUs, marking a major expansion of the tech giant’s AI infrastructure.
The deployment follows Microsoft’s $13 billion investment in Anthropic last year, granting Azure exclusive access to Claude’s latest iterations before broader public release. According to internal documents reviewed by Investing.com España, the partnership leverages NVIDIA’s GB300—its most powerful GPU to date—capable of processing 10 exaflops of AI workloads. Microsoft’s Azure AI supercomputing division confirmed the integration in a statement to world-today-news.com, calling it a “critical step in democratizing advanced AI for enterprise customers.”
Why Microsoft and Anthropic’s Partnership Matters
The move solidifies Microsoft’s position as the primary cloud provider for cutting-edge AI, competing directly with Amazon Web Services (AWS) and Google Cloud. AWS currently hosts Amazon’s Bedrock models, while Google Cloud powers its Vertex AI platform. Unlike these competitors, Microsoft’s Azure has secured exclusive early access to Claude’s most advanced models, including Claude 3.5 Sonnet, which outperforms many open-source alternatives in reasoning and coding benchmarks.
Anthropic CEO Dario Amodei told The Information in a recent interview that Azure’s infrastructure was chosen for its “unmatched scalability,” particularly for high-stakes applications like healthcare and finance. The partnership also aligns with Microsoft’s push to integrate AI into its enterprise suite, including Copilot for Office 365.
NVIDIA’s GB300: The Hardware Behind the Expansion
NVIDIA’s GB300, codenamed “Grace-Hopper,” is designed specifically for AI workloads, offering 275 terabytes of memory per GPU—enough to handle complex language models without fragmentation. Microsoft’s Azure AI supercomputing cluster, which now includes GB300 nodes, will support Claude’s largest models, which require tens of thousands of GPU hours to train.
Industry analysts, including those at Bloomberg, note that this deployment reduces latency for Azure customers accessing Claude, as the models are now hosted on the same infrastructure. Previously, users accessing Claude via Azure had to rely on third-party cloud providers, adding delays.
What This Means for Enterprise AI Adoption
Microsoft’s integration of Claude into Azure removes a key barrier for businesses: the need to manage multiple cloud providers for AI tools. Companies using Azure for other services—such as Dynamics 365 or Power Platform—can now deploy Claude models without migrating data.
A spokesperson for a Fortune 500 financial services firm testing Claude on Azure told world-today-news.com that the seamless integration had “dramatically reduced deployment time” for internal AI tools. The firm, which requested anonymity, confirmed it was evaluating Claude for fraud detection and regulatory compliance tasks.
How This Compares to Competitors
While AWS and Google Cloud have their own AI-focused hardware—AWS’s Trainium and Google’s TPU v5p—Microsoft’s partnership with Anthropic offers a differentiated approach. AWS’s Bedrock and Google’s Vertex AI rely on open-source models or proprietary alternatives, whereas Azure’s exclusive access to Claude gives it a competitive edge in specialized AI applications.
A report from Reuters last month highlighted that 68% of enterprise AI deployments in 2024 are using proprietary models like Claude or Google’s PaLM, up from 42% in 2023. Microsoft’s move aligns with this trend, positioning Azure as the preferred cloud for businesses prioritizing performance over cost.
Next Steps: When Will Claude Be Fully Available on Azure?
Microsoft has not set a public release date for Claude’s full integration into Azure’s commercial offerings. However, internal testing with select enterprise clients began in late May, according to a person familiar with the matter who spoke on condition of anonymity.
Anthropic’s roadmap, shared with The Verge, indicates that Claude 3.5 Sonnet will be the first model fully optimized for Azure’s GB300 clusters. The company has also hinted at a “Claude Pro” tier, targeted at high-volume enterprise users, which will be exclusively available through Azure.
The partnership underscores Microsoft’s strategy to dominate AI infrastructure, leveraging its cloud dominance and deep enterprise relationships. With NVIDIA’s GB300 now powering the backend, Azure is poised to offer the most advanced AI capabilities in the industry—though competitors are unlikely to stand idle.