Anyscale on Azure Enters Public Preview for Scalable AI

Anyscale, the AI infrastructure platform built on Ray, has announced the public preview of its integration with Microsoft Azure. This co-engineered solution allows enterprises to deploy and scale AI workloads directly within their Azure environment, leveraging native Azure governance, security, and billing systems. The move marks a significant step in making enterprise-grade AI infrastructure more accessible, particularly for organizations looking to build proprietary AI systems on massive datasets.

The Ray-powered platform, already used by companies like OpenAI and Shopify, enables capabilities such as multimodal data processing, distributed training, and large-scale inference. By integrating natively with Azure, Anyscale simplifies these workflows for enterprises while tying costs into existing Microsoft Azure Consumption Commitments (MACC). This eliminates the need for separate procurement contracts, a notable advantage for enterprise clients.

Why This Matters

As AI adoption shifts from experimentation to production-scale deployment, enterprises increasingly seek to run AI workloads on their own infrastructure for reasons of cost, control, and data sovereignty. This trend, often referred to as “sovereign AI,” is driving demand for platforms like Anyscale that support proprietary data usage and scalable AI systems.

Key features of Anyscale on Azure include:

• Enterprise-grade scalability: Ray, the open-source framework behind Anyscale, has become a go-to for distributed AI workloads. With this Azure integration, enterprises can now access Ray’s capabilities in a managed environment tailored for production.
• Seamless governance: The platform integrates with Microsoft Entra SSO, Azure RBAC, and Azure Policy, ensuring robust security and compliance mechanisms for AI workloads.
• Streamlined billing: Anyscale usage is billed through Azure, appearing on standard invoices and drawing down MACC commitments, simplifying financial management.

Notably, enterprise-use cases already leveraging Anyscale on Azure include Wayve, which is deploying autonomous driving AI, and Xoople, which processes planetary-scale satellite imagery for supply chain and agriculture insights. Both firms report faster experimentation-to-deployment cycles and improved productivity through the platform.

Context and Competitive Landscape

Anyscale’s integration with Azure comes at a time of rapid growth for the company. Founded in 2019 and emerging from UC Berkeley’s RISELab, Anyscale has been pivotal in scaling AI infrastructure. Its flagship framework, Ray, has seen over 500 million downloads and is used by major players like Uber and Instacart. The company has raised $259 million to date and strengthened partnerships with Azure (November 2025) and NVIDIA (March 2026).

By offering a Python-native environment for distributed AI workloads, Anyscale abstracts much of the complexity associated with scaling foundation models and multimodal data processing. This positions the company as a key enabler for enterprises seeking to operationalize AI without the overhead of managing distributed systems in-house.

What’s Next?

The public preview is part of Anyscale’s broader strategy to capture the enterprise AI market. With demand for sovereign AI systems growing, Anyscale’s Azure integration is well-timed to meet the needs of companies looking to move beyond third-party API approaches and build competitive differentiation through proprietary AI models.

Azure customers can start using Anyscale today via the Quickstart Guide. The platform will also be showcased at Microsoft Build and in an upcoming webinar on June 16, 2026.

Image source: Shutterstock