Microsoft and Anyscale Launch AI-Native Compute Service on Azure Cloud Platform

In a strategic partnership announced today, Microsoft and AI infrastructure company Anyscale have unveiled a new AI-native compute service on Azure, marking a significant advancement in cloud-based artificial intelligence capabilities. The collaboration brings Anyscale's Ray distributed computing framework to Azure as a first-party managed service, targeting enterprises struggling with the computational demands of modern AI workloads.

The service, now in private preview with general availability expected in 2026, represents Microsoft's latest move to strengthen its position in the competitive AI infrastructure market. By integrating Anyscale's technology directly into Azure, Microsoft aims to address what both companies identify as a critical bottleneck in AI adoption: the complexity of scaling AI applications across distributed computing environments.

Anyscale, founded by the creators of the open-source Ray framework, brings specialized expertise in distributed AI computing to the partnership. Ray has become a foundational technology in AI infrastructure, with over 27 million monthly downloads and adoption by major companies including Uber, Spotify, and Coinbase. The framework's ability to coordinate diverse computing resources—from CPUs to specialized GPUs—makes it particularly suited for complex AI workloads involving multimodal data processing and generative AI applications.

The new Azure service promises to deliver up to 10x performance improvements compared to self-managed Ray deployments, according to performance benchmarks shared by the companies. This acceleration comes from the Anyscale Runtime, a performance-optimized engine that maintains compatibility with standard Ray applications while eliminating the operational overhead of cluster management.

For enterprise developers, the integration means streamlined access to distributed AI computing through familiar Azure interfaces. Teams can deploy Ray clusters directly from the Azure Portal, leverage cluster-backed development environments, and maintain their AI workloads within Azure's security and compliance framework. The service runs on Azure Kubernetes Service (AKS), providing cloud-native orchestration for both batch processing and low-latency inference workloads.

The timing of this partnership reflects the growing enterprise demand for turnkey AI infrastructure solutions. As organizations move beyond experimental AI projects to production-scale deployments, many face challenges coordinating the complex mix of data pipelines, model training, and inference services required for modern AI applications. The Microsoft-Anyscale collaboration aims to eliminate these friction points through a fully managed service that maintains the flexibility of open-source tools while providing enterprise-grade reliability.

This announcement positions Microsoft to compete more effectively in the rapidly evolving AI infrastructure market, where cloud providers are racing to offer specialized services for AI development and deployment. The partnership with Anyscale represents a strategic bet that enterprises will prefer integrated, managed solutions over assembling their own AI infrastructure stacks from disparate open-source components.