We’re making it easier to bring production AI with Ray on Azure with availability of Anyscale Platform on Microsoft Azure Marketplace. This gives enterprises a faster, lower-friction path to build, scale, and operate AI workloads with Anyscale on Azure Kubernetes Service (AKS).

LinkAzure customers now get:

• Faster procurement & consolidated billing: Leverage your Microsoft relationship and receive marketplace purchase as part of Microsoft invoice.

• Use existing Azure commitments: Purchase Anyscale Credits that can be used towards your Microsoft Azure Consumption Commitment (MACC). 

• Flexible commercial terms: Work with us on private offers to access volume-based discounts or enterprise support plans.

LinkWhy Anyscale for Ray?

Forward thinking organizations don't just adopt AI, they build AI platforms to maintain control, differentiation and long-term advantage. Ray has become the distributed compute framework for all AI workloads of any scale. Ray is the only framework that can distribute all AI workloads from multimodal data processing and distributed training to LLM inference – batch or online. Anyscale turns a powerful framework into a production-ready platform from day one.

• Innovate faster with purpose-built developer tooling: Interactive development on Ray clusters is complex: you must set up a multi-node backed IDE, manage dependencies across multi-node clusters, and build observability from logs that vanish when clusters shut down. These hurdles slow down debugging and optimization cycles required to accelerate dev to prod. Anyscale simplifies this with developer tools purpose-built for Ray including multi-node IDE (Anyscale Workspaces) and built-in dashboards for Ray libraries (Ray Data and Ray Train). Integrated into the Anyscale web console, these and other dev tools streamline development to deployment cycles.

• Reduce operational costs with reliable, elastic clusters: AI workloads often fail from out-of-memory errors, node failures, or resource contention – issues that demand constant monitoring and manual intervention at scale. Anyscale removes this operational burden with fully managed clusters on Kubernetes both cloud & on-prem. This is done via proactive health monitoring that replaces unhealthy nodes, elastic scaling to meet changing workload demands, and persistent logs accessible via Anyscale managed Grafana or your preferred infra monitoring tool.

• Maximize infrastructure investments with RayTurbo: Getting the most out of GPUs and other hardware requires deep expertise in tuning and optimizing AI workload execution and infrastructure management. Anyscale helps teams scale adoption efficiently with RayTurbo, performance optimizations not available in open-source, that help organizations maximize utilization to more quickly achieve ROI on AI investments. 

• De-risk projects with the leading Ray experts on your team: Deploying AI at scale demands expertise to ensure performance and scalability. Without it, teams risk delays, inefficiencies, or costly mistakes that slow down time to production and reduce ROI. Anyscale is built by the creators of Ray and beyond the platform provides enterprises access to the leading experts via enterprise support plans, training and professional services.

LinkGetting started

• Contact us via Azure Marketplace to create your private offer.

• Learn Ray via Anyscale templates and courses.

• Check out our documentation on Anyscale Kubernetes deployments.

LinkJoin us Ray Summit 2025

See real-world examples and see Ray and Anyscale on Azure product announcements on 2-day, community-led conference that brings together the best engineers, researchers and open source technologies to advance AI.

San Francisco | Nov 3-5, 2025.

Visit Ray Summit 2025 site to learn more about speakers, pricing and agenda.