Anyscale was founded in 2019 by Robert Nishihara, Philipp Moritz, and Ion Stoica in San Francisco. The company is built around Ray, the open-source framework created at UC Berkeley’s RISELab for distributed computing and AI workloads.
Anyscale raised over $260 million, including a $100 million Series C led by Addition in 2021 at a $1 billion valuation. The company’s trajectory reflects the growing need for distributed AI infrastructure as model sizes and training datasets expand.
Ray is used at massive scale by companies like OpenAI (for training), Uber, Spotify, Amazon, and Instacart. The framework handles distributed training, hyperparameter tuning, model serving, reinforcement learning, and general Python parallelism under a unified API.
Anyscale’s commercial platform, Anyscale Platform, provides a managed Ray environment with additional features for teams: job scheduling, GPU cluster management, cost controls, and production deployment tools. The company also offers Anyscale Endpoints for hosted LLM inference.
With Ray being a core part of many major AI training pipelines — including those behind some of the world’s most capable models — Anyscale sits at a critical infrastructure layer of the AI stack. The company employs over 200 people and maintains a large open-source community around Ray, with contributions from hundreds of organizations.
