Anyscale Unveils Ray 2.0 and Anyscale Innovations at Ray Summit 2022; Adds an Additional $99M Funding from Existing Investors Addition, Intel Capital, and Foundation Capital

SAN FRANCISCO, Aug. 23, 2022 (GLOBE NEWSWIRE) -- Anyscale, the company behind Ray, the unified framework for scalable computing, today announced Ray 2.0 and the enterprise-ready capabilities and roadmap for Anyscale’s managed Ray platform at the Ray Summit. This year’s Summit features dozens of organizations scaling their AI initiatives with Ray including Uber, IBM, Meta, Riot Games, Instacart and more. In addition, Anyscale also announced today that it has secured another $99 million in Series C funding co-led by existing investors Addition and Intel Capital with participation from Foundation Capital. The funding follows the $100 million Series C funding round announced in December 2021.

The accelerated adoption of Ray is driven by the growing gap between the demands of machine learning (ML) applications and the limitations of a single processor or a single server. In the past few years alone, the computational requirements for ML training have been growing between 10 to 35 times every 18 months. This fact, combined with the engineering complexity of scaling these workloads, has led to over 85 percent of AI projects failing in production. Ray tackles the cost and complexity of scaling head-on and is the fastest growing open-source, unified distributed framework for scaling AI and Python applications.

“Ray and the Anyscale platform have made tremendous progress in advancing the scaling of machine learning and Python workloads,” said Anyscale CEO, Robert Nishihara. “Thousands of organizations already rely on Ray for their AI initiatives and dozens of them are showcasing their use cases and breakthroughs at this year’s Ray Summit. With new innovations in Ray 2.0 and Anyscale’s platform, we are further accelerating our efforts to ensure Ray is easily accessible to any developer and to organizations of all sizes.”

“Ray is quickly becoming the industry standard for scaling machine learning, Python and AI workloads, solving one of the biggest obstacles today to realizing AI’s full potential,” said Nick Washburn, Senior Managing Director at Intel Capital. “The rapid adoption of Ray positions Anyscale to unlock the growing market opportunity in AI.”

Today, the first day of Ray Summit, the company unveiled major new developments in Ray 2.0. Highlights include:

• Ray AI Runtime (AIR)*

  • Ray AIR adds a unified runtime layer for ML applications and services.

  • The runtime simplifies ML application development, increases developer velocity, and is cloud and framework agnostic, enabling interoperability with popular frameworks such as Tensorflow, PyTorch, Hugging Face and more.

  • Ray AIR overcomes the hurdles of proprietary cloud and vendor solutions and is available now in beta.

• KubeRay

  • KubeRay is an open source project with contributors from Anyscale, Bytedance, Microsoft, Spotify, Union.ai and others.

  • The KubeRay toolkit enhances the robust execution of Ray applications on Kubernetes, an open-source system for managing containerized applications across multiple hosts.
    

• Integration with the ML Ecosystem
  

  • Seamless interoperability with ML ecosystem tools like Weights & Biases and Hugging Face Transformers allows developers to combine the advantages of Ray for scaling with the broader ML ecosystem including ML libraries, data platforms, and MLOps platforms.

• Ease of Development & Increased Scalability

  • New Ray 2.0 optimizations support petabyte-scale, compute-intensive workloads.

  • New observability tooling eases development and debugging, enabling faster iteration cycles for developers and faster time-to-market.
    

Anyscale today also introduced its enterprise-ready Ray platform, a unified compute platform that makes it easy to build, manage, and bring to market scalable AI and Python applications using Ray. Anyscale highlights include:

• Next Generation ML Workspace

  • The Anyscale ML Workspace enables users to develop and scale ML applications from prototype to production and back for debugging. This experience makes moving AI applications and ML pipelines to production easier and reduces context switching.

  • The Anyscale ML Workspace eases deployments by integrating with best-in-class ML tools like Weights & Biases and Arize AI. This allows users to scale compute for machine learning while leveraging the best MLOps tools.
    

• New Enterprise Platform
  

  • The Anyscale Enterprise Platform gives IT and security teams secure cluster connectivity and customer-managed virtual private clouds.

  • The Enterprise Platform also offers users activity auditing, operational monitoring and cost management capabilities to enable organizations to track spending. This functionality is currently in preview.
    

Join Ray users, global organizations, and partners and learn from AI/ML technical experts at Ray Summit 2022. To learn more and view this year’s keynote talks, visit https://www.anyscale.com/ray-summit-2022.

About Anyscale
Anyscale is the company behind Ray open source, the unified compute framework for scalable computing. Ray is the fastest growing open-source unified distributed framework to easily scale AI and Python applications. Ray enables developers to effortlessly scale any workload or application from a laptop to the cloud, without the cost or expertise required to build complex infrastructure. Anyscale is a fully-managed, enterprise-ready unified compute platform that makes it easy to build, deploy, and manage scalable AI and Python applications using Ray, allowing organizations to bring AI products to market faster at significantly lower costs. Backed by Andreessen Horowitz, NEA, Addition, Intel Capital and Foundation Capital, Anyscale is headquartered in San Francisco, CA. www.anyscale.com

Contact and press inquiries: anyscale@launchsquad.com

*We are sunsetting the "Ray AIR" concept and namespace starting with Ray 2.7. The changes follow the proposal outlined in this REP.