Anyscale
Category: AI Infrastructure
Anyscale is the company behind the Ray open-source distributed computing framework, providing a fully managed AI platform for scaling AI workloads across any cloud. Anyscale was founded in 2019. The company is led by Keerti Melkote. Based in San Francisco, California, United States. Team size: 101-500. Total funding raised: $281M. Latest round: Series C. Key investors include Andreessen Horowitz, NEA, Addition, Intel Capital, Foundation Capital, Ant Financial, Amplify Partners, 11.2 Capital, The House Fund.
- Founded
- 2019
- Headquarters
- San Francisco, California, United States
- Team size
- 101-500
- Total funding
- $281M
Value proposition
Anyscale enables developers to build, run, and scale any AI workload on Ray across any cloud without managing infrastructure, accelerating AI development from laptop to production.
Products and solutions
Anyscale Platform (fully managed Ray), RayTurbo (optimized runtime), Anyscale Endpoints, Anyscale Operator for Kubernetes, Ray open-source framework (Ray Core, Ray AIR, Ray Serve, Ray Data, Ray Train, Ray Tune, KubeRay)
Unique value
The only fully managed platform built by the creators of Ray, offering enterprise governance, advanced developer tooling, and expert support for scaling AI workloads across any cloud or on-premise infrastructure.
Target customer
AI/ML engineering teams at enterprises and technology companies that need to scale training, serving, and data processing workloads across distributed GPU clusters.
Industries served
Technology, Financial Services, E-commerce, Media & Entertainment, Healthcare, Autonomous Vehicles, Gaming
Technology advantage
Founded by the creators of Ray (UC Berkeley RISELab); Ray is the fastest-growing open-source distributed computing framework for AI; deep integration with PyTorch, TensorFlow, Hugging Face; cloud-agnostic (AWS, GCP, Azure, on-prem); KubeRay for Kubernetes-native deployment; RayTurbo proprietary optimizations
How they differentiate
Ray is the de facto standard open-source distributed computing framework for AI (used by OpenAI, Uber, Spotify, Netflix); Anyscale offers the only fully managed enterprise platform built by Ray's original creators; cloud-agnostic architecture; deep Python-native developer experience
Main competitors
Modal, RunPod, Northflank, Databricks, AWS SageMaker, Google Vertex AI
Key partnerships
Microsoft Azure (Anyscale on Azure public preview), AWS (AWS Marketplace integration), Google Cloud, NVIDIA, Intel, PyTorch Foundation (Ray joined as a hosted project)
Notable customers
OpenAI, Uber, Spotify, Netflix, Amazon, Microsoft, Shopify, Coinbase, IBM, Meta, Instacart, Riot Games
Major milestones
2016-2017: Ray developed at UC Berkeley RISELab, 2019: Anyscale founded, $20.6M Series A, 2020: $40M Series B led by NEA, 2021: $100M Series C at $1B+ valuation (unicorn), 2022: Additional $99M Series C extension, Ray 2.0 launch, 2024: Keerti Melkote appointed CEO, 2026: Acquired by Nscale for $1.65B
Growth metrics
Revenue grew 70% in most recent sequential quarter (Q2 2026); 4x revenue growth in year prior to CEO transition (2023-2024); ~200 employees at time of acquisition
Market positioning
Market leader in open-source distributed AI compute orchestration via Ray; positioned as the infrastructure layer between raw GPU cloud providers and ML application frameworks; competes with both hyperscaler ML platforms and emerging AI compute startups
Geographic focus
Global (headquarters in San Francisco, US; office in Bangalore, India)
About Keerti Melkote
Founder & CEO of Aruba Networks (IPO 2007, acquired by HPE for $3B); Engineering leadership roles at Intel and Cisco
Latest news about Anyscale
More AI Infrastructure companies
Official website: https://www.anyscale.com
