Analyst(s): Nick Patience
Publication Date: July 31, 2026
Nscale has agreed to acquire Anyscale, the AI compute platform built by the creators of the open-source Ray framework, in a deal expected to close in the second half of 2026. The acquisition pairs Nscale’s GPU, data center, and power infrastructure with Anyscale’s software for training, fine-tuning, inference, and reinforcement learning at scale. Anyscale will keep its brand and customers, and its roughly 200-person team joins Nscale once the deal closes.
What Is Covered in This Article:
- Nscale has signed a definitive agreement to acquire Anyscale, the company built by Ray’s creators, in a deal expected to close in the second half of 2026; financial terms were not disclosed.
- The deal combines Nscale’s GPUs, data centers, and power with Anyscale’s software layer for processing data, training, fine-tuning, running inference, and deploying agents.
- Anyscale continues operating under its own brand, keeps its existing customers, and remains free to run on infrastructure other than Nscale’s.
- Ray, the open-source framework Anyscale is built on, was donated to the PyTorch Foundation in 2025 and stays community-governed; Nscale is joining the Foundation as part of the deal.
- The acquisition follows a broader pattern of AI infrastructure providers moving up the stack to bundle software with compute.
The News: Nscale, the London-based full-stack AI cloud provider, is acquiring Anyscale, the Berkeley, California-based company behind Ray, the open-source framework for scaling distributed AI applications. Anyscale’s roughly 200 employees, spread across the US, Europe, and India, will join Nscale once the deal closes, expected in the second half of 2026, subject to regulatory approval. Anyscale will continue operating under its own brand and serving its existing customer base, including Coinbase, Bedrock Robotics, and Runway, with customers free to choose which infrastructure they run on. Nscale CEO Josh Payne said the acquisition completes a vertically integrated AI cloud spanning power, data centers, compute, and now software, while Anyscale CEO Keerti Melkote described the combination as building the first ‘full-stack AI hyperscaler.’
Nscale Acquires Anyscale: The Neocloud Land Grab Continues
Analyst Take: Nscale acquires Anyscale at a moment when nearly every well-funded AI infrastructure provider is racing to answer the same question: once you have built or leased enough GPUs, what do you actually sell next? Nscale’s answer, like CoreWeave’s expansion into managed inference and fine-tuning services, is to stop selling raw compute and start selling a full-stack AI cloud. Buying Anyscale gives Nscale a genuinely capable software layer, built by engineers who created Ray and understand exactly how demanding distributed AI workloads behave in production. That is real technical depth for Nscale to build on.
Open-Source Software Complicates the Lock-In Story
The obvious question about Nscale acquiring Anyscale is whether owning the software actually buys Nscale anything durable. Ray is open source, was donated to the PyTorch Foundation in 2025, and runs perfectly well on AWS, Google Cloud, Azure, or any GPU cluster a customer chooses. Nscale’s own messaging reinforces this: customers keep the freedom to run Anyscale wherever they like, and only gain Nscale’s infrastructure as an additional option. That is the right call for adoption, since forcing lock-in onto an open-source community would alienate the developers who made Ray valuable in the first place. But it also means the acquisition, on its own, does not hand Nscale a defensible moat. Anyone can still run Ray on a hyperscaler tomorrow.
The Real Value Is Talent and Co-Design, Not Exclusivity
Anyscale contributes a team of approximately 200 professionals, notably engineers with extensive experience optimizing the Ray framework for production-grade training, fine-tuning, and reinforcement learning. Integrating this deep expertise allows Nscale and Anyscale to collaboratively co-design software and infrastructure from the ground up rather than attempting post-hoc integration, aligning closely with the design philosophy highlighted by Anyscale’s Melkote in the announcement. Capturing even a small portion of workloads requiring tightly coupled infrastructure and orchestration represents a significant victory; furthermore, established Anyscale clients like Coinbase, Bedrock Robotics, and Runway provide Nscale with immediate validation extending past its current GPU clientele.
Neoclouds Are Converging on the Same Playbook
This transaction aligns with a broader trend where GPU-centric cloud suppliers advance up the software stack instead of engaging solely in GPU price competition. This direction is illustrated by CoreWeave acquiring Weights & Biases in 2025 and IREN buying Mirantis earlier this year. Similarly, NVIDIA investing in Nebius and CoreWeave expanding into managed services demonstrates the identical shift: infrastructure firms that previously contended mostly on chip availability are now vying to control larger segments of the AI development lifecycle. As the supply of GPUs increases, competitive advantages shift toward organizations capable of transforming raw capacity into an integrated enterprise platform, eliminating the need to combine disparate vendors for data processing, training, inference, and deployment. Over the coming quarters, the critical factor to track will be Nscale’s ability to successfully deliver on this platform integration rather than merely announcing it.
What to Watch:
- Whether Anyscale’s engineering team, including Ray’s core maintainers, stays intact through the integration now that roughly 200 people are joining Nscale.
- How other infrastructure providers that run Ray workloads for customers respond, and whether any pull back from deepening ties with a now Nscale-owned Anyscale.
- Regulatory approval progress ahead of the expected second-half-2026 close, even though undisclosed terms suggest the deal likely falls below major antitrust thresholds.
- Whether Nscale bundles the Anyscale layer into its own pricing or keeps it available as a standalone, infrastructure-agnostic product, as promised.
- Whether competitors such as CoreWeave, Nebius, or Lambda pursue similar software acquisitions to round out their own full-stack AI cloud claims.
See the complete press release on Nscale’s acquisition of Anyscale on the Nscale website.
Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum as a whole.
Other Insights From Futurum:
UK AI Sector Boosted by $40bn+ Investments from Microsoft, NVIDIA & Others
Nebius Designs the Agentic Era of AI Cloud Platforms with NVIDIA Investment
NVIDIA and CoreWeave Team to Break Through Data Center Real Estate Bottlenecks
Author Information
Nick Patience is VP and Practice Lead for AI Platforms at The Futurum Group. Nick is a thought leader on AI development, deployment, and adoption - an area he has researched for 25 years. Before Futurum, Nick was a Managing Analyst with S&P Global Market Intelligence, responsible for 451 Research’s coverage of Data, AI, Analytics, Information Security, and Risk. Nick became part of S&P Global through its 2019 acquisition of 451 Research, a pioneering analyst firm that Nick co-founded in 1999. He is a sought-after speaker and advisor, known for his expertise in the drivers of AI adoption, industry use cases, and the infrastructure behind its development and deployment. Nick also spent three years as a product marketing lead at Recommind (now part of OpenText), a machine learning-driven eDiscovery software company. Nick is based in London.
