Nscale Acquires Anyscale to Build 'Full-Stack' AI Hyperscaler
The UK-based neocloud provider is absorbing the creators of the Ray framework in a $1.65 billion deal.
UK-based AI neocloud provider Nscale has entered into a definitive agreement to acquire Anyscale, the company founded by the creators of the open-source Ray framework. The move signals a strategic shift for Nscale as it evolves from a raw compute provider into a vertically integrated AI hyperscaler.
According to Bloomberg, the acquisition is valued at approximately $1.65 billion. The deal will bring roughly 200 Anyscale employees from the US, Europe, and India into the Nscale fold. Nscale, which reached a valuation of $14.6 billion following a $2 billion Series C funding round in 2026, expects to close the transaction by the end of 2026.
The Shift to Vertical Integration
Historically, the AI infrastructure market has been divided between massive hyperscalers like AWS and Azure and "neoclouds" that focus primarily on GPU availability. Nscale distinguishes itself by owning the entire physical layer, including power and data centers. By absorbing Anyscale, Nscale is moving "up the stack," adding a sophisticated software orchestration layer to its physical assets.
"Most infrastructure providers just buy GPUs and rent them," said Nscale CEO and Founder Josh Payne. "We build and own every layer ourselves: the power, the data centers, the compute, and the software that turns them into an AI cloud."
Implications for AI Neutrality
This acquisition raises critical questions regarding multi-cloud neutrality. Anyscale's Ray framework has served as a standard for scaling distributed AI applications across different providers, preventing developers from being locked into a single vendor's ecosystem. While Ray was donated to the PyTorch Foundation in October 2025 and remains open source and community-governed, the ownership of the company that supports it could influence how managed services are prioritized.
If Nscale optimizes its managed services specifically for its own hardware, it could increase vendor lock-in for developers. However, if the integration succeeds, it provides a blueprint for a new class of AI clouds that can compete with Big Tech by controlling everything from cooling to orchestration.
What's Next
Anyscale CEO Keerti Melkote stated that combining the Ray-based platform with Nscale's infrastructure creates the first "full-stack AI hyperscaler." Market observers will now watch to see if Nscale maintains the cloud-neutral appeal of Anyscale's tools or pivots toward a closed ecosystem to maximize its own infrastructure utilization.