Nscale Agrees to Buy Anyscale in Push Into AI Software

Nscale agreed to acquire Anyscale, adding software for managing AI workloads to the UK company’s growing portfolio of data centers, power supplies and graphics-processing capacity.

Financial terms weren’t disclosed. The transaction is subject to regulatory approval and other closing conditions and is expected to be completed in the second half of 2026, Nscale said in a statement Thursday.

Anyscale will continue operating under its own brand and serving existing customers. Its roughly 200 employees across the US, Europe and India will join Nscale after the acquisition, according to the statement.

The deal would move Nscale further beyond the capital-intensive business of building data centers and renting access to GPUs. Anyscale provides the software layer that developers and corporate AI teams use to distribute data processing, model training, inference and reinforcement-learning workloads across large computing clusters.

Nscale said Anyscale customers would remain free to choose the infrastructure on which they run their workloads. They would eventually gain the option to deploy Anyscale’s software on Nscale’s cloud platform, which combines computing hardware with data-center capacity and power.

Maintaining that infrastructure neutrality may be important for Anyscale, whose platform is designed to manage workloads across different cloud providers and computing environments. It also gives Nscale a route to sell more services to software developers and enterprises rather than competing mainly on the price and availability of GPU capacity.

Anyscale was founded in 2019 by the creators of Ray, an open-source framework designed to scale Python and AI applications across clusters. The company raised $100 million at a $1 billion valuation in 2021, according to its Series C announcement.

Ray was contributed to the PyTorch Foundation, part of the Linux Foundation, in 2025, placing its governance outside Anyscale. Nscale said it would join the foundation following the acquisition. The arrangement means Ray is expected to remain open source and community-governed even as Anyscale changes ownership. The PyTorch Foundation describes Ray as a set of distributed-computing tools for scaling Python applications across different types of computing clusters.

Nscale has raised large amounts of capital to expand its AI infrastructure footprint. The company secured a $2 billion Series C financing in March that valued it at $14.6 billion, following a $1.1 billion Series B round in September 2025. The latest financing was led by Aker ASA and 8090 Industries, with participation from investors including Nvidia, Dell Technologies, Lenovo and Nokia, according to Nscale.

The acquisition reflects a broader effort among AI infrastructure providers to own more of the technology stack, from electricity and data-center real estate to chips, orchestration software and services.

Anyscale Chief Executive Officer Keerti Melkote said combining the company’s platform with Nscale’s infrastructure would allow customers to run AI workloads at greater scale. Nscale CEO Josh Payne said the acquisition would extend the company’s managed services for data processing, training, model fine-tuning, inference and AI agents.