Nscale Buys Ray Creator Anyscale for $1.65B to Build a Full-Stack AI Cloud

Nscale is paying $1.65 billion to acquire Anyscale, the company behind the widely used open-source Ray framework, in a deal that pushes the British AI cloud provider further up the compute stack. The acquisition folds a software layer for training, serving, and scaling large language models directly into Nscale's infrastructure business.

The price tag marks a substantial premium for Anyscale, which was last valued at $1.38 billion during its 2022 Series C round. That round came as the company pivoted toward AI workload scaling in the wake of GPT-3's launch, a bet that appears to have paid off: Anyscale says its most recent quarter delivered 70% revenue growth quarter-over-quarter.

All roughly 200 Anyscale employees are expected to join Nscale as part of the transaction. Notably, Anyscale will keep operating under its own brand, suggesting Nscale wants to preserve the product's identity and developer mindshare rather than absorb it quietly into a broader platform.

What Anyscale brings to the table

Anyscale's platform, built on the Ray distributed computing framework, handles LLM training and inference, data curation, and reinforcement learning workflows. That breadth is the strategic prize for Nscale: rather than just renting out GPUs and data center capacity, it can now offer customers a managed software layer that sits directly on top of its hardware.

The deal follows closely on the heels of Nscale's own funding surge. In March, the company raised $2 billion in a Series C round at a $14.6 billion valuation, backed by Nvidia, Nokia, Blue Owl, Dell, and Aker. That capital appears to be fueling an aggressive buildout, both in infrastructure partnerships and now in software acquisitions.

Nscale has also lined up compute and data center agreements with Microsoft, British Telecom, and Nordcraft, positioning itself as a serious contender in the race to supply AI infrastructure at scale. Buying Anyscale gives it a ready-made software ecosystem with an established customer base rather than one built from scratch.

Owning more of the stack

The logic behind the acquisition fits a pattern playing out across the AI infrastructure sector: providers that once competed purely on raw compute are racing to control the layers above it, from orchestration frameworks to model-serving tools, in order to lock in customers and capture more margin.

Nscale's move signals that owning GPUs and data centers is no longer enough: control over the software that actually runs AI workloads at scale is becoming the real differentiator.

Whether Anyscale's Ray-based tooling can scale alongside Nscale's infrastructure ambitions without losing its open-source community goodwill will be a key test in the months ahead.