Nscale announced this week that it plans to acquire Anyscale, a move that tests whether cloud-neutral software can remain agnostic once it is owned by a GPU neocloud provider. The deal brings together Nscale’s infrastructure capabilities with Anyscale’s software layer for scaling AI workloads across data processing, training, and inference.
The Deal: Infrastructure Meets Software
The acquisition coalesces Nscale’s control systems, which oversee GPUs, datacenters, power consumption, and the application layer, with Anyscale’s orchestration tools. Nscale describes the combination as “two highly complementary companies” that create a full-stack AI offering. According to the company, this allows users to have Anyscale running on the Nscale fleet as a highly-optimized solution.
Chief product officer Dan Bathurst says the Anyscale platform will continue to operate under its own brand. It will still work with bring-your-own-cloud deployments on AWS, Google Cloud, Azure, and other public clouds. “Where we want to win is on performance, not on any sort of vendor lock-in or forcing of someone to choose Nscale as the infrastructure provider,” Bathurst said.
Related: AI Challenges Mount for Big Businesses
Nscale positions this as “neutrality on the platform layer, but differentiation on the infrastructure layer.” Because the companies are vertically integrated, users who choose the first-party option get a stack co-engineered from power to the application. However, the company insists it will not force this choice on customers.
Runs Anywhere vs. Runs Best Somewhere
Not everyone is convinced that the separation between platform and infrastructure remains distinct. Sanjeev Mohan, a principal analyst at SanjMo and former Gartner research VP, argues that the deal creates a conflict of interest. He says Anyscale “stops being a neutral player” the moment its best features and most optimal pricing land on Nscale hardware first.
He notes that “runs anywhere” and “runs best somewhere” are different things. Buyers will likely feel the gap in performance and cost if Anyscale’s optimizations are exclusive to the Nscale environment. At that point, neutrality becomes a label rather than a reality.
Related: Firms shift cybersecurity funds amid new threats
For teams building in-house AI applications, the practical impact of this integration is significant. When the performance gains are quantifiable, developers often prioritize speed and reliability over long-term portability. The integration of scheduling, memory, and networking—something Mohan calls “the strongest part of the deal”—could create a frictionless environment that makes switching providers less attractive, even if the software technically supports multi-cloud usage.
Open Source Origins
Anyscale was founded by the creators of Ray, an open source project that provides a distributed computing framework.
Financial terms of the transaction were not disclosed in the initial announcement. The company says its full stack offering will help companies speed up image processing, fine-tune LLMs, and deploy AI agents. Keerti Melkote, CEO of Anyscale, defined the combination of the two organizations as the “first full-stack AI hyperscaler.” Representatives from AWS, Google Cloud, and Microsoft Azure were contacted for comments on the acquisition.
