Nscale Joins the PyTorch Foundation as a Platinum Member

The PyTorch Foundation announced on 20 October 2026 that Nscale has joined as a Platinum member. Nscale describes itself as a full-stack AI cloud platform that combines power, data centers, compute and software in one vertically integrated offering, spanning a unified cloud platform for AI training and inference through to the facilities and low-cost power beneath it. The announcement was dated London, U.K.
The relationship predates the membership. Nscale entered the Foundation's project ecosystem through Ray, the open-source framework for distributed computing and scaling AI workloads, which has been governed under the Foundation since October 2025. Nscale is planning to acquire Anyscale, the company that created Ray, and has committed to continued investment in Ray's development and its contributor community. That commitment is the substantive part of the news for open-source users: an infrastructure provider is promising to keep funding a framework it is about to own.
Platinum membership carries one seat on the PyTorch Foundation Governing Board, which sets policy through the Foundation's bylaws, mission and vision statements and defines the scope of its initiatives, technical vision and direction. Nick Jones, VP of Cloud Native Platform Engineering at Nscale, takes that seat. He leads engineering strategy for Nscale's high-performance GPU cloud and brings more than 15 years of experience across cloud-native infrastructure, AI and high-performance computing, with prior roles at SUSE, Rancher and Mesosphere and contributions to projects spanning Kubernetes, OpenStack and distributed systems.
Eyal Lantzman, VP of AI Services at Nscale, joins the Foundation's Technical Advisory Council. He leads Nscale's AI services stack, focused on serving and optimizing large open-weight models in production, and has more than 20 years of experience across AI/ML platforms, cloud infrastructure and developer tooling, including previous roles at JPMorgan Chase, AWS and Microsoft, with a long-standing focus on cloud security, governance and sovereignty.
In a statement, Jones framed the membership as vertical integration rather than sponsorship: most infrastructure providers, he said, treat the framework and the data center as separate problems, while Nscale treats them as one. He described PyTorch as the base layer of AI because it kept the Pythonic style researchers know while getting serious about performance, and Ray as the piece that gets that work running across thousands of GPUs. With Anyscale set to join Nscale, he said the company is building the infrastructure underneath that ecosystem rather than only consuming it.
Mark Collier, Executive Director of the PyTorch Foundation, said frontier AI workloads are scaling to multi-trillion parameters and that tight integration between PyTorch, distributed frameworks such as Ray and the underlying cluster infrastructure is essential. He described Nscale's membership as a commitment to open source, to removing friction for developers, to optimizing multi-node performance and to making PyTorch run out of the box on advanced hardware.
The PyTorch Foundation is hosted by the Linux Foundation and positions itself as a vendor-neutral home for the open-source intelligence layer used to train, optimize, serve and run models on any chip, in any cloud, for any agent. Beyond the core PyTorch framework, its portfolio includes vLLM for inference serving, DeepSpeed for training optimization, Ray for distributed computing, Helion for kernel authoring and Safetensors for model weight serialization.
This is where the membership becomes more than a logo. The Foundation's projects sit at different layers of a modern AI stack: PyTorch defines the model and training loop, Ray distributes that work across a cluster, DeepSpeed and Helion squeeze performance out of it, vLLM serves it, and Safetensors handles the weights. A cloud provider that also controls a distributed-computing framework and operates the data centers underneath has a strong interest in how those layers are tuned for one another. In practice, the useful signals for developers will be concrete: whether Ray's scheduling and fault tolerance improve on large multi-node jobs, whether PyTorch builds stay healthy on Nscale's GPU fleet, and whether the Anyscale acquisition closes on the terms implied here.
A likely trade-off is governance attention. Platinum seats buy influence, not control, and the Governing Board and TAC are where vendor interests get balanced against the community's. Nscale's commitments to Ray's contributor community are statements of intent; how they hold up once Anyscale is inside the company will be the thing to watch. The source gives no timeline for closing the acquisition and no specifics on the size or shape of the planned Ray investment.
Ray's presence in the Foundation since October 2025 already gives the project neutral ground, and Nscale says its contribution builds on that governance rather than replacing it. The open question, from the outside, is whether an owner-operator of both the framework and the hardware produces tighter integration or quieter divergence from other clouds that also run PyTorch and Ray.
Why it matters: Teams running distributed training and inference on PyTorch and Ray are the audience most directly affected, since a major infrastructure provider now has board-level and technical-council influence over both projects, and its stated priorities are multi-node performance and out-of-the-box hardware support. Developers should expect, as an inference from the stated commitments rather than a confirmed roadmap, more emphasis on cluster-level integration between Ray and PyTorch. For product teams choosing a cloud, the practical test is whether that integration shows up as measurable throughput and reliability gains on Nscale and elsewhere, and whether Ray's open development stays open after the Anyscale deal completes.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
London, U. K., 20 October 2026 – The PyTorch Foundation, the vendor-neutral home for the open source intelligence layer, is announcing today that Nscale has joined as a Platinum member.... The post Nscale Joins the PyTorch Foundation as a Platinum Member appeared first on PyTorch .