NVIDIA said on 13 October 2026 that it is making the full NVIDIA Run:ai platform available as open source software. In a post on its Technical Blog, it calls Run:ai an open foundation for AI workload and GPU orchestration within its NVIDIA DSX AI Factory Platform.
NVIDIA says the complete open source platform is also offered on NVIDIA NGC as a Community Product, and that no paid NVIDIA Run:ai subscription or NVIDIA support entitlement is required for it. It says commercial, NVIDIA-supported versions remain available.
Business Pill · SWITCHING COSTS
A one-minute explainer of switching costs: what it costs a customer to move from one supplier to another. It teaches the general idea only and says nothing about any organisation in this story.
The key insight: As we read it, NVIDIA is opening the software that decides how its GPUs are shared: it says the full Run:ai platform is now open source, with a Community Product that needs no paid Run:ai subscription, while supported commercial versions remain available.
NVIDIA says the release includes the complete NVIDIA Run:ai platform and four components that can be adopted on their own: KAI Scheduler, GPU Fractioning Operator, KAI Resource Management and Workload Map. It says the source code sits in the NVIDIA Run:ai repository within the NVIDIA DSX AI Factory GitHub organization.
The post says the source code is governed by its applicable open source licenses, and that downloading the NGC Community Product requires accepting NVIDIA’s software license terms. It says NVIDIA enterprise support is not included in the Community Product.
NVIDIA describes Run:ai as software that helps organisations run AI workloads on Kubernetes at scale, by dynamically allocating GPU resources and managing workloads from data processing and training to inference.
What NVIDIA open-sourced on 13 October 2026, per its Technical Blog: the full NVIDIA Run:ai platform plus four components that can be adopted on their own, KAI Scheduler, GPU Fractioning Operator, KAI Resource Management and Workload Map.
The Four Components
KAI Scheduler is the Kubernetes scheduler that powers Run:ai, NVIDIA says. It says it released KAI Scheduler under the Apache 2.0 license and contributed it to the CNCF as a Sandbox project.
GPU Fractioning Operator lets several workloads share one GPU with hard memory limits, NVIDIA says. The post says Run:ai’s earlier proprietary implementation has been retired in favour of the native capability in CUDA 13.4.
KAI Resource Management handles projects, departments, quotas and node pools, NVIDIA says, so GPU capacity can be guaranteed and shared across teams. Workload Map, NVIDIA says, is a Go library that describes the structure of different AI workload types so tools can manage them consistently.
Who Is Building On It
NVIDIA names eight DSX ecosystem partners integrating the open source platform: Crusoe, Mirantis, Nebius, OpenNebula Systems, Rafay, Saturn Cloud, Spectro Cloud and vCluster.
Among them, the post says Nebius offers Run:ai as a one-click deployment in Nebius AI Cloud, Mirantis extends its k0rdent AI stack with it, and Saturn Cloud uses it to help AI cloud operators launch white-label inference services with per-token metering and billing.
For NVIDIA Enterprise AI Factory deployments, NVIDIA says Run:ai stays available through supported commercial offerings, including as a standalone product, through NVIDIA Mission Control and NVIDIA AI Enterprise, and in select cloud marketplaces.
Why NVIDIA Says It Is Doing This
NVIDIA says open sourcing the platform lets every DSX ecosystem partner and AI cloud increase workload throughput, improve GPU utilisation and run shared AI infrastructure more efficiently, and bring differentiated solutions to market faster.
As we read it, the two moves point the same way: NVIDIA is opening the layers around its GPUs, the rack design and now the scheduling software, while the GPUs themselves remain its product.
The Structural Read
As we read it, this is the same direction as NVIDIA’s 12 October rack contribution to the Open Compute Project: the layers around the GPU are being opened, and the GPU stays the product.
As we read it, the partner list is the distributionplan: NVIDIA names eight AI clouds and platform vendors, from Nebius to Crusoe, that are building the open platform into their own offerings.
As we read it, the support line keeps a paid tier: NVIDIA says the Community Product comes without NVIDIA enterprise support, which stays with its commercial offerings.
NVIDIA Technical Blog, 13 October 2026
“No paid NVIDIA Run:ai subscription or NVIDIA support entitlement is required for the Community Product, and NVIDIA enterprise support is not included.”
Three Implications
AI CLOUDS NVIDIA says DSX ecosystem partners and AI clouds can use the platform to increase workload throughput and improve GPU utilisation.
PLATFORM TEAMS NVIDIA says teams can adopt Run:ai as a whole, build on it, or take individual components such as KAI Scheduler.
ENTERPRISE BUYERS NVIDIA says supported Run:ai stays available as a standalone product, through Mission Control and AI Enterprise, and in select marketplaces.
The framework puts it this way: “Close the SCARCE layer (keep proprietary) + Open the ABUNDANT layer (commoditize)”.
As we read it, NVIDIA’s Run:ai move fits that pattern: it opens the orchestration software around its GPUs, the layer it says partners and AI clouds will build on, and keeps the GPUs and supported offerings as the paid layer.
The eight partners NVIDIA names as integrating open source Run:ai, per its 13 October 2026 post.
What Is Not Established
The post does not name the licence that covers the full platform beyond KAI Scheduler’s Apache 2.0, and it gives no adoption, revenue or pricing figures for Run:ai. It does not say how the change affects existing paying Run:ai customers beyond saying commercial offerings continue.
NVIDIA says it has made the full Run:ai GPU orchestration platform open source, with a Community Product on NGC that needs no paid Run:ai subscription and comes without NVIDIA support, eight named partners integrating it, and paid, supported versions still on offer.
A note on sourcing. On 13 October 2026 we read NVIDIA’s Technical Blog post of the same day in full. What the release includes, the licence terms and the partner list are NVIDIA’s own statements. Nothing here is investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
