Yes—NVIDIA has released the complete Run:ai platform as open source. Announced October 13, 2026, the release lets teams inspect, build, modify and deploy the platform from source, or install an integrated package from NVIDIA NGC as a Community Product. That Community Product does not require a paid Run:ai subscription or a paid NVIDIA support entitlement. It also does not include NVIDIA enterprise support, and any cloud, infrastructure, implementation or separately contracted enterprise services can still cost money.
What NVIDIA released
The open-source release sits within the NVIDIA DSX AI Factory Platform and is modular rather than a single inseparable application. NVIDIA identifies four components that can be adopted independently or deployed together:
As an Amazon Associate I earn from qualifying purchases.
| Component | What it does |
|---|---|
| KAI Scheduler | GPU-aware scheduling for Kubernetes clusters. |
| GPU Fractioning Operator | CUDA-native sharing of GPUs among workloads. |
| KAI Resource Management | Organizational quotas, hierarchy and resource governance. |
| Workload Map | A consistent contract for managing workloads across the platform. |
The NVIDIA Run:ai repositories are hosted in the DSX AI Factory GitHub organization. Teams can inspect the implementation, contribute changes, build their own distribution or select only the components they need. NVIDIA also describes Model Streamer and Grove as complementary open projects for model loading and multi-component inference orchestration; they are related projects, not additional Run:ai subscription tiers.
License scope
KAI Scheduler is explicitly released under the Apache 2.0 license and contributed to the Cloud Native Computing Foundation as a Sandbox project. The wider platform contains multiple source components, so the license applicable to each component governs its use. “Open source” therefore means the code is available under its stated open-source license, not that every surrounding NVIDIA service or commercial product is automatically free.
#1 Best Overall
- AI Performance: 767 AI TOPS
- OC mode: 2632 MHz (OC mode)/ 2602 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Axial-tech fan design features a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- A 2.5-slot design maximizes compatibility and cooling efficiency for superior performance in small chassis
What is free
Source-code adoption
There is no software purchase required to obtain the published repositories. Subject to each component’s license, a team may read the code, compile it, adapt it, contribute upstream changes or run a self-managed deployment. The team remains responsible for Kubernetes, GPU drivers, cluster security, upgrades, observability and day-to-day operations.
NVIDIA NGC Community Product
NVIDIA packages the integrated stack in NGC as a Community Product, including a Helm chart for deployment. NVIDIA states that installing this Community Product does not require a paid NVIDIA Run:ai subscription or a paid NVIDIA support entitlement.
NGC access still has a legal acceptance step. Users must accept the applicable NVIDIA Software License Agreement and Product-Specific Terms for NVIDIA AI Products. Accepting those terms is a licensing requirement; it is not evidence that a fee is being charged.
What is not included
Enterprise support
NVIDIA’s announcement draws a clear boundary: enterprise support is not included with the Community Product. The free path gives you the software and the community-oriented distribution, but not an NVIDIA enterprise support contract, guaranteed response times, commercial service-level agreement or other support entitlement unless separately arranged.
Infrastructure and operations
Open-source software does not remove the cost of the environment that runs it. Organizations may still pay for GPU servers, storage, networking, Kubernetes operations, security work, staffing, monitoring, electricity or a cloud provider’s instance and data-transfer charges. Those costs are external to the Run:ai Community Product license.
Rank #2
- Powered by the NVIDIA Blackwell architecture and DLSS 4. System Requirements: Minimum 850W PSU with 16-pin 12V-2x6 (12VHPWR) connector required. Verify before purchasing.
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability. Compatibility: 348mm (13.7") length, 3.6 slots, 4.3 lbs. Confirm case clearance and slot spacing. GPU bracket included.
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.6-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
Commercial NVIDIA products and services
NVIDIA’s licensing FAQ distinguishes open-source software and free Community Products from paid Enterprise Products. Its general enterprise model can include several billing structures:
- Annual subscriptions priced per seat or per GPU-year.
- Consumption or pay-as-you-go charges, often calculated per GPU hour through cloud marketplaces.
- Concurrent-user licensing for products that offer that model.
Those are general NVIDIA enterprise licensing patterns, not a published Run:ai price list. The October 2026 Run:ai announcement does not state a Run:ai subscription price, paid feature tier or enterprise-support rate. A dollar figure presented as the Run:ai price would therefore be unsupported.
Choose a deployment route
| Route | What you receive | Who carries the work or cost |
|---|---|---|
| Self-managed source | Individual repositories and the ability to build or modify components. | Your team handles builds, Kubernetes, GPUs, upgrades and support. |
| NGC Community Product | An integrated stack distributed through NGC with a Helm chart. | No paid Run:ai subscription is required; you still operate the environment and accept NVIDIA’s applicable terms. |
| Managed or partner service | A provider may operate GPU infrastructure or an AI-factory deployment for you. | Provider pricing, cloud consumption, implementation fees and support terms must be confirmed directly. |
| Enterprise contract | Potentially contracted support and other NVIDIA enterprise services. | Pricing, entitlements and service levels depend on the negotiated product and agreement. |
Who can provide managed help
NVIDIA names eight DSX ecosystem partners integrating the open platform: Crusoe, Mirantis, Nebius, OpenNebula Systems, Rafay, Saturn Cloud, Spectro Cloud and vCluster. Their integrations may provide managed GPU-cloud or AI-factory services, but NVIDIA’s announcement does not publish affiliate, referral, reseller or commission terms. It also does not establish prices or support-level commitments for those offerings. Treat each as a separate vendor engagement and verify its consumption rates, implementation scope, data handling, maintenance duties and SLA before signing.
How to evaluate the free option
- Decide whether you need the whole stack. Start with KAI Scheduler, GPU Fractioning Operator, KAI Resource Management or Workload Map if one capability solves the immediate problem; use the integrated package when you want the platform’s components together.
- Choose source or NGC. Source deployment maximizes control and customization. NGC’s Helm chart can simplify an integrated installation, but it still requires you to operate the cluster.
- List operational ownership. Assign responsibility for Kubernetes upgrades, GPU drivers, access control, quotas, monitoring, backups and incident response.
- Budget the non-license costs. Include GPU capacity, cloud or data-center charges, engineering time and any external consulting or managed service.
- Decide whether support is a requirement. If your workload needs NVIDIA enterprise response commitments, obtain a separate commercial proposal; do not assume the Community Product includes it.
- Check the governing terms. Review the open-source license for each repository and accept NGC’s Software License Agreement and Product-Specific Terms when using the packaged distribution.
The practical answer for buyers
For an engineering team able to run Kubernetes and GPU infrastructure itself, Run:ai can now be adopted without paying for a Run:ai subscription or NVIDIA support entitlement through the Community Product. The trade-off is operational responsibility and the absence of enterprise support. A managed partner or an NVIDIA enterprise agreement may reduce that burden, but its price and service terms are separate from the open-source release and are not specified in the launch announcement.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
