NVIDIA announced an expansion of its AI infrastructure in an announcement dated November 10, 2026. The donation of OpenShell and new DSX components are the key points. Developers can evaluate NVKE, which is in preview.
This article follows NVIDIA's announcement and official documentation, with information checked as of November 12, 2026.
OpenShell donation and new DSX components
NVIDIA announced that it would donate OpenShell to CNCF. CNCF stands for Cloud Native Computing Foundation. The same announcement also introduced new DSX components. Details are available in the official NVIDIA announcement.
The following components have been added to the NVIDIA DSX AI Factory Platform.
- NVIDIA Kubernetes Engine, abbreviated as NVKE
- NVIDIA Kubernetes Recipes, abbreviated as NVKR
- NVIDIA Snapshot
OpenShell is an open-source runtime for running groups of autonomous AI agents. A runtime is the environment in which a program runs. OpenShell uses sandboxes isolated at the kernel level. This description is based on the official OpenShell documentation.
NVIDIA announced the donation of OpenShell to the Cloud Native Computing Foundation.AI-generated image
NVKE is available for evaluation in preview
NVKE is available as a preview, not as a production release. Developers can evaluate it through the public project. However, NVIDIA advises against using it for production workloads. Production refers to the environment where live services run.
NVKE's APIs, configuration, and behavior may change between releases. These limitations should be checked first when considering its use in live services. The preview warning is stated in the official NVKE repository.
Apache 2.0 licensing and commercial support pricing are separate
NVKE and NVKR are licensed under Apache 2.0. Both projects also accept external contributions. Commercial support pricing for the new components has not been specified.
| Item to check | Information confirmed in official documentation |
|---|---|
| NVKE license | Apache 2.0 |
| NVKR license | Apache 2.0 |
| External contributions | Allowed for both NVKE and NVKR |
| Commercial support pricing for the new components | Not specified |
The published license alone does not mean commercial support is also free. When reviewing a contract, the scope and cost of support should be checked separately. NVKR's license and contribution guidelines are available in the official repository.
Support terms in Korea and full release schedules have not been disclosed
Official sources do not specify availability by region within Korea or support terms in Korea. Full release schedules for NVKE and NVKR have not been provided either. Publicly available software therefore does not, by itself, guarantee support in Korea.
The exact time and time zone of the announcement have not been confirmed either. There is no basis for converting the date shown in the announcement to Korean time. The time a feed was detected must not be presented as the time of the company's announcement.
No independent measurement results for Snapshot from Artificial Analysis have been provided either. There is no basis for concluding that the organization has verified its performance. Company announcements and independent measurement results should be read separately.
Check operational readiness and support scope before adoption in Korea
Developers and operators in Korea should distinguish between evaluation and use in live services. The official NVKE documentation does not recommend use for production workloads. Support in Korea and costs cannot be determined from the public license alone.
- Check the preview status and potential for changes in the official NVKE documentation. Follow the company's warning that production use is not recommended.
- Check the terms of use in the NVKE and NVKR license documents.
- If considering commercial support, check the scope of support in Korea and pricing.
- Decide on a timeline for moving to production after official release guidance is published.