--- title: Ubuntu Pro is now available on AWS Deep Learning Amazon Machine Images (DLAMI) description: In September 2026, Amazon Web Services (AWS) published its first DLAMIs built on Ubuntu Pro in the Base GPU family. By choosing Ubuntu Pro, enterprises launching DLAMIs gain a stronger security maintenance posture at launch thanks to Expanded Security Maintenance (ESM). What are the DLAMIs? The DLAMIs are AWS-maintained machine images for […] url: https://canonical.com/blog/ubuntu-pro-available-on-dlami --- 1. [Blog](https://canonical.com/blog) 2. Article --- [Carlos Bravo](https://canonical.com/blog/author/carlos-bravo "More about Carlos Bravo") 5 November 2026 # Ubuntu Pro is now available on AWS Deep Learning Amazon Machine Images (DLAMI) [AI](https://canonical.com/blog/tag/ai) [AI/ML Infrastructure](https://canonical.com/blog/tag/ai-ml-infrastructure) [AWS](https://canonical.com/blog/tag/aws) [AWS EC2](https://canonical.com/blog/tag/aws-ec2) [cloud](https://canonical.com/blog/tag/cloud) [Public cloud](https://canonical.com/blog/tag/public-cloud) --- Share the article In September 2026, Amazon Web Services (AWS) published its first DLAMIs built on Ubuntu Pro in the Base GPU family. By choosing Ubuntu Pro, enterprises launching DLAMIs gain a stronger security maintenance posture at launch thanks to Expanded Security Maintenance (ESM). ## What are the DLAMIs? The DLAMIs are AWS-maintained machine images for GPU and accelerator instances. Each ships the NVIDIA driver, CUDA, the Elastic Fabric Adapter (EFA) networking stack, and the surrounding tooling already installed and tested together. This means a GPU node is productive on first boot, rather than after a day of driver work. These images cover 31 actively built lines: Base AMIs carrying the drivers alone, framework images with PyTorch preinstalled, and Neuron images for Trainium and Inferentia. ## What is included in Ubuntu Pro DLAMI images? Ubuntu Pro DLAMI is available as a base GPU image for typical AI workloads on the latest GPU hardware. There are four versions of the image split across 2 Ubuntu releases and 2 architectures. This enables a broad support of instance types, including Grace Blackwell architecture. It ships with NVIDIA driver 595.91.07 and several CUDA toolkits. | **Image** | **Architecture** | **Kernel** | **CUDA** | | --- | --- | --- | --- | | Base GPU AMI (Ubuntu Pro 24.04 Long Term Supported (LTS)) | x86\_64 | 7.0.0-1013-aws | 12.8, 12.9, 13.0, 13.2 (13.2 default) | | Base GPU AMI (Ubuntu Pro 22.04 LTS) | x86\_64 | 6.8.0-1064-aws | 12.8, 12.9, 13.0, 13.2 (13.2 default) | | ARM64 Base GPU AMI (Ubuntu Pro 24.04 LTS) | ARM64 | 7.0.0-1013-aws | 12.8, 12.9, 13.0, 13.2 (13.2 default) | | ARM64 Base GPU AMI (Ubuntu Pro 22.04 LTS) | ARM64 | 6.8.0-1064-aws | 12.8, 12.9, 13.0, 13.2 (13.2 default) | The x86 images cover 14 GPU families, from G4dn and G5 through P4d, P5, and P5en, to the Blackwell P6-B200 and P6-B300. The ARM images cover G5g and the Grace Blackwell P6e-GB200 and P6e-GB300 instance families. ## Enterprise grade images with Ubuntu Pro GPU images have been preconfigured with a pinned kernel for stability and compatibility purposes. The NVIDIA and EFA drivers are compiled against it, so AWS holds the kernel packages and ships security fixes as a new image every week or two. [**Ubuntu Pro**](https://ubuntu.com/pro) **comes with ESM** by default, which expands security updates from the Main repository (the primary Ubuntu repository, maintained by Canonical’s engineers) to Universe (the community-maintained repository of Ubuntu) and extends the overall security coverage from 5 to 10 years. This solves a major compliance challenge for enterprises that cannot rely solely on community support. The Universe repository contains many key AI/ML dependencies, which the default DLAMI already has included at launch. It also enables [**Kernel Livepatch,**](https://ubuntu.com/security/livepatch) delivering in-memory security updates for the active kernel to prevent unexpected reboots. This is especially important for GPU accelerated machines, which are typically dedicated to continuous, long-running operations. ## Supporting NVIDIA Grace Blackwell P6e-GB200 and P6e-GB300 instance types For the first time on AWS, Ubuntu Pro DLAMIs provide out-of-the-box NVIDIA Grace Blackwell support across its two variants: GB200 and GB300.  While Ubuntu Pro 24.04 LTS supports both P6e-GB200 and P6e-GB300, Ubuntu Pro 22.04 LTS supports only P6e-GB200. Kernel livepatch on ARM64 is available from Ubuntu 26.04 LTS onwards. ## What ESM covers on the image If you’re curious to know whether the Ubuntu Pro variant is for you, you can either run `pro security-status --esm-apps` on the Ubuntu Pro 24.04 LTS Base GPU image, or continue reading. The Ubuntu Pro DLAMI base image comes with 38 pre-installed packages from the Universe repository. These packages are covered by Ubuntu Pro from launch, and include key packages commonly used on AI workloads at scale: | **What it is** | **Packages covered by ESM** | | --- | --- | | MPI and RDMA fabric | libopenmpi3t64 libfabric1 libpmix2t64 libucx0 libpsm2-2 libpsm-infinipath1 libhwloc15 libhwloc-plugins libibdm1 libopensm9 libosmcomp5 libosmvendor5 ibutils libmunge2 | | GPU and accelerator runtime | libamdhip64-5 libamd-comgr2 libhsa-runtime64-1 libhsakmt1 ocl-icd-libopencl1 | | Python packaging | python3-pip python3-pip-whl python3-wheel python3-venv python3.12-venv python3-setuptools-whl | | Everything else | apt-transport-https check environment-modules fuse libfuse2t64 libsnmp-dev libsnmp-perl libnetsnmptrapd40t64 libsubunit0 libsubunit-dev pahole pigz | However, these packages are only the starting point. With every deployment built, more packages and dependencies will be drawn in, the majority of which often come from the Universe repository. These packages receive support only through a Canonical Pro subscription, through ESM. ## How to get the images Each image line has a public AWS Systems Manager (SSM) parameter that resolves to its latest build. You can get the images by running the following commands with AWS CLI. For AMD64 images: ``` aws ssm get-parameters --names /aws/service/deeplearning/ami/x86_64/base-oss-nvidia-driver-gpu-ubuntu-pro-24.04/latest/ami-id ``` ``` aws ssm get-parameters --names /aws/service/deeplearning/ami/arm64/base-oss-nvidia-driver-gpu-ubuntu-pro-24.04/latest/ami-id ``` For ARM64 (Graviton and NVIDIA Grace Blackwell): ``` aws ssm get-parameters --names /aws/service/deeplearning/ami/arm64/base-oss-nvidia-driver-gpu-ubuntu-pro-24.04/latest/ami-id ``` If you want to check whether your Ubuntu Pro subscription is active, run pro status on a live instance to see your active entitlements and configurations. ## Conclusion and Next Steps Bringing Ubuntu Pro to the AWS DLAMIs provides enterprises with an ESM posture right out of the box. With Canonical’s commitment to support and maintain Universe packages for 10 years, and Kernel Livepatch enabling seamless patching without reboots, organizations can confidently run AI workloads while meeting stringent security and compliance standards. Ready to enhance your GPU workloads? Get started today by launching the new Ubuntu Pro DLAMIs on AMD64 images or ARM64, or visit the [AWS Deep Learning AMI Documentation](https://docs.aws.amazon.com/dlami/) for detailed setup guides and deployment options. If you have any questions, [contact us here](https://ubuntu.com/aws#get-in-touch). Learn more: * [AWS Deep Learning Base GPU AMI (Ubuntu Pro 24.04 LTS)](https://docs.aws.amazon.com/dlami/latest/devguide/aws-deep-learning-x86-base-gpu-ami-ubuntu-pro-24-04.html) – AMD64 * [AWS Deep Learning ARM64 Base GPU AMI (Ubuntu Pro 24.04 LTS)](https://docs.aws.amazon.com/dlami/latest/devguide/aws-deep-learning-arm64-base-gpu-ami-ubuntu-pro-24-04.html) – Graviton and Grace Blackwell * [AWS Deep Learning AMI Documentation](https://docs.aws.amazon.com/dlami/) * [DLAMI Support Policy Table – AWS DLAMIs](https://docs.aws.amazon.com/dlami/latest/devguide/dlami-support-policy-table.html) * [Supported Instance types, packages installed and changes for AMD64 Base Image](https://docs.aws.amazon.com/dlami/latest/devguide/aws-deep-learning-ami-gpubaseoss-ul2404pro-2026-09-24.html) * [Supported Instance types, packages installed and changes for ARM64 Base Image](https://docs.aws.amazon.com/dlami/latest/devguide/aws-deep-learning-ami-gpubaseossarm64-ul2404pro-2026-09-24.html) ## Sign up for our newsletter Get the latest Canonical news and updates in your inbox. Work email: \*I agree to receive information about Canonical's products and services. In submitting this form, I confirm that I have read and agree to [Canonical's Privacy Notice](https://canonical.com/legal/data-privacy/contact) and [Privacy Policy](https://canonical.com/legal/data-privacy). Sign up ## Share on --- ## Related posts [### Beyond the 10-year mark: Extending Ubuntu Pro 16.04 LTS security coverage](https://canonical.com/blog/extending-ubuntu-pro-16-04-coverage) A decade ago, Canonical launched Ubuntu 16.04 LTS (codenamed “Xenial Xerus”). As a Long-Term Support (LTS) release, it comes with 5 years of standard security coverage, which... [Carlos Bravo](https://canonical.com/blog/author/carlos-bravo) 17 September 2026 [### AI at the edge: simplifying infrastructure with Cisco and Canonical](https://canonical.com/blog/ai-at-the-edge-simplifying-infrastructure-with-cisco-and-canonical) Legacy infrastructure was not designed for the requirements of the AI era. While large-scale model training remains centralized in data centers, test-time inference is rapidly... [Pedro Lazzarotto](https://canonical.com/blog/author/pedro-lazzarotto) 11 June 2026 [### A new way to upgrade to Ubuntu Pro on AWS](https://canonical.com/blog/upgrade-ubuntu-lts-instances-to-ubuntu-pro-in-aws) Upgrading existing Ubuntu LTS instances to Ubuntu Pro using AWS License Manager just got better with a new destination license model. We are happy to announce a new way to... [Carlos Bravo](https://canonical.com/blog/author/carlos-bravo) 23 October 2026 [### Securing open source software dependencies in the public cloud](https://canonical.com/blog/securing-open-source-software-dependencies-in-the-public-cloud) Building stable and secure software requires understanding build systems and having a plan for vulnerabilities in your software dependencies. [Aaron Whitehouse](https://canonical.com/blog/author/aaron-whitehouse) 5 October 2026