Pure open source AI.
Fully self-governed.
Build, deploy, and manage your AI lifecycle with Canonical. From AI infrastructure to production-ready model serving, LLMOps, and edge AI – all on Ubuntu, the open source platform that puts you in control of your data and compliance.
Your AI, your terms
- Run your models on-premise, in your private cloud, or across public clouds
- Total control over your digital infrastructure, software, and data
- Maintain full governance, strict compliance, and the long-term freedom to scale your architecture on your terms
Enterprise-grade AI stack
Powerful tools for every stage of your AI journey, from raw silicon to the edge.
AI infrastructure
A solid foundation for computing. Run AI workloads smoothly using Canonical Ubuntu, optimized specifically for all leading hardware. Scale efficiently with Canonical Kubernetes and robust private cloud automation - including bare metal management with Canonical MAAS.
Model serving & MLOps
Move from experimentation to production fast. Build, train, and track models reliably using Charmed Kubeflow for end-to-end lifecycle management. Serve large language models (LLMs) using vLLM. Scale GPU accelerated model serving using LLM-d.
Edge AI
Bring intelligence directly to your hardware. Deploy compact, efficiently packaged AI models on remote devices, factory floors, or smart machines. Keep your edge operations safe with snap sandboxing, automatic, transactional over-the-air updates, and robust offline stability.
Why enterprise leaders choose Canonical
- True open source: Built on public, auditable code that guarantees flexibility and community innovation.
- Enterprise security: Backed by regular security updates, automated patching, and strict compliance standards.
- Global support: Benefit from round-the-clock expert support for your entire infrastructure stack.
Built with Canonical AI
Turn raw technology into intelligent applications. Here is what you can build with our open-source stack:
Retrieval-augmented generation (RAG)
Connect open source LLMs safely to private databases. Deliver answers without leaking sensitive corporate data to external APIs.
Agentic workflows
Create smart, multi-step systems where AI agents make decisions, plan tasks, and execute complex workflows without constant human supervision.
Coding agents
Speed up software development safely. Deploy autonomous code assistants within your private network to write, test, and debug code securely.
Autonomous contact centers
Power the next generation of customer support. Build highly reliable, human-like voice and chat agents that handle customer issues instantly while keeping all records strictly private.
Security – never an afterthought
Confidential AI protects your models and data
Your models and sensitive data need to be secured at every stage of the ML lifecycle. With confidential AI on Ubuntu, you can protect your data and model at run-time with a hardware-rooted execution environment. Strengthen your compliance posture and safely fine-tune your model with your enterprise data.
Fast-track compliance with trusted AI software
Rapidly access all the open source software you need for AI, and run it with peace of mind. We maintain and support all of the key open source software in the AI lifecycle so you don’t have to. Leave the CVE fixes to us, so you can fast-track compliance and focus on your AI models.
What our customers say
“We needed a cloud solution that was stable, reliable and performant. Canonical allowed us to do this by helping to design and deploy our cloud - and they helped us do this quickly.”
Peter Blain
Director of Product and AI, Firmus
“Partnering with Canonical lets us concentrate on our core business. Our data scientists can focus on data manipulation and model training rather than managing infrastructure.”
Machine Learning Engineer
Consumer entertainment company
All the open source you need for AI
Learn more about enterprise AI solutions
-
What are large language models (LLMs)?
LLMs and generative AI are dominating much of the current AI/ML discourse, and their potential goes far beyond chatbots. Our blog breaks down LLM use cases, challenges and best practices.
-
GenAI with vector databases and RAG
Go deeper on GenAI in this webinar explaining how to enhance your model outputs with RAG.
-
Check out how we enable open source in the world's leading silicon
Explore how Canonical partners with silicon vendors to optimize our solutions with certified hardware.
-
MicroK8s with Charmed Kubeflow on NVIDIA EGX platform
Download the reference architecture for a detailed guide to infrastructure for machine learning use cases.