Getting a GPU-accelerated Python environment up and running with the right NVIDIA CUDA dependencies, the right framework versions, and the right security posture can take days of troubleshooting—time better spent on productive work. And once teams finally move from experimentation to production, they often find their environments aren’t reproducible, their models aren’t governed, and their compliance teams have serious questions.

Anaconda and NVIDIA have been working together to close that gap. By combining Anaconda’s trusted distribution and governance capabilities with the NVIDIA accelerated computing stack, we’re giving enterprise teams an expedited, secure path from their first line of code to a scaled production AI environment.

What’s New: Nemotron Models in AI Catalyst

AI Catalyst now includes NVIDIA Nemotron 2 and NVIDIA Nemotron 3 models, with quantizations.

For those unfamiliar, Nemotron is NVIDIA’s family of open source foundation models designed for building efficient, accurate, and specialized agentic AI systems. These models are among the most capable open-source options available for enterprise use cases. Now they’re available directly within AI Catalyst with the same governance controls that cover every other model in our catalog.

That means Nemotron models in AI Catalyst come with:

  • Comprehensive AI bill of materials to enable full transparency, governance and use case selection of runtime optimized models
  • Vulnerability scanning and compliance documentation: the same rigor Anaconda applies to Python packages, extended to foundation models
  • CUDA stack compatibility validation: ensuring models work seamlessly with your GPU environment, no matter where it is
  • Reproducibility controls: the environment that works in development is the one that deploys to production

The governance capabilities Anaconda has built for Python-based development don’t stop at the model boundary anymore. Whether you’re managing packages or foundation models, AI Catalyst gives your security, compliance, and infrastructure teams a single, consistent framework.

DGX Spark Support: From Desktop to Production

This expansion also includes official support for NVIDIA DGX Spark, the Grace Blackwell AI Supercomputer that fits on a desk, delivering up to one petaFLOP of AI performance with 128GB of unified memory. For developers who want to build and iterate locally before scaling to cloud or on-prem infrastructure, DGX Spark represents a compelling new option.

Anaconda provides the governed software environment that turns DGX Spark into a production-ready AI development environment. Developers can go from setup to training models in minutes, with pre-compiled, version-aligned CUDA binaries distributed through conda, eliminating the manual configuration and dependency conflicts that typically slow teams down.

See It Live at NVIDIA GTC

If you want to see what this looks like end-to-end, come find us at NVIDIA’s GTC conference March 16-19, 2026 in booth #3001. We’ll be demonstrating how a developer working from a local workstation like a DGX Spark, or from a cloud-based VPC, can spin up a secure, GPU-accelerated Python environment, load an AI/ML notebook, and deploy a Nemotron model for inference from AI Catalyst. No manual configuration. No dependency conflicts. Everything tracked, governed, and reproducible.

This is what AI development in the enterprise should look like: fast enough to keep up with the pace of innovation, governed enough to meet the demands of IT and compliance.

The Bigger Picture

GPU-accelerated AI is no longer a niche capability; it’s quickly becoming a baseline expectation for enterprise AI teams. By integrating the NVIDIA CUDA stack, DGX Spark support, and Nemotron models into the Anaconda platform, we’re ensuring that teams can take advantage of the best hardware and models available without sacrificing the security, reproducibility, and governance that production AI demands.

We’re just getting started. Stay tuned for more updates as we continue to expand AI Catalyst’s model catalog and deepen our NVIDIA integrations throughout 2026.

Want to learn more? Explore AI Catalyst or request a demo to see how Anaconda can help your team move from GPU-accelerated experimentation to governed AI production.