Numerically Speaking
This month: why your AI strategy doesn’t touch your enterprise apps, a commissioned report on governing open source risk, and why AI labs are buying the tooling layer.
Featured
Why can’t your organization actually execute its AI strategy?
Odds are, you’ve already written a framework for how your team should use AI. But policies that can’t be enforced at runtime don’t govern anything.
To build a strategy that actually works, you need to impose technical controls before AI reaches your developers’ environment.
Analyst Report
The 3 things you need to govern open-source AI at scale
AI depends on open source software. Does your organization have the controls in place to use it safely?
In this commissioned report, Omdia principal analyst Mark Beccue reveals what enterprise governance of open source AI actually requires.
Trending
AI labs are buying your devs' toolchain
OpenAI last week announced it will acquire Astral, the startup behind uv, Ruff, and ty. The move comes just three months after Anthropic’s acquisition of Bun, the JavaScript runtime that ships Claude Code.
“OpenAI will continue supporting our open source tools after the deal closes,” writes Charlie Marsh, founder and CEO of Astral, in a blog post. “We’ll keep building in the open, alongside our community – and for the broader Python ecosystem – just as we have from the start.”
Free Preview
Build AI agents that ship—now with 3 new chapters
Our most popular e-book of 2026 just got even bigger.
In this expanded preview of AI Agents in Action, 2nd Edition, author Micheal Lanham adds chapters on memory and RAG, agent evaluation and feedback, and production deployment.
In the News
Can I pay you in tokens?
At NVIDIA GTC 2026, CEO Jensen Huang floated a new compensation model: Giving engineers an AI token budget on top of their base salary as a productivity multiplier.
“It is now one of the recruiting tools in Silicon Valley: ‘How many tokens comes along with my job?’” Huang says. “And the reason for that is very clear: Because every engineer that has access to tokens will be more productive.”
Available Now
Nemotron models now available in AI Catalyst
Anaconda’s latest release adds NVIDIA’s Nemotron foundation models to AI Catalyst’s curated model catalog. Two other updates ship with it:
- Quantifiable security scoring
- A full audit trail for every package decision in Anaconda Core
Connect with Anaconda
Meet us at PyCon DE and PyData 2026!
Anaconda is sponsoring two of the biggest Python conferences this spring:
📍 PyCon DE + PyData 2026
We’ll be in Darmstadt, Germany, from April 14-16 for Europe’s largest Python, AI & software engineering conference. Visit our booth and catch our sessions:
📍 PyData Texas 2026
Texas’s largest gathering of Python developers is taking place April 17-19 in Austin, and Anaconda is a platinum sponsor. If you’re planning to attend, don’t miss the keynote session with Dawn Wages!
Learn with us
In Conversation: AI Essentials for Tech Executives
Hamel Husain, Greg Ceccarelli, and Jess Haberman outline the frameworks you need to turn AI ambition into a working strategy.
Remove the #1 Blocker to GPU Adoption
Torsten Volk and Alan McCarty tackle the biggest hidden obstacle slowing AI adoption—and how to eliminate it for good.
Vetting AI Models: From DIY to Enterprise
Join us Thursday, March 26, at 1PM ET for practical strategies to vet AI models for enterprise deployment.