Here’s the common reality about how AI gets adopted today. A team picks a model, an agent, they build fast, and weeks or months later someone asks: do we know how this was trained, and can we prove it to a regulator, a customer, or our own board? All too often the answer is no.
That’s the true fight underneath the Open Weights and American AI Leadership letter. Anaconda joins Microsoft and over 230 organizations with our full support of this message. Open weights are the difference between building on a foundation you can inspect, and renting capability you can never fully see inside of. Closed models ask the market to trust first and verify never. Open weights models let you verify so you can trust from day 1. In a world where AI agents are now running in banks, hospitals, and government agencies, that’s a lot of unknown risk in some of the most regulated markets.
It also decides who gets to compete. A closed default favors whichever handful of frontier labs get there first. Open weight models are what let AI leadership come from an entire ecosystem instead of three or four companies picking winners for everyone else.
But there’s more to it. Openness by itself does not make you secure. An open model nobody can benchmark, red team, or govern is a different kind of exposure. The real promise of open weights only gets delivered when someone builds the layer that makes openness secure and safe to run at enterprise scale.
That’s been Anaconda’s work for over a decade. Long before AI, we built the trusted open source foundation that over 52 million AI, ML, and Python developers rely on every day. Today we’re doing the same thing for AI-native development. Through our recent Kilo Code acquisition, we give enterprises an open, agent engineering layer that routes across 500+ models with no single-provider lock-in. And across the Anaconda Platform, we pair that openness with the trust and governance layer enterprises need: verified packages, security scanning, and controls that turn compliance frameworks into something enforceable instead of paperwork.
When we say we support this letter, we are not signing on as a sympathetic voice from the sidelines. We are the part of this story that shows how it works in practice. Openness gives everyone a foundation to build on instead of a toll to pay. Anaconda is the foundation that makes that foundation safe to stand on.
We are proud to add our support to this group of companies and AI leaders building this industry. The case this letter makes in Washington is the same one we have been building in the market: the organizations creating real value with AI should get to keep it, verify it, and govern it, all on their own terms.