First Pour.
From besties to frenemies? Enterprise companies and frontier AI labs might be entering their frenemy era.
There’s a tension brewing that I think we need to talk about: frontier AI labs vs. enterprise companies.
Might they actually be competing?
[Prefer to watch? My full take is below.]
Let me explain with Anthropic as an example.
Figma and Anthropic were deep partners. So much so that Anthropic CPO Mike Krieger joined Figma’s board.
Then Anthropic launched Claude Design. A direct competitor.
Cursor helped make Claude the coding model of choice. Then came Claude Code.
It’s clear that frontier labs are interested in moving up the stack—commoditizing the very companies they’re selling to.
No wonder open-weight AI is suddenly looking much more attractive.
Microsoft, NVIDIA and hundreds of other organizations recently signed an open letter supporting open-weight models. Yes, it’s about American innovation—but let’s not ignore the incentives.
Open weights give enterprises more control and reduce their dependence on frontier labs. Why should they open the corporate kimono when frontier labs won’t do the same?
Frontier labs may need to reevaluate their strategy—especially if their IPOs are counting on those enterprise dollars.
Something to brew on.
