Ryan Henry came back from Y Combinator’s demo day and didn’t invest in a single company, which for him is abnormal. By the numbers the batch was the strongest in YC history, and that’s exactly what gave him pause. A million dollars in ARR is cheap to hit now, and fast, spiky revenue is not the same as a durable company. The SaaS spreadsheet that used to tell you whether something was venture scale is gone, and nothing has replaced it.

Ryan runs Sand Hill North, the venture arm of a family office. Half of this one is about what you bet on when you can’t model the business anymore (it keeps coming back to the person). The other half is the ground everyone’s building on: AI as private, rentable infrastructure that can be pulled overnight, the Fable and Mythos saga, and why “sovereign AI” is a harder question than it sounds.

In this episode:

• Why the “strongest YC batch ever” is a warning, not a green light

• A million in ARR is cheap now, and how that breaks the SaaS playbook

• Selling the work, not the software (and a uranium mine as a business model)

• Why the model is never the moat when you’re renting the foundation

• Sovereignty isn’t a hard drive: you can’t out-spend the frontier

Chapters:

0:00 Back from hiatus

1:20 The YC batch Ryan skipped

3:15 $175M rounds and the Carta data

7:44 “Strongest batch ever,” or just better selection?

10:51 The SaaS spreadsheet is dead

13:00 Sell the work, not the software

22:47 Uranium, and the king of Spain

24:40 Agent infrastructure and why trust is the moat

29:31 Results as a service (OpenPhone to Quo)

34:04 Fable, Mythos, and a model the government pulled

44:39 Sovereignty isn’t a hard drive

50:29 Alex Karp, Palantir, and the middle layer

Ryan Henry runs Sand Hill North: https://shn.vc

Read the companion post and subscribe: https://decelerator.media/

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.decelerator.media