1. We took a hard-problem, peer-to-peer networking and IdM, and (mostly) solved it.
2. We're hoping this will drive people to build apps that leverage the unique capabilities of authenticated p2p mesh networks. It doesn't even have to be specifically for Tailscale.
3. People will want to use those apps and (if we're good at our jobs) choose to pay us to run the network for them over our competitors or building something in-house.
4. $$$
I'm not sure I would say this is as nefarious as the tone of the comments here suggest. Wanting a "killer app" for your software platform is pretty normal which is really what he's talking about. I would be nervous declaring victory or an inevitability without being to name what that killer app actually is but trying to figure it out/build it is a good strategy. It's one of those times where the desire for engineers to solve their pet-problems, play with shiny new toys, and build Halo LAN Party over Tailscale is aligned with the business.