Replit is a company that offers free in-browser REPLs (read-eval-print loop) in a variety of different languages. Some of the best software businesses give some primitive away for free (storage, compute, email, stock trades, or even source code) in exchange for aggregating user demand and capturing value differently.
But who is Replit for, and how can Replit capture value from giving away free compute?
Bottoms-up demand generation for enterprise sales. There's a tier of free and hobby applications that move from provider to provider – first, they were hosted on Heroku; now, it's Vercel or a next-generation PaaS. I'd say that most of these applications will never upgrade to an enterprise plan – they are just taking advantage of the free hosting. But it causes developers to learn these systems and possibly use that knowledge in their day jobs.
It's difficult to build a runtime that's user-friendly and enterprise ready. The trap is that companies inevitably have to start with a simpler product, and simpler means more user-friendly. However, as these enterprise SaaS companies need to move upmarket, they find that they have to build out boring but required enterprise features – many of which introduce developer friction.
Education tool. After testing a few markets, Replit found some success in the education market – programming environments for kids (and, on the flip side, interview environments for adults). It's easy to use, and there are tons of kids running Minecraft servers on it – what better signal that you're product is easy to use if kids are using it? Of course, you can see how school administrators might take pause if the site is being used for both education and games (some schools banned the domain). Unclear how big this market is.
Long-tail of creator developers. I wrote about this at the height of the bull market last year in the phenomenon of MicroSaaS. SaaS is becoming commoditized, but it's possible that you could build a lifestyle business on Replit – only knowing the very basics of coding. This would put Replit against companies in the low-code/no-code space (which seems to be correlated with good/bad markets). I'm not sure how much gets captured by specialized platforms (e.g., Shopify, WordPress).
I don't think these are the most compelling business models, but I think a company that figures out how to give away free compute in exchange for demand has a chance of being a generational company in the next cycle. Even with a slowdown in Moore's law, skate where the puck is going.