Hacker News Reader

Comment by candiddevmike | Hacker News Reader

Reading the tea leaves, Series D means they opted for more funding vs IPO. They claim to have 2 million users, but they're open core so how many are paying? Maybe their books aren't looking that great. Wall street doesn't understand database vendors outside of "big data", so they're probably hoping for acquisition. Not sure who would buy them though, as PostgreSQL vendors are kind of a dime-a-dozen these days...

View on HN

If lovable, bolt.new, etc kept integrating with them, that's a money maker without needing to do much sales. There's a wave of AI tools that require somehow save state and Supabase provides that. I'm absolutely amazed others haven't jumped in the same ship yet.

That definitely seems to be the play. Keep funneling in users from Lovable/bolt.new and keep building revenue or hope to be acquired if one of those vibe coding tools gets huge.

In the short term yes, but why would you as an end user of lovable and similar tools prefer expensive Supabase? If you already have an AI developer at your disposal you might as well make it figure out how to properly set up and maintain AWS, right? Maybe it can't do it now -- too complicated because there's no simple AI-accessible interface, and LLM models are simply not smart enough yet, but I imagine it will change soon.

> Not sure who would buy them though, as PostgreSQL vendors are kind of a dime-a-dozen these days...

Supabase defo has a much higher mindshare.

Sure but ultimately they're still just selling something that is already free and wrapping AWS. These business models aren't sustainable unless you trash your free product, which also isn't sustainable. Presumably they have a good deal with their cloud vendor, AWS I think, but I think its safe to assume they lose A LOT on their free products.

The whole premise of cloud hosting businesses is that people don't want to manage stuff themselves.

They end up managing stuff themselves anyway. Plus managing another kind of bills.

AWS seems to be doing fine.

> so how many are paying

This is like if Google Spanner were open sourced tomorrow morning: realistically how many people are going to learn how to deploy a thing that was built by Google for Google to serve an ultra-specific persona?

Maybe you might get some Amazon-sized whale peeking at it for bits to improve their own product, but the entire value prop is that it's a managed service: you're probably going to continue paying for it to be managed for you.

IMO it also depends on how the whole process is tied together.

I always loved Vercel for their easy hosting of Next.js with included CI/CD, but I recently switched to self-hosting - their pricing switched from a flat, worry-free $20/month to an unpredictable whatever-it-may-cost plus it sent me 10+ emails every single month about hitting some quotas that they introduced and I couldn't find a good way to stop that.

A lot are paying, including me for multiple projects. They have a pretty good offering. I used to use them for dev and prod, but now using neon for dev. Supabase still for prod. I had switched from mongo to supabase. I may switch to neon for prod but not in a rush.

They also offer so much more than just postgres. Though I use them only for postgres myself.

Since you use both supabase and neon, any particular strength or weakness to keep supabase for prod? I just moved my app to neon today (easy enough to test it!) and am enjoying the auto-scaling features and UI is great on neon. But I'm curious about how supabase stacks up.

Supabase feels less flexible. Also it tries to do many other things so I don't think they can focus fully on the db side. However it still works well enough for production so cannot complain too much. I haven't done benchmarks for performance latency etc though. I should!

A lot of people don't self-host it, even though it is open core. This is due to their docs being garbage and tons of differences between the offerings, so you can't even rely on the main docs if you're self-hosting.

It's easier to just become familiar with a DB UI tool like Beekeeper or DataGrip and spin up your own things. I'm also not a huge fan of being "locked-in" to so many things (including their auth). I think most projects would be better off keeping these parts separated, even if they are using third-party services to handle them, as it would be way less overhead to migrate out.

it's an aggressive preemptive round, so i'd guess 2b/50 = 40M of revenue. Probably low margins since the free tier/ hosting postgres nature of the business.

Reply on news.ycombinator.com