AI codegen favors Firebase defaults

The recommendation to move from Supabase to Firebase is really a recommendation to trade database flexibility for a safer default workflow. In this view, teams using AI to assemble apps often do better with Firestore because the app can be modeled as a few nested documents, auth is bundled, and the security model stays closer to the application layer, while Supabase asks users to keep a Postgres schema, row policies, and API exposure disciplined over time.

• Supabase puts a full Postgres database behind instant APIs, and its docs stress that any table exposed through the Data API needs Row Level Security enabled and maintained with policies. That is powerful, but it also means a weak schema or weak policy design can become the app’s core failure mode for non engineers.
  3 supabase 4 supabase 5 supabase
• Firebase was built around the simpler backend as a service idea, app code talks directly to managed services for database, auth, and storage. In the vibe coding wave, that simpler mental model has become newly attractive because many app builders are optimizing for getting a working product, not for designing normalized relational data models.
  2 sacra 6 google 7 sacra
• This does not mean Supabase is losing relevance. Supabase grew to an estimated $70M ARR by August 2025 as the default backend for many AI generated apps. The tension is that what makes it great for fast starts, Postgres plus bundled backend tools, can also create more cleanup work later than a more opinionated document store path.
  1 sacra 2 sacra 7 sacra

Going forward, backend winners in AI app creation will be the ones that make the safe path the default path. Firebase is strong where the app is mostly users, documents, and permissions. Supabase is strong where a team wants SQL, relational joins, and deeper database control. As AI builds more of the code, the product that requires less architectural judgment at day one should keep gaining share.