Platform Bundling Threatens Convex

The real risk is distribution loss, not just feature competition. Convex grows when AI coding tools suggest a backend at the moment an app is first created, but those same platforms now have reasons to keep the database, auth, storage, and payments layers inside their own product so they can earn recurring revenue from every app that keeps running, not just one time app generation fees.

• This playbook already worked for Supabase. As Bolt.new, Lovable, and Replit scaled in 2024 and 2025, their code generation flows frequently pointed users to bundled backend services, helping push Supabase from about $13M ARR in 2023 to $70M by September 2025. That shows how much control the app creation layer has over backend demand.
  1 sacra 4 replit
• The incentive is economic. Backend services monetize compute, storage, bandwidth, MAUs, and seats over the life of an app, which is usually stickier than charging for initial creation. Bolt Cloud added native backend infrastructure in August 2025, and Replit documents now distinguish between legacy Neon databases and newer Replit managed databases inside the product.
  1 sacra 2 sacra 5 bolt
• Convex is especially exposed because its product is optimized for frontend developers who want a ready made reactive backend, the same user the AI app builders already own. If Bolt, Lovable, or Replit make their default database good enough, Convex has to win through superior developer experience after the platform has already steered the user elsewhere.
  2 sacra 1 sacra 3 sacra

The next phase of this market is vertical integration. AI coding platforms will keep pulling more backend primitives into one workflow, and independent backends will need to become the best execution layer behind those platforms, or the obvious upgrade path once apps outgrow the bundled default.