Cash doesn't equal success: how Supabase rode three waves of AI to 10M developers
Key takeaways from YC Root Access interview with Supabase co-founder and CEO Paul CopplestoneDealroom has a profile for this one. Try Dealroom → (August 2026):
Scale. Supabase now has around 10 million developers, up from roughly 6.5 million at the start of 2026 — Copplestone estimates developer growth of 6–7x over the past 12 months, off an already-large base.
Third venture, right timing. Supabase is Copplestone's third venture-backed startup; his two earlier companies (both in Southeast Asia) were acquired. Founded in 2020 as the first fully remote YC batch, it started as open-source Postgres-as-a-service and now positions as an open-source Firebase alternative.
Betting on Postgres and open source. The founders deliberately chose 30-year-old Postgres — spotting a growing groundswell on Hacker News — and shipped everything under MIT, Apache 2 or Postgres licences rather than an open-core model, contributing back to tools like PG Vector (created by Andrew Kane) instead of building rivals.
Three chapters of AI. Copplestone frames Supabase's acceleration as three waves: vector databases / PG Vector and RAG (early 2023), the vibe-coding wave with Lovable and Bolt (late 2024), and Claude Code (Christmas 2025) — the third being the most transformative, lifting every metric from activation to conversion at once.
Agents are now the primary user. Over 60% of databases are launched by an agent (likely closer to 90%, since CLI-driven launches go uncounted), pushing Supabase to make MCP and CLI first-class citizens and to keep dashboard and code in lockstep via a git-friendly workflow.
Supabase for Platforms. A dedicated product lets other platforms — Lovable among them — build on top of Supabase, now serving over 50 (often large) customers, including enterprises adopting it internally.
Lean, distributed, AI-run ops. The company runs on roughly 360 people (up from 220 at the start of the year) across 60+ countries, fully written-down on Slack and Notion since inception; it feeds six years of company context into AI bizops, and runs a sales model where reps are paid on measured uplift over a control cohort.
Funding as a means, not the goal. Copplestone points to the $500M Series F at a $10.5B valuation as operating capital to bridge the gap between launching databases and customers paying — stressing YC's lesson that cash in the bank doesn't equal success, since better-funded competitors still lost.
Read more: Supabase
