~$50M+ ARR run-rate · OSS Firebase
Postgres + Auth + Storage + Realtime + Edge Functions. The default backend for indie SaaS in 2026 — ShipFast, every Lovable export, half of v0 outputs land on Supabase.
Track Atlas · OPC ATLAS
Database / Serverless DB: 7 Capital-Heavy Plays and the Indie Sidestep
Supabase, Neon, Turso, Convex won this layer. The honest indie play is built on top of them — not against them.
The serverless / managed-database wave from 2020-2024 has consolidated into clear winners. Supabase is the biggest open-source Firebase alternative on the market with reported ARR in the eight-figure range and ~$2B valuation, last raise around $200M+. Neon was acquired by Databricks for ~$1B in 2025 after raising $104M and crossing ~1M projects. Turso pulled in $34M+ on libSQL and SQLite-at-the-edge. PlanetScale (~$1.05B valuation, MySQL/Vitess) is the enterprise tier; Convex raised $30M+ for reactive backend-as-a-service. The lesson for solo builders: this whole layer is capital-heavy infra. You don't build a new serverless Postgres unless you have $50M and a distributed-systems team. The actual indie opportunity is in the workflow layer ON TOP — schema migration tools, query optimizers, vector helpers, DX wrappers — selling shovels to the people who already standardized on Supabase, Neon, or Turso.
01 · 2026 Market Reality
Three sub-layers. (1) Hyperscalers and acquirers: AWS Aurora Serverless, Cloudflare D1, Vercel Postgres (powered by Neon under Databricks). They're the "default if your platform already runs there" option — distribution moat, not technology moat. (2) Independent winners: Supabase (open-source Firebase, $2B valuation, ~$50M+ ARR run-rate), Neon (acquired by Databricks ~$1B), Turso (libSQL on edge, $34M+ raised), Convex (reactive backend, $30M+ raised), PlanetScale (MySQL/Vitess, ~$1B). All defensible at this scale, but the gates closed at ~$25M Series A in 2023; new entrants now need a vertical edge (vector, time-series, embedded). (3) Vertical / niche databases: Pinecone (vector, raised $138M but lost momentum), Tigris (object + database hybrid), Xata (hybrid Postgres + search, raised $35M+). Three forces in 2026: AI workloads keep pulling vector + agent-memory specialization (this is where the new $10M-Series-A capital is going); embedded SQLite revival via Turso/libSQL is real but tiny TAM unless you bundle DX; and consolidation through 2024-2025 (Neon → Databricks, Heroku Postgres ossified, Render Postgres shrinking) means the open field for solos is the data-tooling layer, not the data-engine layer.
02 · Companies to Watch
~1M projects pre-acquisition · $104M raised
Serverless Postgres with instant branching. Acquired by Databricks May 2025 — became the Postgres infra for Vercel Postgres. The textbook strategic exit for the category.
libSQL fork of SQLite · edge replicas
SQLite-at-the-edge with multi-region replication. The clearest "capital + open source" bet that's NOT just another Postgres. Real adoption in mobile + edge runtimes.
Reactive backend-as-a-service
TypeScript-first reactive database + functions. Loved by AI app builders for tight loop reactivity. Smaller than Supabase, sharper opinion — winning niche dev mindshare.
MySQL/Vitess · enterprise tier
YouTube's database tech, productized. Killed the free tier in 2024 — caused massive migration to Neon/Supabase. Cautionary tale about pricing the indie ecosystem out.
SQLite at the edge, built into Workers
Free tier so generous it's a price-floor for indie data. Distribution moat: if your stack is already Workers, you'll never leave. Squeezes everyone above on price.
Vector DB · ~$100M valuation peak
Was the default vector DB through 2023. Lost ground as Postgres + pgvector + Supabase became "good enough" for 90% of indie AI apps. Specialization isn't always defensible.
26k+ GitHub stars · indie tooling
TypeScript-first ORM beating Prisma for indie devs. The cleanest example of the indie sidestep: don't build a database — sell developer ergonomics on top of one.
03 · Green Lights & Red Flags
🟢 Green light · Consider entering
You're building dev tools ON the platforms
Migration tools, schema visualizers, query optimizers, vector libs, RLS dashboards — every Supabase/Neon/Turso user needs these. Indie surface, big buyer pool.
You have a real vertical edge (vector, time-series, geo)
If you actually do PhD-grade distributed systems and know one vertical workload better than anyone, there's still a $30-50M Series A waiting. But you need the credentials.
You have a capital partner + 5+ year horizon
Supabase took 5 years to $50M ARR. Convex still pre-PMF after 4. This is not a 12-month side project — it's a decade-long infra bet.
🔴 Red flag · Hold off
You're an indie solo trying to compete with Supabase
Honest math: cloud bills are $5-50k/mo from day one, you need SOC 2 to land any real customer, and Supabase will hire your top 3 engineers. Don't fight here.
Your differentiation is "faster Postgres"
Postgres is fast enough. The fight is on DX, replication, edge, branching, billing — not raw QPS. If your pitch is "1.3x speedup," you've lost before slide three.
You think Cloudflare D1's free tier won't affect you
D1's pricing is the new floor. Any indie database tool needs to be either free + paid-pro, or solve a workload D1 explicitly doesn't — analytics, vector, multi-region writes.
04 · Three Ways In
Funded: vertical database
Distributed-systems team with academic or FAANG infra cred
Indie sidestep: data tooling
Senior backend dev who lives in Supabase/Neon daily
Advisor / DB angel
Ex-database company founder or eng director
05 · Founders Who Fit Best
The Infra Architect
If you've shipped a query planner or written WAL replication for a living, this is your category — but you'll be raising VC, not bootstrapping. Plan for a 7-year arc.
The One-Person Dev Team
Don't build a database. Build the tool every Supabase user needs and doesn't have. Drizzle is the template — open source + monetize on hosted features later.
The Product Speaks for Itself
If you can compound a Twitter audience around schema-design, performance war stories, and clean docs, you can find a way in via tooling — even without distributed-systems pedigree.
06 · What's Next
Worth reading
Communities
People to follow
- Paul Copplestone (Supabase) @kiwicopple
- Nikita Shamgunov (Neon) @nikitabase
- Glauber Costa (Turso) @glcst
Adjacent tracks
- SaaS BoilerplatesEvery boilerplate ships with Supabase/Drizzle by default. Indie data tools sell into the same ICP.
- AI Coding / DevToolsVector DB and agent-memory tooling overlaps heavily with the AI coding side.
- Eval & ObservabilityAdjacent infra play — same buyer (senior backend), same Series A gates, lighter capital.
Which kind of founder are you?
5 min · 12 questions · Free · Get your archetype + top 3 matching tracks
Take the quiz →