How Supabase Grew to Nearly 10 Million Developers

Supabase started in 2020 with a simple idea:

Give developers the ease of Firebase, but build it on Postgres and make it open source.

By June 2026, Supabase said nearly 10 million developers were using the platform.

Its database launches had grown 600% year over year, and more than 60% of new databases were being launched through AI tools.

Supabase did not get there through one growth hack.

Its growth came in waves.

First, it positioned itself as "the open-source Firebase alternative."

Then it made Postgres easy enough for anyone to start using.

Open source helped Supabase spread through GitHub and developer communities.

Launch Weeks turned product releases into recurring marketing events.

Technical content helped developers find Supabase when they were trying to solve a problem.

Then AI coding tools created an even bigger distribution channel.

Products like Lovable and Bolt could create Supabase databases while building apps for their users. Tools like Cursor and Claude Code could interact directly with Supabase.

Supabase no longer had to convince every developer to choose it manually.

The software building the app could choose Supabase for them.

That is the main story behind Supabase's growth.

Supabase growth timeline

Year	Milestone
2020	Supabase launched publicly
May 2020	Hacker News brought 30,000 visitors and 1,400+ signups in a week
2021	40,000+ developers and 50,000+ hosted databases
2022	110,000+ developers and 150,000+ hosted databases
2024	More than 1 million managed databases
2025	AI builders became a major source of new Supabase projects
April 2026	Supabase passed 100,000 GitHub stars and reported 8 million developers
June 2026	Supabase reported nearly 10 million developers
June 2026	Database launches were up 600% year over year
June 2026	More than 60% of new databases were launched through AI tools

The numbers above are company-reported unless otherwise noted.

Supabase has not described its nearly 10 million developers as monthly active developers.

1. Supabase found positioning developers understood immediately

Paul Copplestone and Ant Wilson started working on Supabase in 2020.

The early product was built around Postgres and realtime functionality.

At first, the team described what it was building as:

"Realtime Postgres."

It was technically correct.

But it was not a great pitch.

It explained the technology without telling developers why they should care.

Shortly before Supabase's first big Hacker News moment, the description changed to:

"The open-source Firebase alternative."

That was much stronger.

Firebase had already taught developers why a backend-as-a-service product was useful.

It made it easy to add things like databases, authentication, storage, and backend services without managing everything yourself.

But Firebase also came with tradeoffs.

Developers could worry about:

vendor lock-in
proprietary databases
unpredictable costs
migrating later if an application grew

Supabase offered a simple alternative:

What if you got a Firebase-like developer experience, but with Postgres and open source?

Developers already knew Firebase.

They already knew Postgres.

Supabase connected the two ideas.

That meant another developer could understand the product in a few seconds.

The positioning also helped Supabase decide what to build.

If Supabase wanted to become a Firebase alternative, a database and realtime updates were not enough.

It would eventually need authentication, storage, APIs, functions, permissions, and a good developer experience.

The comparison became part of the roadmap.

2. Hacker News gave Supabase its first big growth spike

Supabase did not plan its first major public launch.

An early GitHub follower found the project and submitted it to Hacker News in May 2020.

The post took off.

It stayed on the Hacker News front page for more than a day.

According to Supabase's alpha launch postmortem, the following week brought:

30,000 new website visitors
more than 1,400 signups
rapid growth in GitHub stars
more than 1,000 databases

Supabase said the number of databases it was hosting increased around 10x overnight.

The infrastructure was not ready.

The team ran into limits with DigitalOcean and Cloudflare and eventually moved roughly 1,800 servers to AWS while new users were still arriving.

But the launch did something more important than generate traffic.

It gave Supabase thousands of developers to learn from.

One request kept appearing:

authentication.

Supabase spent the following months building Auth using PostgreSQL, PostgREST, and GoTrue.

When Auth launched, Supabase said its acquisition, activation, and retention improved.

Hacker News had created two growth loops at once.

It brought new users.

And those users told Supabase what to build next.

3. Supabase made Postgres easy to start with

Postgres was already one of the world's most popular databases.

Supabase did not need to convince developers that Postgres was useful.

It needed to make Postgres easier to start with.

Building a modern application usually needs more than a database.

A developer might also need:

authentication
APIs
storage
permissions
realtime updates
server-side functions
connection management
a dashboard

Firebase bundled many of these pieces together.

Supabase wanted to offer a similar experience while keeping Postgres underneath.

A developer could create a Supabase project and get a Postgres database plus the services around it.

That gave Supabase an important advantage.

Starting was easy.

But the underlying database was still Postgres.

It was not a database format that only Supabase understood.

If the project became important, developers could still use SQL, Postgres tooling, migrations, and the wider Postgres ecosystem.

That reduced the fear of making the wrong infrastructure choice early.

Supabase eventually captured the idea in a simple line:

"Build in a weekend. Scale to millions."

The product needed to answer two questions.

Can I start quickly?

And:

If this works, will I regret choosing it?

Using Postgres helped Supabase answer both.

4. A free tier made experimenting cheap

Most developers starting a project do not know whether it will become a business.

It might become a company.

It might stay a side project.

It might be abandoned two days later.

Charging before the project has any value creates friction.

Supabase originally considered a paid-only product.

It changed direction.

Its first Launch Week introduced a free tier alongside a paid Pro plan.

Now a developer could create a real database without making a financial commitment first.

That created a simple growth path:

start for free → build something → use more Supabase products → application grows → upgrade

Supabase has said many of its largest databases started on its free plan.

The useful activation event was not creating an account.

It was creating a database.

An empty Supabase account has almost no value.

A running database can slowly become part of an application's infrastructure.

The application might begin using Auth.

Then Storage.

Then Functions.

Then more compute.

By the time the project needs a paid plan, Supabase is already useful.

The free plan did not only generate leads.

It gave projects time to become valuable.

5. Open source became a distribution channel

Supabase was open source from the beginning.

That helped the company in several ways.

Developers could inspect its code.

They could self-host parts of the stack.

They could contribute fixes.

They could create SDKs.

And they could discover Supabase through GitHub.

Supabase also built around existing open-source software instead of creating every layer itself.

Its stack included projects such as:

Postgres
PostgREST
GoTrue
Phoenix
Elixir
pgvector
Deno

Supabase hired maintainers, sponsored projects, and contributed to the ecosystems it depended on.

The company sometimes describes this strategy as being a "community of communities."

This created distribution outside Supabase's own audience.

Someone searching for Postgres tooling, Next.js authentication, Flutter integrations, vector search, or another related problem could eventually discover Supabase.

The developer community grew quickly.

When Supabase announced its Series A in 2021, it reported:

40,000+ developers
50,000+ hosted databases
roughly 19,000 GitHub stars

By August 2022, it reported:

110,000+ developers
150,000+ hosted databases
roughly 36,000 GitHub stars

In April 2026, Supabase crossed 100,000 GitHub stars.

Supabase itself has called GitHub stars a vanity metric.

They do not tell you how many people are paying.

They do not tell you how many production databases exist.

But they do show attention inside the developer community.

The real value came when that attention turned into hosted projects.

6. Supabase turned launches into a repeatable growth system

One of Supabase's best-known growth tactics is Launch Week.

The idea came from Y Combinator.

During YC, startups have a fixed deadline: Demo Day.

Supabase liked how the deadline forced the team to ship.

After YC ended, the founders tried to recreate that pressure internally.

Every few months they would choose a date and work toward a big launch.

The idea was:

fixed timeline, flexible scope.

If something was not ready, the date did not move.

The feature moved.

Eventually, a single launch became a week of launches.

The first Supabase Launch Week ran from March 29 to April 2, 2021.

The team announced products and updates including:

pricing
Storage
the Supabase CLI
UI components
connection pooling
development workflows

Then it repeated the format.

Again.

And again.

By 2025, Supabase had reached Launch Week 15.

Launch Week worked because it was more than a marketing campaign.

It created a shipping rhythm.

The team knew another deadline was coming.

Developers knew new products were coming.

Every Launch Week created more:

product announcements
technical posts
social content
Hacker News discussions
community conversations
product feedback

Supabase said in 2021 that its product-led growth strategy had helped the number of databases it managed grow 47% month over month for its first 18 months.

That does not mean Launch Week alone caused the growth.

Open source, product improvements, community, word of mouth, and other channels were growing at the same time.

But Launch Week gave Supabase a recurring way to turn new product work into distribution.

7. Engineers helped market the things they built

Supabase did something else that worked particularly well with developers.

The people building a feature often helped explain it.

An engineer could write the launch post for the product they had worked on.

That made the content more technical and useful.

Instead of saying:

We are excited to announce our revolutionary new feature.

Supabase could explain:

how it worked
why it was built
architecture decisions
benchmarks
tradeoffs
examples
migration steps

The company also changed its launch strategy based on what it was shipping.

A frontend product might be shared with:

frontend developers
designers
Product Hunt

A database infrastructure product might be better suited to:

Hacker News
GitHub
backend communities
technical newsletters

Supabase did not treat every launch the same.

It went where the relevant developers already were.

It also learned that the same product could be launched more than once.

A feature could quietly ship.

Then appear in a monthly update.

Then get a larger announcement during Launch Week.

Most companies assume people saw the first announcement.

Usually, most people did not.

Supabase kept giving useful products more chances to be discovered.

8. Technical content captured developers with intent

Launches create spikes.

Technical content can keep bringing users for years.

Supabase built a large library of documentation and tutorials around topics such as:

Next.js
React
Flutter
authentication
Postgres
row-level security
migrations
vector search
AI applications
storage
example apps

This content served a different purpose from Launch Week.

A developer reading Hacker News might discover Supabase before they need it.

A developer searching:

how to add Supabase Auth to Next.js

already has a problem to solve.

That person is much closer to creating a project.

Technical education helped Supabase move developers from:

knowing Supabase exists

to:

building with Supabase.

It also gave Supabase long-term search distribution.

A useful tutorial can continue bringing developers to the product long after a launch ends.

The best Supabase content usually helped somebody create or fix something.

It did not simply tell them Supabase existed.

9. AI gave Supabase another growth wave

Supabase was already growing quickly before the AI coding boom.

Then AI created a new source of demand.

At first, the connection was vector search.

Supabase made pgvector easy to use and published a tutorial for storing OpenAI embeddings in Postgres in early 2023.

Developers could store:

normal relational data
vector embeddings

inside the same Postgres database.

That made Supabase useful for early AI applications.

But vector search was only the beginning.

The much bigger change came from AI coding tools.

Products like:

Lovable
Bolt
Cursor
Claude Code
Replit
v0
Figma Make

made it much easier to create software.

But the apps they generated still needed infrastructure.

They needed:

databases
authentication
APIs
storage
backend logic

Supabase already bundled those things together.

Its product suddenly fit one of the fastest-growing software trends in the market.

10. AI builders became a new distribution channel

Before AI builders, Supabase usually had to win developers one at a time.

A developer would start building something.

Then they would research backend options.

Then they might choose Supabase.

AI builders changed that flow.

A person could open Lovable and type:

Build me a SaaS app.

Lovable still needed somewhere to store the application's data.

If Supabase was integrated, the platform could connect the application to Supabase.

The user might become a Supabase user without ever searching Google for:

best backend for my app

The platform had already made part of that decision.

This created a much more powerful distribution model.

One integration could create thousands or millions of Supabase projects.

Supabase went deeper into this model with Supabase for Platforms.

Supabase for Platforms lets another company provision Supabase backends for its own users.

Instead of acquiring every developer directly, Supabase can acquire an entire platform.

By late 2025, Supabase said platforms including Lovable, Bolt, and Baidu's MeDo had created millions of Supabase projects.

By 2026, Supabase called Supabase for Platforms its fastest-growing product.

A product integration had turned into a growth channel.

11. Coding agents started creating databases themselves

Supabase did not stop with AI app builders.

In April 2025, it released an official MCP server.

MCP, or Model Context Protocol, lets AI tools connect to external products and take actions.

Supabase's MCP server allowed compatible coding agents to do things like:

create projects
run SQL
inspect logs
manage database branches
retrieve project configuration
generate TypeScript types

The important change was simple:

AI could operate Supabase directly.

A coding agent did not have to tell the developer:

Go to Supabase, create a project, then paste these credentials.

The agent could increasingly do the setup itself.

Supabase also began writing more documentation and product interfaces for both humans and agents.

By June 2026, the change was showing up in the company's numbers.

Supabase said:

nearly 10 million developers were using the platform
database launches were up 600% year over year
more than 60% of new databases were being launched through AI tools

CNBC also reported that Claude Code had become Supabase's largest source of new database creation in 2026.

AI had gone from being a use case for Supabase to being one of its biggest distribution channels.

12. Supabase moved from human distribution to machine distribution

This is the most important part of the Supabase growth story.

Its first five years were mostly about convincing people.

The Firebase positioning made developers curious.

Hacker News gave the company attention.

Open source created trust.

A free tier made experimenting easy.

Launch Weeks repeatedly brought people back.

Technical content reached developers when they had a problem.

But AI changed where the decision happened.

The old flow looked like this:

developer starts app → developer researches backends → developer chooses Supabase

The new flow can look like this:

developer asks AI to build app → AI provisions Supabase → database gets created

Supabase is still serving the developer.

But it increasingly needs to win over the tools the developer uses too.

That is machine distribution.

And it could become one of Supabase's strongest advantages.

Supabase's growth loops

Supabase did not have a single growth loop.

It stacked several.

Positioning loop

clear Firebase comparison → developer understands Supabase → developer tries it → more word of mouth

Product loop

free database → developer builds application → application uses more Supabase products → usage grows → project upgrades

Open-source loop

developer discovers Supabase → uses or contributes to it → community creates more integrations and content → more developers discover Supabase

Launch loop

build → launch → traffic → feedback → improve → build → launch again

Content loop

developer searches for solution → finds Supabase tutorial → creates Supabase project → ecosystem grows

Platform loop

AI platform integrates Supabase → platform users create apps → Supabase projects get created automatically → usage grows

Agent loop

developer asks coding agent to build app → agent creates Supabase infrastructure → more apps run on Supabase

Each loop strengthened the others.

Why did Supabase grow so fast?

There were six main reasons.

1. Clear positioning

"The open-source Firebase alternative" made Supabase easy to understand and easy to share.

2. Postgres

Supabase built on a database developers already knew and trusted.

Users could start quickly without committing to a new proprietary database.

3. A low-friction free tier

Developers could try Supabase before their project had users or revenue.

4. Open-source distribution

GitHub, community contributors, integrations, self-hosting, and the wider Postgres ecosystem gave Supabase reach and credibility.

5. A repeatable shipping system

Launch Week continuously turned product development into marketing and community attention.

6. AI distribution

AI app builders and coding agents turned Supabase from something developers manually chose into infrastructure software could provision automatically.

Supabase growth by the numbers

Metric	Reported figure
Visitors from May 2020 Hacker News breakout	30,000 in one week
Signups from that week	1,400+
Developers in 2021	40,000+
Hosted databases in 2021	50,000+
Developers in 2022	110,000+
Hosted databases in 2022	150,000+
Managed databases by April 2024	1M+
GitHub stars by April 2026	100,000+
Developers in April 2026	8M
Developers in June 2026	Nearly 10M
YoY database-launch growth in June 2026	600%
New databases launched through AI tools	60%+

Supabase has not publicly disclosed its revenue.

Third-party ARR estimates should not be treated as confirmed company figures.

The company also has not said that its nearly 10 million registered developers are all monthly active users.

What did not go perfectly?

Supabase's growth was not a straight line.

Its first Hacker News spike caused infrastructure problems almost immediately.

Self-hosting Supabase can also be more complicated than the phrase "open source Postgres backend" suggests.

Supabase is not a single Postgres binary.

It contains several services that need to work together.

Security is another challenge.

Supabase gives developers powerful tools such as row-level security.

But they still need to be configured correctly.

This became even more important as AI tools started generating applications for people with less database experience.

Researchers have found cases where applications built on top of Supabase exposed data because security policies were configured incorrectly.

That was not a breach of Supabase itself.

But it showed the other side of AI-driven growth.

AI can scale database creation.

It can also scale configuration mistakes.

Supabase has also had outages, including a regional incident in February 2026 that lasted several hours.

The larger the platform becomes, the more reliability matters.

Growth does not reduce infrastructure problems.

It increases the cost of getting them wrong.

The biggest lesson from Supabase

It is easy to look at Supabase and copy Launch Week.

That would miss the more important lesson.

Supabase kept getting closer to the moment when infrastructure gets chosen.

In 2020, that meant showing up on Hacker News when developers were looking for an alternative to Firebase.

Then it meant being present on GitHub, in technical tutorials, inside framework ecosystems, and across developer communities.

Today it increasingly means being available inside the AI tool building the application.

At every stage, Supabase reduced friction.

Open source reduced the fear of lock-in.

The free tier reduced the cost of trying it.

Postgres reduced the risk of outgrowing it.

Launch Week gave developers repeated reasons to pay attention.

Technical content helped developers use it.

AI integrations eventually removed the need for some developers to choose Supabase manually at all.

The growth strategy looks like this:

clear positioning → easy product → free tier → open-source distribution → repeatable launches → technical content → platform integrations → agent-driven distribution

Supabase started by making Postgres easier for humans.

Its latest growth wave came from making Supabase easy for machines to use too.

Frequently asked questions

How did Supabase grow?

Supabase grew by positioning itself as the open-source Firebase alternative, making Postgres easy to start with, offering a free tier, building an open-source community, repeatedly launching products through Launch Weeks, publishing technical content, and later integrating deeply with AI builders and coding agents.

Why did Supabase become popular?

Supabase combined the ease of a backend-as-a-service product with Postgres and open source. Developers could start quickly while keeping their data inside a familiar relational database.

How many developers use Supabase?

Supabase reported nearly 10 million registered developers in June 2026. The company has not said that all of them are monthly active developers.

How many databases does Supabase manage?

Supabase reported more than one million managed databases when the platform reached general availability in April 2024. It has since said database launches grew 600% year over year, but it has not published a newer total database count.

Did Hacker News make Supabase successful?

Hacker News created Supabase's first major growth spike. Its May 2020 breakout brought around 30,000 visitors and more than 1,400 signups in a week. Later growth came from product improvements, open source, Launch Weeks, technical content, integrations, and AI coding tools.

What is Supabase Launch Week?

Launch Week is Supabase's recurring product launch format. The company announces several new products or features over a fixed period. It began as a way to recreate the shipping pressure the founders experienced during Y Combinator.

Is Supabase open source?

Supabase builds much of its platform in the open and relies heavily on open-source technology such as Postgres. The company also contributes to and supports projects across the wider open-source ecosystem.

How did AI help Supabase grow?

AI coding tools created a new distribution channel for Supabase. Platforms such as Lovable and Bolt could provision Supabase backends for users, while coding agents could interact directly with Supabase through MCP. By June 2026, Supabase said more than 60% of new databases were being launched through AI tools.

What is Supabase's revenue?

Supabase has not publicly disclosed its revenue. ARR figures published by private-company databases and analysts are estimates rather than confirmed company figures.