The name was a placeholder.
Paul Copplestone had been working through a list of database name options. Ultra Base, Hyper Base, something with “base” at the end. He needed a throwaway name to share ideas with his future co-founder Ant Wilson, so he typed Supabase, with a U instead of an E, half as a joke, half because he and Ant liked the Nicki Minaj song. It was the kind of name you pick when you’re certain you’ll rename the company before anyone sees it.
Then someone posted it to Hacker News. The post went to the top. Within hours, Copplestone realized they couldn’t change it. Everything inside the company became Super something. He calls it one of his regrets. The company is now worth $5 billion.
This is the Supabase success story: a New Zealand farm kid, a technical problem he hit at a startup in Southeast Asia, a tagline swap that changed everything, and one of the most remarkable valuation trajectories in recent startup history. From $765 million to $5 billion in under 12 months. Four funding rounds in a year. The default backend for the tools that are rewriting how software gets built.
The Founder and the Farm
Paul Copplestone grew up on a farm near Kaikoura, New Zealand. He started coding at 18 when he left school. No formal computer science education, no prestigious university pipeline. He taught himself. That self-taught background shaped how he thought about developer tools: if it was hard to understand, it was a design problem, not a user problem.
He eventually made his way to Singapore through Entrepreneur First, a talent investor that brings developers and founders together to build companies. At Entrepreneur First he met Ant Wilson, a systems engineer who would become Supabase’s co-founder and CTO.
Copplestone left the Entrepreneur First program before it finished, which is not the path, to start a different company called Nimbus for Work, an office management platform built with a former colleague from ServisHero, another Singapore startup he’d worked at. The detour matters because of what happened while he was building Nimbus.
The team needed a chat feature with real-time functionality, similar to WhatsApp. They used Firebase, Google’s backend-as-a-service product, which had become the default starting point for developers who needed real-time data and didn’t want to manage server infrastructure.
Firebase hit a wall. The specific problem was a one-query-per-second limit on individual Firestore documents, which meant the chat application was not updating properly for users. They had to refresh to see new messages. This was not a minor UX issue. It was a fundamental constraint built into the product’s architecture.
Copplestone went to investigate Postgres as an alternative. Postgres was the better database for relational data, flexible queries, and scale. But Postgres did not have Firebase’s killer feature: real-time subscriptions. With Firebase you insert data and any number of devices listening to the database receive the update instantly. Postgres did not have that natively.
He built a workaround, attaching serverless functions to Firebase to replicate data into a Postgres database. The hack worked. And then he thought about everyone else who had hit the same wall and had to build the same hack.
That was the thesis. Firebase had made backend development fast and accessible. Postgres was the better database for almost any real application that needed to scale or query flexibly. The gap between them was the product: a Firebase-level developer experience built on Postgres instead of a proprietary document store.
He pitched Ant Wilson on the idea. Ant joined as co-founder and CTO in January 2020.
The Eight Databases
The first version of Supabase was positioned as “real-time Postgres.” That was technically accurate and immediately interesting to a narrow audience of developers who already understood the problem it was solving.
By April 2020, three months after they started building, Supabase had eight hosted databases.
Eight. Not eight hundred. Not eighty. Eight.
Copplestone and Wilson had raised a $100,000 pre-seed from angel investors and joined Y Combinator’s Summer 2020 batch. The Hacker News buzz had been real. The investor validation was real. The actual usage was not.
The product was fine. The positioning was wrong.
Copplestone had been browsing Hacker News when the idea came to him: change the tagline from “real-time Postgres” to “the open-source Firebase alternative.” It was a small change in words that was a complete change in message. Instead of describing a technical capability that a developer had to already understand to appreciate, it described a relationship to something developers already had opinions about. Firebase was the thing they were using. Firebase was the thing frustrating them. Here was an alternative.
He made the change and someone posted the new Supabase to Hacker News. The post got over 1,100 upvotes. It hit the front page two days in a row. Copplestone woke up to an overloaded Slack, a free tier being consumed at a rate they had not anticipated, and a user count that had jumped from 80 to 800 in a single night.
Eight databases to hundreds in days. Not because the product changed. Because the message finally matched how developers actually thought about the problem.
Why Open Source Was the Bet
Most startups building developer tools have a choice to make about open source. You can build a proprietary product and keep the code closed. You can open source everything and try to monetize around the edges. You can open source the core and keep commercial features proprietary, which is what Hashicorp did with Terraform and what MySQL had done before Oracle acquired it.
Supabase chose a specific version of this: everything they built was open source, and the business model was hosting. You could run Supabase yourself on your own infrastructure. You could also pay Supabase to run it for you, and the hosted version was polished enough that most developers, even those capable of self-hosting, chose the managed product.
This was not the obvious commercial choice. Open source means your code is available to anyone, including competitors who can fork it and offer their own hosted version. It means you cannot lock customers in through technical barriers. It means some percentage of your most capable users will self-host and never pay you anything.
The bet Copplestone and Wilson were making was different. In developer markets, trust is the moat. Developers are intensely skeptical of vendor lock-in because they’ve been burned before. Firebase itself was an example: adopt it early, build on its proprietary data model, and discover later that migrating was far harder than starting with Firebase had been. Developers who had been through that experience once were permanently more cautious about proprietary tools.
Open source is a trust signal. If the code is public, you can read it, audit it, understand exactly what it’s doing with your data, and know that if the company disappeared, you could still run the software. That trust, compounded over time with a developer community that talked to itself constantly on GitHub, Twitter, Discord, and Hacker News, became the primary distribution channel.
Supabase’s first approximately 50 employees came from the open-source community. This was deliberate. When you hire the maintainer of PostgREST, which Supabase uses to auto-generate APIs from Postgres schemas, you bring the person who already understands the technical decisions most deeply into the team. Steve Chavez, a PostgREST maintainer, was an early hire. The pattern repeated across other open-source tools Supabase integrated.
The community didn’t just provide distribution. It provided product direction, bug reports, and technical contributions that accelerated development faster than a closed team could have managed.
Launch Week and the Developer Marketing Playbook
Supabase’s marketing approach was as unconventional as its product positioning.
The primary channel was Twitter, specifically Ant Wilson’s willingness to post memes. “Anybody who knows Ant, knows that he loves memes,” Copplestone wrote in a blog post. The meme-friendly name they had accidentally kept turned out to be a genuine marketing asset. Developer humor spreads through Twitter in ways that press releases do not.
The format that became most associated with Supabase’s growth was Launch Week: a structured week of daily feature releases, each announced separately, building toward a cumulative product update that felt larger than any individual release would have. The format created appointment viewing for the developer community. Every day of Launch Week was a new reason to share, discuss, and engage. At the end of each Launch Week, Supabase saw measurable spikes in GitHub stars and Discord membership.
The Discord community grew to over 20,000 developers, with Copplestone, Wilson, and engineers participating actively rather than leaving it to a community manager. In the early days, Copplestone and Wilson personally messaged every user to collect feedback. The 1-on-1 attention at small scale built the reputation for being genuinely responsive that persisted as the community grew.
This is an approach that doesn’t scale in the traditional sense but builds something that scales in a different way. When developers trust that the people building the tool are actually listening to them, they advocate for it. They write tutorials, record videos, share it in team Slack channels, recommend it to developers at other companies. The organic spread of developer tool adoption, when it works, is exponential in a way that paid acquisition cannot replicate.
By July 2020, two months after the positioning change and the Hacker News front page hits, Supabase was hosting over 3,000 databases. From eight to three thousand in three months.
The Postgres Bet
Building on Postgres was a specific technical decision that had implications far beyond the database itself.
Postgres is the world’s most deployed open-source relational database. It has been in active development for decades. It has an enormous ecosystem of tooling, extensions, experts, and documentation. Virtually every developer who has worked on a web application for more than a few years has some Postgres experience. When something goes wrong, there are answers on Stack Overflow.
Building on Postgres rather than a proprietary database layer meant Supabase inherited all of that. It also meant Supabase could position itself as the opposite of Firebase on the data portability question. Your data is in Postgres. You can export it, move it, query it with standard SQL tools, and take it anywhere. There is no proprietary data format designed to make migration difficult.
The Postgres bet also set up Supabase well for AI workloads in a way that proprietary database choices would not have. Postgres has pgvector, an extension that adds vector storage and similarity search natively into the database. As AI applications increasingly needed vector databases to store embeddings for retrieval-augmented generation and other patterns, Supabase could add this capability as a feature of its existing product rather than building a separate system. The vector toolkit became a standard part of the Supabase stack.
Copplestone has spoken about the Postgres bet being made when the database’s leadership was less certain than it is now. Postgres has only gotten more dominant in the years since Supabase launched. The bet aged well in a way that a proprietary database architecture would not have.
55% of the YC Batch
The most telling statistic about Supabase’s position in the developer ecosystem is not its valuation or its funding. It is this: 55% of the most recent Y Combinator batch uses Supabase as their backend. More than 1,000 YC companies total.
Y Combinator companies are not average startups. They are companies that have been through an intense three-month program specifically designed to help them identify what matters, move fast, and survive. The founder profile skews heavily technical. The environment rewards tools that work and punishes tools that require maintenance overhead.
When more than half of a YC batch independently chooses the same database, that is product-market fit at a specific and influential segment of the developer market. Those YC founders recommend it to other founders. They take it to companies they join or advise later. The tool spreads through the alumni network without Supabase spending anything to put it there.
The dynamic is similar to Stripe’s early adoption within YC, or Vercel’s among Next.js developers. When you become the default choice for a high-credibility technical community, the community does your marketing for you.
Vibe Coding and the Multiplier Nobody Planned For
In 2024 and into 2025, the AI coding tools boom changed the growth curve for Supabase in a way that the team describes as being in the right place at a fortuitous time.
Vibe coding tools, the term for AI-assisted development platforms where a user describes what they want in plain English and receives a working application, needed backend infrastructure. Lovable, Bolt, Replit, Cursor, Figma Make, and others were each building platforms where developers and non-developers could build functional web applications with AI assistance. Each of those applications needed a database, authentication, file storage, and real-time subscriptions.
The Supabase product was precisely what these platforms needed. When a vibe-coding tool scaffolds a new project, it can provision a Postgres database, configure authentication, set up storage, and wire real-time subscriptions, all automatically through Supabase, in seconds. No infrastructure decisions. No database administration. No DevOps.
Supabase became the default backend for the vibe-coding movement. Lovable reached $200 million in ARR in 2025. Bolt, Replit, and Cursor each added millions of users. Every one of those projects generated Supabase databases. User growth surged 700% year-on-year during the 2024-2025 vibe coding boom. The user base went from 1 million to over 4 million developers in under a year.
The network effect compounded: more vibe-coding users meant more Supabase databases, which meant more revenue, which funded deeper infrastructure investment, which made Supabase better for the next wave.
Copplestone describes the AI builders as representing approximately 30% of Supabase signups. The other 70% are traditional developers using it the way they would have used any Postgres hosting service, just with a better developer experience and a full suite of adjacent tools included.
From $765M to $5B in 12 Months
The funding trajectory is worth stating directly because it captures the pace at which the market updated its assessment of Supabase’s position.
Series C, September 2024: $80 million led by Peak XV and Craft Ventures. PitchBook estimated the post-money valuation at approximately $765 million.
Series D, April 2025: $200 million led by Accel. Valuation: $2 billion.
Series E, October 2025: $100 million led by Accel and Peak XV. Valuation: $5 billion.
That is $380 million raised and a 554% valuation increase in roughly 12 months. ARR went from $20 million at end of 2024 to $70 million by mid-2025. The company had 124 employees generating $70 million in ARR, which is an unusually efficient revenue-per-employee ratio for a company at this scale.
The Series E was notable for two reasons beyond the valuation. First, Copplestone deliberately restricted it to existing investors despite significant outside interest, bringing in only Figma as a new institutional investor. His reasoning was direct: he had investors who could help Supabase reach a $50-100 billion outcome and adding more investors at this stage would not meaningfully change that math. Second, Supabase allowed its developer community to co-invest in the round, turning the open-source ethos of shared ownership into a literal equity stake for users who had helped build the company’s position.
The company has now raised $500 million total.
What the Supabase Story Is Really About
The Supabase success story is a story about the right positioning at the right moment, but the positioning only worked because the product was genuinely good.
The “open-source Firebase alternative” tagline that triggered the Hacker News flood in April 2020 would not have converted those 800 users into long-term customers if the product had been a disappointment. What made the Supabase story compound was that every time a developer tried it, they found something they wanted to keep using and tell people about.
Postgres was the right bet. Open source was the right trust mechanism for a developer audience that had been burned before. The community-first approach, founders personally messaging users, engineers active in Discord, Launch Weeks as developer events rather than press releases, built a reputation for responsiveness that scaled as word of mouth spread.
And then the AI coding wave arrived, and Supabase was already in position. The developer community that had been built organically over four years was the same community building the vibe-coding tools. The database infrastructure that had been refined across millions of projects was exactly the backend that AI-assisted development platforms needed.
Paul Copplestone wanted to find a good database name. He had settled on Supabase as a placeholder. Then someone posted it to Hacker News and the name got locked in. He calls it one of his regrets.
The company is worth $5 billion. The name sounds like a Nicki Minaj reference. Everything inside the company is called Super something and it makes the CEO cringe.
Build in a weekend. Scale to millions.
Leave a Reply