How Supabase Became One Of The Fastest Growing DevTool Companies In The World
In a Nutshell
Supabase succeeded by building an open-source PostgreSQL platform positioned as a Firebase alternative, achieving $7.5M ARR within two years through superior developer experience that reduced database setup from 8.5 minutes to 5 seconds. The company's remote-first model with 360 people across 60+ countries created a comprehensive written knowledge base that now enables AI agents to access six years of decisions, while 60-90% of new databases are currently launched by AI agents. Key principles include rejecting cash as the primary competitive advantage and betting on self-driving databases as the next 5-10 year focus rather than competing in the saturated build-stage AI tooling market.
These notes were generated by AI and may contain inaccuracies.
Money will not solve your problems, which is a grounding principle for early-stage startups. AWS has significantly more cash than new entrants, yet newer companies have succeeded for multiple reasons.
Supabase was founded in 2020 as an open-source tool to host PostgreSQL online. It has become one of the fastest growing dev tool companies globally and achieved decacorn status within a few years of founding. Paul Copplestone serves as co-founder and CEO.
Paul Copplestone is from New Zealand and currently spends most of his time in Europe. His father was an entrepreneur, making startups a natural career path. He worked one corporate consultancy job which he hated, then pursued only startups and entrepreneurship.
This is his third venture-backed startup. His first startup was as CTO rather than CEO, where he left after two years due to differing visions with the CEO about business direction. His second startup was one he helped launch with a co-founder while incubating Supabase, and that company remains profitable but was never intended as a lifelong venture.
Copplestone worked on Supabase as a side project for two years before launching in 2020. He knew he wanted to build a dev tools company and specifically a database company, but was uncertain about the exact approach. The sequence began when using both PostgreSQL and Firebase in his previous company, hitting scaling limits on Firebase, then migrating to PostgreSQL.
PostgreSQL was approximately 40 years old when Supabase began. The choice was based on several factors: Copplestone wanted an open-source database company, PostgreSQL was trending on Hacker News, and it had strong existing reputation essential for database companies. Additionally, since no one owns PostgreSQL, everyone contributes to it, creating a flywheel effect where hyperscalers must offer the best PostgreSQL services, ensuring continued improvement.
Initial positioning as "real-time PostgreSQL" did not gain traction. Switching the tagline to "open source Firebase alternative" drove significant growth. The decision to be open source was philosophical rather than tactical—both founders liked open source and building communities.
Monetization was considered from the start, though it took approximately 18 months to earn the first dollar. Databases were attractive because they create natural growth and moats—once established, usage expands naturally as customers start small and grow with the platform.
The initial customer base targeted high-growth, venture-backed startups, specifically focusing on Y Combinator companies. Over 60% of every YC batch uses Supabase, though their own batch had zero users initially due to lack of trust and remote format preventing networking.
Supabase operates as 100% open source with no enterprise version behind paywalls. The only unavailable component is the code for hosting millions of databases, which is intertwined with billing logic and presents security concerns. Major cloud providers have attempted to host and resell Supabase but face challenges since they already host PostgreSQL databases, and Supabase represents an integrated stack of tools rather than a single instance like Redis.
Community complaints about not being open source enough or not providing enough help are inevitable. The approach is to listen to feedback while making judgments about what makes business sense, acknowledging that pleasing everyone is impossible.
Trust building occurred through credibility with larger startups. First-year revenue reached approximately $1 million ARR, growing to $7.5 million the following year. Firebase remained competitive as users appreciated its experience but sought PostgreSQL and open source capabilities.
Community building focused on developer experience rather than acquiring followers. The key metric tracked was "time to value"—measuring how long it took to launch an RDS instance, connect a tool, and insert a row (8.5 minutes on AWS versus under 1 minute target for Supabase). Current performance achieves this in 5 seconds.
By end of December 2024, Supabase experienced significant database launches from AI companies, primarily Bolt and Lovable. Currently, 60-90% of databases are launched by agents, totaling millions monthly. Initial skepticism about AI use cases existed because early applications were small and growth potential was uncertain.
The "Supabase for Platforms" product emerged from AI success, enabling companies to launch millions of databases. Over 50 companies now build on this product. The same infrastructure serves enterprise needs for managing multiple databases with security, visibility, and governance features.
By early 2025, developer count grew from 6.5 million to 10 million. Charts showed acceleration across all metrics: user growth, conversion rates, and activation rates increased simultaneously. This occurred without marketing spend, as the preference is giving free databases to developers over marketing expenditures.
Three phases of developer experience emerged: initial dashboard with click-to-add functionality, then simplified dashboard with SQL editor and chat interface for LLMs, and now focus on CLI and MCP interfaces for agents. The goal is unified experience where dashboard functionality translates to code through declarative schemas, infrastructure-as-code within Supabase folders, git integration, and branching capabilities.
If starting today, founders should target agents from the beginning. However, becoming a default choice for agents is rare and luck-dependent. The recommendation is focusing on word-of-mouth, ease of experience, and being open source, while accepting uncertainty about market outcomes. Being open source likely helped Supabase significantly as LLMs consumed community discussions and GitHub content.
The concern that only AWS wins when agents dominate is acknowledged, but the response is to remain competitive. Future differentiation may come from "vibes" and user preference rather than purely technical advantages, assuming AWS can match experience quality.
The question addresses how to continue running the company after raising $500 million at over $10 billion valuation and how to make sense of life and priorities in that situation. The money should not change how the company operates because the jobs to be done remain the same and the money is in service of those jobs. If leadership starts thinking about doing something else because they now have this money, they are probably not going about it the right way.
Grounding Principles for Founders
What keeps founders grounded is going to Buddhist monasteries and maintaining the fundamental understanding that money will not make you succeed. The founders have to really know this and cannot get caught up in the money. While money can help succeed temporarily, building a generational company requires doing a bunch of things correctly. The key realization is that money will not solve your problems, which is actually a good thing for startups because it means they are limited not by the amount of cash in the bank.
This creates competitive opportunities because competitors can compete on many other vectors beyond cash. The example given is that by all rights AWS should never have allowed Superbase to exist given their cash advantage, but Superbase has been able to compete for many different reasons. Cash is really just in service of all the things that would allow the company to win even if they did not have the cash.
Investor Relationship Dynamics
The relationship with investors changes as the company grows. Normally when starting off, initial seed rounds come from seed investors, but Superbase had a different dynamic where a hedge fund invested initially. These hedge funds have $70 billion under management and at the start they invested $6 million with relatively little interest. As the company became more successful and it looked like they might go to public markets, the investors suddenly became very interested.
This creates an opposite dynamic from most seed investors, where as companies become more successful the seed investors become less interested because they have made their return, but the last investors are very interested in how much return the company will generate for them.
The company started in New Zealand, though the company was actually founded from Singapore. The team is fully distributed across over 60 countries with 360 people, meaning only a handful of people per country. There are no offices and everything is completely online.
Why Fully Remote Works
The recommendation is to commit fully to the remote model rather than attempting a hybrid approach with some people distributed and some in an office. The key is doing everything written online, which becomes particularly valuable with AI integration. Everything that has ever happened at Superbase has been written down on Slack, Notion, and other platforms, including every decision made throughout the company's history.
This creates a significant advantage because AI agents can access decisions made six years ago through prompts, whereas most people have agents that only know everything from the last six months. This means the fundamental principles of how the company operates can be encoded into agents from the very start.
The major bet being considered for the next 5-10 years involves self-driving databases. While current AI agents focus on the build stage, such as prompting them to build applications, there is less emphasis on the much harder problem of how to continue operating systems once they are built. A database represents one of these ongoing operational challenges.
The vision is for databases to operate themselves so that teams do not have to wake up at night dealing with issues like databases being down, security problems, or patches that need to be applied. This focus on the operate stage rather than just the build stage represents a harder area to launch in, which is strategically preferable because many startups launching now in the build stage are very easy to replicate.
Keep Y Combinator in your library
Save the videos and channels worth coming back to, and find them again in one place.
