Superbase Growth: AI Agents, Builder Boom, and Personal Databases

•

 48 min video

•

 5 min read

YouTube video ID: vl7HiMdGfjo

Source: YouTube video by Y Combinator — Watch original video

PDF

Paul Coppstone, co-founder and CEO of Superbase, an open-source PostgreSQL backend, discussed the company's explosive growth, its shift towards supporting AI agents and non-engineer builders, and the future of personal databases. Superbase has grown from 3 million signups to 13 million in the past year, with revenue quadrupling and databases under management increasing from 8 million to nearly 50 million. The company is now spinning up over a million databases per week.

The Rise of the "Builder" and AI Agents

Coppstone highlighted a significant shift in Superbase's user base. While software engineers continue to be core users, a new category of "builders" is emerging—individuals without traditional software engineering backgrounds who are leveraging AI to create applications. This trend is driven by:

  • Accessibility of AI: AI tools are making it possible for non-engineers to build and customize software, removing the need to rely on developers.
  • Code-First Approach: Superbase has evolved from a dashboard company to an API company, and now a "code-first" company, aligning with how AI agents prefer to build, often using command-line interfaces (CLIs).
  • Default Choice in LLMs: Superbase has become a default database choice within large language models (LLMs), significantly boosting its adoption by agents.
  • Open Source Advantage: Being open source is critical, as agents prefer to inspect source code when issues arise, and enterprises increasingly demand transparency for security.

This shift has led to an acceleration in Superbase's top-of-funnel signups, alongside an increase in activation and conversion rates, which is unusual during periods of rapid growth.

Personal Software and the "End of One"

Coppstone shared his personal experience building a comprehensive personal application called "Control Tower" using Superbase. This app manages various aspects of his life, including:

  • Email Triage: A UI to manage and prioritize emails without needing to read his inbox.
  • Finances: Tracking spending and managing personal finances, offering features comparable to commercial apps like Mint or Monarch, but with personalized context (e.g., identifying transactions by cross-referencing his "Gbrain" for contact information).
  • Health: Tracking weight, workouts, and other health metrics.
  • File Uploads: A simple interface to upload or email files, which his agent then processes and stores on Superbase, determining relevance and appropriate actions.

He emphasized that such personalized applications, especially for finance and health, are becoming increasingly valuable as individuals seek to own and correlate their own data. The ability to sync various devices to a single database and prompt agents to find correlations (e.g., linking weight fluctuations to travel from his calendar) provides unique and powerful insights.

This trend highlights a future where individuals will likely have their own personal databases, serving as a "store of record" for their lives, allowing them to build applications and attach different agents on top, rather than relying on third-party databases.

Scaling for a Billion Builders

Superbase's mission has expanded from serving YC companies to servicing "everyone in the world with access to AI" who will become a builder. The company currently launches a million databases weekly, but projects a future demand of 1.2 billion databases per month if all ChatGPT users were to launch something monthly.

To meet this unprecedented scale, Superbase is:

  • Architectural Innovation: Exploring new architectures beyond traditional PostgreSQL, including the acquisition of Terso, which offers a PostgreSQL and SQLite-compatible architecture.
  • Superbase Spark: Introducing a new tier called "Superbase Spark" for prototyping and SQLite-type workloads, aiming for a "prototype to petabytes" scaling experience.
  • Dual Product Strategy: Maintaining two distinct product lines: "Multigresses" for petabyte-scale enterprise needs and "Terso" for smaller, prototyping workloads.

Internal AI Adoption and Agent-Centric Development

Superbase itself has seen a dramatic increase in AI adoption internally. Within six months, the percentage of engineers writing 80% of their code with coding agents jumped from 10% to 60%. This is attributed to:

  • No AI Policy: Allowing engineers freedom to use any tools they choose.
  • Model Improvement: The increasing capabilities of AI models.
  • Testing and Tooling: Adopting robust testing tools to ensure the quality of AI-generated code.

The company is now actively designing its product for agents, not just humans. This includes:

  • Code-First and Git-Native: Ensuring that dashboard functionalities work seamlessly with code and infrastructure-as-code principles.
  • Agent-Friendly Error Messages: Re-engineering error messages to provide agents with actionable instructions on how to fix problems, rather than just reporting an error.
  • Agent Testing: Using an internal evaluation suite and public eval website to test how different AI models interact with Superbase, identifying areas for improvement.

The Future of Entrepreneurship and Work

Coppstone expressed excitement about the future, envisioning a surge in entrepreneurship driven by the accessibility of building tools. He believes that many individuals, including those previously in blue-collar roles, will discover new skill sets and opportunities, leading to a more fulfilling lifestyle.

He also noted a current trend where vertical SaaS founders, even with significant ARR, are struggling to raise Series A funding. However, this also means companies can achieve substantial profitability with smaller teams (e.g., 10-15 people reaching $10-50 million ARR), making them less reliant on external funding.

Lessons for Future Builders

Coppstone offered advice for those looking to emulate Superbase's success:

  • Durable Categories: Choose categories with long-term relevance, like databases, which have proven their longevity.
  • Adaptability: Be highly adaptable and maintain capacity within teams to shift quickly, as the pace of change is accelerating.
  • Dogfooding: Actively use your own tools to identify gaps and areas for improvement.
  • Ecosystem Partnerships: Collaborate with key players in the ecosystem to gain feedback and explore new ideas, understanding that not all bets will pay off, but the successful ones can lead to significant growth.
  • Focus on the Future: Always keep an eye on the next wave of building, prioritizing innovation and developer experience.

The conversation concluded with a vision of a future where personal agents attend virtual conferences, exchanging tips and integrating new features into codebases, further blurring the lines between human and AI development.

  Takeaways

  • Superbase expanded from 3 million to 13 million sign‑ups in a year, quadrupling revenue and scaling databases from 8 million to nearly 50 million, now launching over a million databases weekly.
  • The company is witnessing a new “builder” segment—non‑engineers using AI to create apps—driven by AI accessibility, a shift to a code‑first API model, and its status as the default database for large language models.
  • Coppstone’s personal “Control Tower” app demonstrates how AI‑augmented personal databases can unify email, finance, health, and file management, offering users a single “store of record” for their life data.
  • To support a projected 1.2 billion monthly databases, Superbase is introducing architectural innovations like the Terso acquisition and the “Superbase Spark” tier, while maintaining separate product lines for petabyte‑scale and prototyping workloads.
  • Internally, Superbase engineers now write 80 % of code with AI agents, prompting the company to redesign error messages, adopt agent‑friendly testing, and build products explicitly for AI agents rather than just human developers.

Frequently Asked Questions

Why is Superbase described as the default database choice within large language models?

Superbase is called the default database for LLMs because the models have been trained on its open‑source PostgreSQL schema and API patterns, making it the most familiar and easily callable data store when agents generate code. This familiarity speeds up agent‑driven app creation and boosts adoption.

What is Superbase Spark and how does it support scaling from prototype workloads to petabyte‑scale databases?

Superbase Spark is a new tier designed for lightweight, SQLite‑compatible workloads that lets developers prototype quickly and then scale seamlessly to petabyte‑level storage. By leveraging a hybrid PostgreSQL/SQLite architecture, Spark bridges the gap between fast iteration and massive data capacity, supporting the projected billions of databases.

Who is Y Combinator on YouTube?

Y Combinator is a YouTube channel that publishes videos on a range of topics. Browse more summaries from this channel below.

Does this page include the full transcript of the video?

Yes, the full transcript for this video is available on this page. Click 'Show transcript' in the sidebar to read it.

Helpful resources related to this video

If you want to practice or explore the concepts discussed in the video, these commonly used tools may help.

Links may be affiliate links. We only include resources that are genuinely relevant to the topic.

Full transcript is not shown on this page

This page focuses on the summary and original notes. For full verification, refer to the original YouTube video.

PDF