Supabase is acquiring Turso

Date: 2 Oct 2026 | Reading Time: ~3 minutes | Author: Paul Copplestone, CEO & Co-Founder

The rise of Artificial Intelligence is empowering developers to generate software at an unprecedented velocity. Currently, AI agents are deploying millions of databases to support a flood of new prototypes, dashboards, explorations, and full-scale applications.

This shift toward massive-scale deployment necessitates a fundamental change in how we handle database infrastructure. While Supabase will maintain its focus on Postgres, Turso will continue advancing SQLite. Together, we are merging these strengths to provide a seamless experience for every AI agent.

---

🤖 Databases for Agents

Supabase is currently seeing a surge in demand, launching over one million databases every single week. We anticipate that as agentic AI continues to build software, the demand for databases will exceed the current global capacity.

To survive this growth, we need a new architectural approach. The goal is to make database creation as frictionless as creating a simple text file.

The Vision for Agentic Infra

"Agents should be able to create a database as easily as creating a file, with just as little concern about cost."

To achieve this, we are moving away from legacy constraints:

• Provisioning a dedicated machine for every small workload $\rightarrow$ On-demand, shared infrastructure.
• High overhead costs $\rightarrow$ Near-zero cost for initial creation.
• Fragmented workflows $\rightarrow$ Unified DX from prototype to production.

Scaling Logic

If $D$ represents the number of databases and $C$ represents the capacity, we are currently seeing: $\lim_{t \to \infty} \frac{D(t)}{C(t)} > 1$ This indicates that demand will eventually outpace existing infrastructure unless we evolve.

---

⚡ The Turso Advantage

Turso has already engineered the exact solution required for this new era. By rebuilding SQLite using Rust, they developed a cloud platform capable of managing millions of databases on a single server.

How it Works

The system uses a "load-on-demand" mechanism:

1. Active: Database is loaded into memory when needed.
2. Idle: Database is suspended when not in use.

This efficiency allows companies like Superhuman, Sauna.ai, CTO.new, and Mastra to provision databases for agents instantly, whether on Turso's cloud or within their own private clouds.

Infrastructure Comparison

---

🌐 Open Source & Future Integration

Both Supabase and Turso are driven by a shared philosophy: a deep commitment to open source, a relentless focus on developer experience (DX), and a passion for solving the most difficult infrastructure challenges.

Integration Roadmap

• Acquire Turso team and technology.
• Integrate Turso's SQLite capabilities into the Supabase ecosystem.
• Establish a seamless migration path for workloads as they scale from SQLite to Postgres.

We are thrilled to welcome Glauber Costa, Pekka Enberg, and the entire Turso team to the Supabase family.

Technical Implementation Example

In the future, spinning up an agent database might look as simple as this:

// Conceptual API for Agent DB creation
const agentDb = await supabase.agents.createDatabase({
  engine: 'sqlite', 
  template: 'prototype-v1',
  autoSuspend: true
});

console.log(`Database ready at: ${agentDb.url}`);

Our ultimate mission remains the same: to store the world's data. With the explosion of AI agents, there is more data to store than ever before.