Convex raises $57M to harden agent backends
Convex closed a $57M Series B led by Insight Partners, betting that AI coding agents need a more reliable backend than what most app builders wire together.
Convex raised $57 million in a Series B led by Insight Partners, with Etna Labs, Spark Capital, a16z, and Justin Kan also participating. The San Francisco company, founded by former Dropbox infrastructure engineers, builds a backend platform designed for software written by AI coding agents, with ACID transactions and automatic data synchronization instead of the usual assembled-from-parts database-plus-API layer.
Convex says millions of its instances are now running in production across roughly 500,000 developers, with customers including OpenAI, Tripadvisor, Solana, Zapier, and Reducto. The company plans to use the new funding to close the gap between its current implementation and the scale its growing customer base needs, including multi-tenant enterprise support and EU hosting, which is now live.
What it means for operators
Convex’s pitch is a specific bet: as more backend code gets written by an agent rather than a person, the reliability problems shift from “did the agent write correct code” to “does the infrastructure under that code hold up without a human reviewing every deploy.” That’s a real gap. Fake bug reports have already been shown to hijack coding agents into making bad changes, and a backend built for agent-written code to fail safely is a different design target than one built for a careful human PR review. If you’re running AI coding agents against a hand-rolled backend and hitting sync bugs or transaction issues an agent introduced without noticing, that is exactly the gap Convex is selling into, alongside the broader wave of AI-tooling funding rounds betting infrastructure, not just model quality, is where the next round of AI product value gets built.