Relace raises $23M Series A to build AI agent infrastructure — used by Lovable, Figma, and 38 other companies

Summary

Relace has closed a $23 million Series A, positioning itself as infrastructure for AI coding agents rather than a direct competitor to tools like Cursor or Cognition. The company, founded by Preston and co-founder Eitan Borgne — both PhD dropouts, from Harvard and UChicago respectively, who met as undergrads at Caltech — launched shortly after ChatGPT's release and has since signed 40 customers, including Lovable and Figma.

Relace's core product thesis is that agents require a different set of auxiliary models than human developers do. Where a developer uses tab autocomplete, an agent needs fast diff-application models and embedding and re-ranking systems that can surface relevant code from a codebase in under a second. That infrastructure layer is what Relace sells.

Relace Repos, announced alongside the Series A, embeds retrieval directly into source control. The product is a bet that version control itself needs to be rebuilt for agents — with model-native capabilities baked in from the start rather than bolted on.

The broader strategic conviction is that the definition of software will expand significantly over the next two to three years as model capability increases. Relace points to customers already building PowerPoint-style presentations in TypeScript and React, and consumer TikTok-style filters expressed entirely in code, as early signals. The company frames its long-term position as "the rails for software on demand," anticipating that coding agents will generate and run much of what is not yet considered software today.

Over the next twelve months, Relace plans to use the raise to co-optimize its models and underlying infrastructure, drawing an analogy to how Cursor tightened the integration between its apply model, tab autocomplete, and IDE product to achieve a seamless user experience.