Daytona raises $24M Series A to build agent-native compute infrastructure

Croatian-founded Daytona has secured a significant $24 million Series A funding round to develop agent-native compute infrastructure, a crucial development for the growing field of AI agents. Led by FirstMark Capital, with participation from prominent venture firms and strategic investors like Datadog and Figma Ventures, this investment underscores the increasing demand for specialized computing solutions to support advanced AI workloads. The company aims to build infrastructure capable of handling the massive scale required by these sophisticated software agents. Daytona's innovative approach centers on its concept of "sandboxes" as a foundational infrastructure primitive. Unlike traditional cloud infrastructure optimized for static production workloads, Daytona's sandboxes are designed for the dynamic and experimental nature of agent tasks. These programmable, composable environments can be provisioned on demand, offering rapid startup times, the ability to branch into parallel executions, support for snapshots, and efficient scaling across numerous concurrent instances, all vital for AI agent operations. The implications of Daytona's technology are far-reaching, particularly for developers and organizations leveraging AI agents for tasks ranging from coding to complex decision-making. By providing infrastructure that allows agents to launch sandboxes, persist state, and explore alternative execution paths at scale, Daytona is enabling more robust and efficient AI agent development and deployment. This development also signals a broader industry trend towards infrastructure that is increasingly optimized for automated workloads rather than purely human-centric workflows, potentially boosting European digital sovereignty in AI infrastructure. Following this funding, Daytona plans to expand its offerings, scale its systems for higher agent workload volumes, and deepen integrations with existing developer and agent tooling.