Labelbox Secures $100 Million to Scale AI Agent Infrastructure

The company reached profitability in 2026, marking a fivefold increase in growth year over year. Labelbox intends to funnel these resources into its two flagship products: Horizon and Recursion Managed Agents. Horizon provides the necessary infrastructure for frontier labs to conduct reinforcement learning, allowing models to practice complex tasks—such as coding and research—within simulated environments.

Recursion Managed Agents shifts the focus to enterprise utility, enabling companies to deploy fleets of AI agents that operate within existing business software. These agents improve their decision-making capabilities with every iteration, learning from the specific outcomes of their previous tasks. CEO Manu Sharma noted that the platform aims to integrate sandbox infrastructure with post-training systems, ensuring that AI models evolve alongside the requirements of the businesses they serve.