Open-source infrastructure companies monetizing via cloud, founded 2024 or later

1. Pydantic — Maintains the open-source Pydantic/PydanticAI Python infrastructure stack and sells hosted observability through Logfire. Why it fits: Pydantic Ltd. was established in 2024; Logfire has moved the widely adopted open-source project toward a cloud-commercialization model.
2. Letta — Builds the open-source Letta framework for stateful, long-running AI agents, with managed deployment and operations through Letta Cloud. Why it fits: Launched in 2024 out of the MemGPT project; the open-source runtime provides a clear funnel into hosted agent infrastructure.
3. Composio — Provides open-source tooling and an integration execution layer for connecting AI agents to SaaS applications and developer tools. Why it fits: Founded in 2024 and already offers a managed platform/API alongside its open-source developer product.
4. Mastra — An open-source TypeScript framework for building AI agents, workflows, tools, and model-powered applications, paired with a hosted Mastra Cloud offering. Why it fits: Created in 2025 by the Gatsby team; its cloud product is being positioned as the operational layer for the open-source framework.
5. OpenMeter — Open-source usage-metering infrastructure for tracking events, aggregating consumption, and powering usage-based billing, with a managed cloud service. Why it fits: Commercialized as a new company in 2024; it targets a foundational infrastructure problem for API, AI, and developer-tool businesses.
6. Doubleword — Builds open-source and hosted infrastructure for serving and operating large language models, including inference deployment and scaling. Why it fits: Founded in 2024; its cloud offering monetizes the operational complexity around an open model-serving stack.
7. Tambo — An open-source toolkit for building generative user interfaces and AI-native React applications, with hosted services intended to simplify deployment and production operations. Why it fits: Founded in 2024 and still early, with an open-source developer entry point and a cloud layer aimed at teams moving beyond prototypes.
8. Motia — An open-source backend and workflow framework that combines APIs, event-driven execution, background jobs, and stateful processes, with Motia Cloud for deployment. Why it fits: Emerged as a 2024 project/company; it follows the increasingly common “open-source application runtime plus managed cloud” model.
9. Lunary — Open-source LLM observability and prompt-management infrastructure covering traces, evaluations, analytics, and production monitoring, with a hosted SaaS version. Why it fits: Founded in 2024 and focused on an infrastructure category where self-hosting is attractive but managed monitoring is easier to adopt.
10. Mem0 — Provides an open-source memory layer for AI applications, with APIs and hosted infrastructure for storing, retrieving, and managing user- and agent-level memory. Why it fits: Founded in 2024; the Mem0 Platform converts an open-source persistence primitive into a managed service for production AI systems.

White-space gaps noticed

• Open-source infrastructure for agent reliability: There is still limited startup coverage for durable execution, permissions, sandboxing, rollback, and multi-agent coordination as a cohesive cloud platform rather than as separate tools.
• Cloud-native data planes for AI applications: Open-source projects are emerging around memory, metering, and observability, but there is room for a neutral, managed data layer combining event storage, evaluation data, lineage, policy, and usage-based billing without locking customers into a model vendor.