How Modal is Developing AI-Native Infra for AI-Native Companies

Modal, an AI-infrastructure for developers, led by Erik Bernhardsson and Akshat Bubna, has secured $80 million in a Series B round, led by Lux Capital, with existing investors participating as well, bringing the total funds raised to $111 million.

Compute infrastructure is transforming similarly to electricity’s shift from big engines to grids. Early AI and cloud workloads strain existing systems, from GPU management to large models. A new platform addresses this by pooling global compute, enabling rapid iteration, low-latency performance, code-first programmability, and usage-based serverless pricing. Built from the ground up with custom file systems, runtimes, and schedulers, it now supports thousands of complex applications, delivering an unparalleled developer experience.

Modal’s foundation is its container platform and storage layer, supported by robust primitives that enable teams to build powerful AI applications. Modal’s product suite now supports diverse AI workloads, including inference for LLMs and generative models, secure sandboxes running tens of thousands of containers, massive parallel batch jobs, high-throughput training clusters, and collaborative notebooks with near-instant GPU starts.

The goal is to be the infrastructure provider for every stage of developing and running AI in production, as AI’s infrastructure demands grow increasingly complex. Modal empowers developers to deploy applications quickly, leveraging a deep bench of talent, with many new innovations planned in the coming years.