Modal raises $750m Series D as demand for AI compute grows
The Accel-led financing values Modal at $15.75bn after the investment, five months after its Series C. Isolated environments for AI-generated code have become a substantial business alongside model training and inference.
Agent workloads expand the business
Modal says its Sandboxes product has reached a $300m annualized run rate, within a company-wide figure above $1bn. Those figures describe the current pace of revenue, rather than sales booked over a completed year. Usage has tripled since the preceding financing, according to the company.
Sandboxes give AI agents somewhere isolated to execute generated code. Their growth adds a second source of demand alongside the GPU workloads associated with training and serving models. Modal has also made virtual machines generally available for sandbox users.
From experiments to production
Accel describes Modal as a Python-based execution platform that handles provisioning and scaling across workloads with different CPU, GPU and networking needs. Its customer examples include DoorDash, Cognition and Oura, illustrating demand from enterprise teams as well as specialist AI companies.
Use of funds
Develop infrastructure for AI applications, including reinforcement learning, post-training, inference and isolated code execution.