On October 14, 2026, Modal, which runs serverless infrastructure for AI inference, training and code sandboxes, announced a $750 million Series D at a $15.75 billion post-money valuation. Accel led the round and Kleiner Perkins joined; existing investors including Menlo Ventures, General Catalyst, Redpoint, Lux Capital, Amplify Partners and Bain Capital Ventures all participated.
Modal said it “recently crossed $1B in annualized run rate,” that usage has tripled since its Series C about five months earlier, and that it has grown more than 15 times year over year. Its Sandboxes product, launched in 2023 for running code generated by language models, is now “a $300M run rate product just on its own,” used for background agents, reinforcement learning rollouts and vibe-coding platforms. The company says it has spun up close to 3 billion sandboxes and that its capacity pool now spans more than 20 clouds.
Modal built its own file system, container runtime, scheduler and networking from the ground up, running them on capacity pooled across many cloud providers rather than a single cloud.
Why it matters: agents and reinforcement learning need huge numbers of short-lived, isolated computers, and that demand is creating a new layer of infrastructure companies between the hyperscalers and the AI developers. A sandbox product reaching $300 million run rate in three years shows how fast agentic workloads are turning into real spend.
What it does not show: run rate is an annualized snapshot, not audited revenue, and Modal does not disclose margins. Much of its capacity is rented from other clouds, so its economics depend on GPU prices it does not control.