Modal raises $750 million at a $15.75 billion valuation for its AI cloud platform

Modal is drawing fresh capital for the computing work beneath AI applications, from running models to executing code they generate. The company announced a $750 million Series D on October 14 at a $15.75 billion post-money valuation. Accel led the financing, which accompanied Modal’s claim that it had crossed a $1 billion annualized revenue run rate.

Modal’s basic proposition is to let developers write code while its platform handles the machinery needed to run it. Accel describes a Python-native workflow, meaning developers work in Python rather than separately managing the underlying computing resources. Modal handles provisioning—setting up those resources—and scaling, or adjusting capacity as workloads change.

The challenge is that AI applications do not produce just one kind of computing job. Transcribing audio, serving a language model and running an agent’s code can require different mixes of processors, storage and networking. Requests can also arrive in unpredictable bursts. Accel frames Modal as one execution platform spanning experiments and production applications.

Modal says it built its own file system, container runtime, scheduler and networking, with a capacity pool spanning more than 20 clouds. Its first building block, Functions, lets developers submit code for Modal to run. Uses include generating images and music, training models and inference: running a trained model to produce an output.

Where customers use the platform

• Company applications: Modal says DoorDash’s AI platform team uses it to move applications across the company from experiments into production.
• Background agents: Cognition, Ramp, Linear and Hebbia are building agents on Modal that work in their respective fields, according to the company.
• Robotics and research: Modal names Physical Intelligence, Generalist AI and 1X as robot-training users, and Chai Discovery and Achira as drug-discovery users.

Another part of the business addresses what happens after a language model writes code. Modal launched Sandboxes in 2023 to provide environments where that code can run separately from other work. The company lists background agents, reinforcement learning and AI-assisted coding platforms among the uses. Generating code and executing it are separate jobs in this setup.

Three years after that launch, Modal says Sandboxes alone has reached a $300 million revenue run rate. That gives the financing announcement a more specific business signal than overall AI demand: isolated code execution has become a substantial product within a platform that also runs models and training workloads.

In its announcement, Modal puts cumulative sandbox creation at close to 3 billion and says the service can scale to 1 million concurrent sandboxes. Those figures describe different things: environments created over time versus how many can operate at once. Neither is a count of distinct customers.

The company-wide $1 billion milestone is an annualized revenue pace, not a statement that Modal collected that amount over a completed year. Accel places the crossing in September 2026. It is also distinct from the usage growth above: a rise in computing activity and a revenue run rate are not interchangeable measures.

Alongside Accel, Kleiner Perkins joined the financing. Modal says all existing investors participated, including Menlo Ventures, General Catalyst, Redpoint, Lux Capital, Amplify Partners and Bain Capital Ventures. Accel’s lead investment deepens an earlier stake in the company, rather than marking its first backing of Modal.

Founder Erik Bernhardsson frames the new capital as funding more infrastructure building blocks for future software. His announcement describes a team of more than 200 across New York, San Francisco and Stockholm. The investment backs that platform-building strategy, with model execution and isolated code execution already supplying two distinct parts of the business.