Modal Labs Inc. has raised $750 million in late-stage funding, the artificial intelligence infrastructure provider announced. Accel led the Series D round, with Kleiner Perkins and about six existing Modal investors participating. The company said it is now valued at $15.75 billion, more than triple its valuation after its last round in May.
The jump reflects fast growth at the startup. Modal said its annualized revenue has more than tripled since the May round to more than $1 billion. Close to one-third of that revenue comes from Sandboxes, a cloud service that supplies hosted containers tuned for AI workloads.
Sandboxes is designed so developers can use its containers as model training environments and to run AI agents along with the code those agents produce. Modal said the service can provision tens of thousands of containers per second, an ability it attributes to a custom infrastructure management system, or scheduler. That scheduler spreads container provisioning requests across multiple servers in parallel, and Modal said it ran 1 million concurrent containers in a recent internal test.
When Sandboxes hosts AI agents, it sometimes connects them to a second container, called a sidecar, that carries auxiliary software, such as a tool that detects and filters hallucinations. Modal said it runs sidecars on the same physical server as the containers they support to cut latency.
A feature called Secret Injection lets agents log into external applications using login credentials without seeing them, which Modal said protects those credentials from rogue agents and cyberattacks. Modal sells Sandboxes alongside a cloud service optimized for AI training workloads that lets developers spin up graphics card clusters through relatively simple scripts, with the chips in those clusters exchanging data over RDMA.
Modal also competes in the inference market with Endpoints, which lets developers deploy AI models on hosted infrastructure and expose them through an application interface. Enterprises that do not want to build their own model can use a library of open-source algorithms from Modal. The company's services run on hardware rented from more than 20 cloud providers, so customer workloads can keep running on other clouds if one has technical problems.
Modal said it will put part of the new funding into product development and plans to launch services focused on reinforcement learning and post-training, phases of the AI development workflow meant to improve reasoning, safety and cost-efficiency.
More company and startup news from TechManNews.
