Modal Raises $355M Series C, Launches Agent‑Centric Cloud Platform
Explore Modal's new agent cloud features and evaluate migrating agent workloads to its elastic inference and sandbox infrastructure.
Set up a test agent on Modal to evaluate performance and cost.
Summary
Modal, a cloud platform focused on AI workloads, secured a $355M Series C round, positioning itself as a leader in the emerging agent cloud market. The platform offers elastic inference, sandboxes, GPU burst, background agents, and infrastructure that agents can operate, supporting 17 cloud providers and enabling multi‑node training and post‑training research. Key innovations include serverless functions, decorator‑based infrastructure, elastic inference for custom models across audio, video, robotics, and computational biology, GPU snapshotting, DeFlash, speculative decoding, and Auto Endpoints. Modal also introduced networked sandboxes, private IPv6, RDMA, and a supercloud strategy that pools capacity across providers.
The new features address the limitations of Kubernetes for bursty, compute‑heavy workloads and provide a developer‑friendly experience that supports fast iteration, debugging, and observability for agent workloads. Modal’s infrastructure is designed to handle large numbers of sandboxes, such as the 100,000 required for reinforcement learning rollouts, and to offer specialized guardrails for production agents.
Key changes
- Modal raised $355M Series C
- Platform offers elastic inference, sandboxes, GPU burst, background agents
- Supports 17 cloud providers and multi‑node training
- Serverless functions and decorator‑based infra introduced
- Elastic inference for custom models across audio, video, robotics, and computational biology
- GPU snapshotting, DeFlash, speculative decoding, Auto Endpoints added
- Networked sandboxes, private IPv6, RDMA, and supercloud strategy
- Designed to handle 100,000 sandboxes for RL rollouts