Modal's rise: sandboxes, cold starts, and what the SEC paperwork reveals (henrymichaelson1.substack.com)

Modal, an emerging AI infrastructure company, has gained attention for simplifying and scaling the process of building AI applications. Founded by Erik Bernhardsson, a former Spotify engineer, and Akshat Bubna, a lead engineer at Scale AI, Modal aims to make cloud computing as effortless as local execution. Its platform offers composable execution primitives, such as customizable sandboxes for running code, elastic workload management, and a focus on ownership over deployed models—all designed to streamline the developer experience and eliminate the complexities associated with traditional systems like Kubernetes. The significance of Modal's innovations became evident when its infrastructure effectively managed an explosive demand surge from the AI application builder Lovable, which created 250,000 applications in a single weekend without requiring pre-provisioned capacity. Modal's custom architecture, including a unique filesystem for rapid cold starts and a GPU memory snapshotting feature, positions it as a critical player in the AI/ML landscape, particularly for companies navigating the fragmented GPU supply market. Recent SEC filings also highlight strong investor confidence and strategic partnerships, further solidifying Modal’s growing influence in the AI infrastructure sector as it prepares to host a conference focusing on AI technologies in October 2026.