Stanford MS&E435 Economics of the AI Supercycle | Spring 2026 | Applications, Coding AI

Modern software development is transitioning from human-centric web applications to autonomous "agentic" systems, fundamentally changing cloud infrastructure requirements. Guillermo Rauch, founder of Vercel, explains that coding agents—which now handle tasks ranging from software maintenance to complex business operations—require specialized environments to deploy and scale effectively. This "agentic infrastructure" includes new primitives like AI gateways for token management and ephemeral sandboxes for compute. As AI models increasingly write, repair, and secure code, the traditional software-as-a-service model is shifting toward token-based pricing and highly tailored, malleable applications. By providing the foundational layer for these agents, Vercel enables developers to move beyond static web pages toward dynamic, self-optimizing systems. This evolution marks a significant expansion in the total addressable market for software creation, as AI-driven tools lower the barriers to deploying functional, scalable digital products.