Claude Code Dominates Vercel Deployments

This reveals that Vercel is becoming infrastructure for whichever coding agent already owns the developer’s attention, and right now that market is concentrating fast. Once an agent can write code, run tests, push a deploy, and hand back a live URL, the IDE matters less than the agent doing the work. That is why Vercel is seeing agent initiated deployments consolidate around Claude Code, with Cursor contributing only a small share despite its scale as an AI IDE.1 2 3 4

• On Vercel, agents moved from less than 3% of deployments in late 2025 to more than 50% by June 2026. Within that agent traffic, Claude Code drove 75% of deployments in April 2026, while v0 and Lovable together were only 6%, showing a winner take most pattern at the agent layer.1
• Cursor is a large business, estimated at $4B annualized revenue by May 2026, but its product started as an AI native IDE. Claude Code started as a terminal agent that can autonomously edit, test, and debug code, which makes it better suited to own the full path from prompt to deploy.2 4
• This concentration is spilling into adjacent infrastructure. Mintlify saw agent traffic rise from about 15% in January 2026 to 66% by July, as Claude Code, Codex, Replit, Lovable, Cursor, and Vercel increasingly choose tools and APIs themselves. The deployment layer is becoming part of a broader agent selected stack.4

From here, the platforms that win will be the ones easiest for the top agents to operate end to end. That favors Vercel expanding beyond hosting into workflows, sandboxed code execution, backend services, and agent specific auth, because the dominant coding agents are starting to behave less like editor features and more like autonomous software operators.1 3