Agentic Infrastructure
30% of Vercel deployments now initiated by coding agents—up 1000% in six months. Infrastructure is evolving to let machines deploy without humans.
Why it matters
As AI agents move from experimental to production-grade deployment engines, infrastructure platforms are fundamentally redesigning themselves for autonomous software delivery. This shift from human-gated to agent-native deployment surfaces represents a critical inflection point in how AI companies compete.
The key facts
6 to knowWeekly deployments on Vercel doubled in 3 months, driven by agents
30% of Vercel deployments now initiated by coding agents (up 1000% YoY)
Claude Code accounts for 75% of agent deployments, Lovable/v0 6%, Cursor 1.5%
Projects deployed by agents 20x more likely to call AI inference providers vs. human-deployed projects
Agents require new infrastructure primitives: long-lived execution, multi-step orchestration, model routing, cost controls, sandboxed execution
Vercel unifying AI primitives: AI SDK, Agent SDK, Chat SDK, AI Gateway, Fluid Compute, Workflows, Queues, Sandbox, Observability
Go to the source
Vercel Blogvercel.com
Publisher excerpt: Every generation of software eventually demands a new generation of infrastructure. LLMs and coding agents are driving the next transition, and it's happening fast. In just three months, weekly deployments on Vercel have doubled, and agents are driving the growth. Today, over 30% of deployments are…