No one is an individual contributor anymore.
That is what Guillermo Rauch, the CEO of Vercel, said at a conference in San Francisco last week. Not everyone will be a manager. Everyone already is.
Rauch is not a fringe figure. He runs a company last valued at $9.3 billion, with $340 million in annual recurring revenue and a sales team that, as of last year, consisted of one human managing a autonomous bot. The rest of the positions had been eliminated.
Vercel is a web hosting and deployment platform that competes with Amazon Web Services and Cloudflare, except its primary customers are frontend developers and, increasingly, AI agents. The case for Vercel as an AI infrastructure play rests on a single data point that Rauch shared at the HumanX conference: 30 percent of the applications running on its platform are now created by AI agents, up 1,000 percent from six months ago. That is not a forecast. It is a measured shift in who is actually using the product, right now.
The clearest evidence of what this means in practice is the sales team. Vercel had a 10-person sales organization. It replaced it with one person and a bot. The bot handles the relationship; the human oversees the bot. Rauch frames this as an upgrade: when you no longer have individual contributors executing discrete tasks, everyone becomes a manager of the system doing the work. The title does not disappear. It changes meaning.
This is the transition Rauch has been describing for months. At a different event earlier this year, he said AI agents are turning everyone into what he called a mini CEO, someone who sets direction for autonomous systems rather than executing tasks themselves. The language is deliberately provocative, and Rauch knows it. But the headcount math underneath it is not spin.
The agent-initiated deployment data is where the argument becomes structural rather than anecdotal. Vercel's own blog, published four days before the conference, disclosed that more than 30 percent of its deployments are initiated by coding agents, a figure that grew 1,000 percent in six months. Of that agent traffic, 75 percent comes from Claude Code, the Anthropic tool that lets developers delegate implementation to an autonomous agent. Smaller shares come from Cursor and v0, Vercel's own vibe coding tool. What this tells you is that Anthropic is winning the developer tooling race by a wide margin, and Vercel is the infrastructure layer that benefits regardless of which model provider wins the foundation model battle.
That is the part of the story that does not get said enough. Every AI summit debates which model is best. Vercel is the equivalent of being the company that sold picks and shovels during every gold rush simultaneously, and the gold rush happens to be automated.
The IPO context is where the story gets complicated. The pipeline for new public listings was expected to open in 2026 after two years of drought. It has not. A broad sell-off in software stocks, driven by investor anxiety about AI disruption, has frozen most go-public ambitions outside of a handful of names like SpaceX, Anthropic, and OpenAI. Vercel is signaling it is ready to list anyway, framing itself as what Rauch calls a working public company, one that already operates with the discipline and disclosure standards of an SEC-filed entity even before it files. Whether public market investors will value a company whose primary growth driver is agent-initiated rather than human-initiated activity is the open question. Nobody has priced that before.
The skeptical case is straightforward. Vercel is reporting its own agent usage figures at a moment when it has every incentive to look as agent-native as possible. The human-to-bot sales replacement ratio is a single data point from a company that employs fewer than 500 people. The 30 percent figure does not tell you what portion of revenue comes from agents versus human customers who pay for more seats. It is possible that agent-initiated deployments are a leading indicator of genuine growth and that the trend compounds. It is also possible that the numbers are being used to tell a story that serves the IPO narrative more than it describes a durable shift.
What Rauch is describing, whether or not the numbers hold, is a version of work where the individual contributor, as a category of employment, ceases to exist. Your AI agent is the worker. You are its manager. That transition, if it plays out as he describes, changes the economic relationship between technology companies and the humans who use them in ways that have not been fully thought through. Vercel's IPO, whenever it happens, will be the first time public market investors are asked to value a company that sells that proposition at scale.
