Summary: Vercel CEO Guillermo Rauch said the company is ready for an IPO without naming a timeline, with ARR rising from $100 million in early 2024 to a $340 million run rate by the end of February 2026. He said agents have already built 30% of platform apps, making Vercel a deployment destination for AI-generated software. Competition from larger clouds and the state of the IPO market remain the open constraints.
TechCrunch’s report dated April 13, 2026 carries the news: Vercel CEO Guillermo Rauch took the stage at the HumanX conference in San Francisco the week prior and put the decade-old developer platform forward as ready for a public listing — “There’s no perfect timeline or quarter I can give. The company’s ready and getting more ready for it every day.” The confidence rests on a growth curve few companies can show: ARR stood at $100 million in early 2024 (The Information’s reporting) and had reached a $340 million run rate by the end of February 2026 (as Forbes reports).
Where that growth originates is the more interesting story. By Rauch’s account, agents have already built 30% of the apps on the platform, and agents are “very prolific at deploying.” Software companies at large are busy worrying that AI will make customers buy less software; Vercel instead made “everyone is producing software” the engine of its own revenue.
The Message at HumanX
Software stocks are selling off on AI fears and the IPO pipeline is frozen, so listing plans are a subject most CEOs dodge. Rauch went the other way and declared publicly that his company is set — a rare move in its own right. He frames Vercel as already “very much a work-in-public company,” run to the standards of a public cloud infrastructure business. TechCrunch adds the backdrop: the AI listings the market entertains as candidates to reopen the window are SpaceX, Anthropic, and OpenAI — that’s the entire list.
The Figures: An ARR That More Than Tripled
Line up the three data points: ARR near $100 million at the start of 2024; a $300 million Series F closed in September 2025, led by Accel, at a $9.3 billion valuation; a $340 million annualized run rate by the end of February 2026. In a little over two years the figure more than tripled — and it got there with no foundation model and no compute business, carried by a developer platform that now hosts AI-generated applications.
The competitive setup reads just as plainly: on one side sit giants of the Cloudflare and AWS caliber; on the other, Vercel sells v0, a vibe-coding tool of its own, and so stands on the supply side of “software written by AI.” It is, in effect, placing money on both ends at once — more people building apps, and agents building apps for them.
The Engine: Agents Turning Out Software by the Batch
The number Rauch put up from the stage: agents have already created 30% of the apps on Vercel. The reasoning behind it: agents will speed up software production because “generating custom solutions becomes easier than buying existing software” — a frontal challenge to the SaaS business model, and an opening for infrastructure. “All of that software … it needs to go somewhere, and we think it’s going to be Vercel.”
On the size of the prize he was blunter still: “The total addressable market of infrastructure has now grown, and it simply has no ceiling.” In the company’s early days only tens of millions of people could deploy an app; today “everybody in the world can create an app.” Our opening-of-year outlook flagged the same 2026 storyline: the cost structure of producing software is being rewritten.
The Window Ahead — and What Could Spoil It
The bull case runs like this: whichever model or agent ends up writing the software, the demand for deployment and hosting lands on Vercel, and the wave of AI-generated apps shows up straight in its billings. The risks are just as tangible. One, quality across agent-built apps is all over the place; should a big chunk prove to be one-off, low-retention micro-apps, revenue quality thins out. Two, Cloudflare and AWS fight this same market on price and scale, and whether Vercel’s premium survives is unsettled. Three, the reopening of the IPO window rides on how the market feels about AI capital spending — a dial no company gets to turn.
The note to carry back to product and platform teams: once agents, not people, are producing the software, deployment becomes a business that scales. The player that makes it least effort for agents to push things live takes the first layer of margin off this supply wave.