Vercel

The same company that makes the framework your developers write in also runs the cloud it deploys to. That's the whole game.

What Vercel actually does

Vercel runs web applications so a team doesn't have to run servers. A developer pushes code to Git, and Vercel builds it, deploys it across a global network, and scales it up and down on its own. The older way meant hiring a DevOps team to wire up a content delivery network, serverless functions, build pipelines, and server-side rendering by hand, then babysitting them. Vercel turns that into one git push. Underneath, it sits on top of cloud infrastructure like AWS and adds the developer experience layer that makes it usable. The company also builds Next.js, the open-source React framework that has become the default way to build serious web apps, plus a growing set of AI tools: v0, which turns a text prompt into a working interface; the AI SDK, one way to call more than 60 AI models; and an AI Gateway that routes between them. The pitch is that building, scaling, and securing a modern app, including an AI app, happens in one place.

The analogy

The closest comparison is Stripe, and the parallel is specific. Before Stripe, accepting payments online meant months of work that only large companies could afford to do well. Stripe turned it into a few lines of code. Vercel's founder describes the same starting point for the web: Amazon, Google, and Facebook had internal infrastructure that let them build and ship fast, and everyone else was locked out. Vercel packaged that advantage, deploying and scaling a web app without a platform team, and sold it to any developer. The mechanism is identical in both cases: take a capability that used to require a dedicated internal team and turn it into something a small team can switch on in an afternoon.

Who they serve

Vercel serves the people who build web frontends, mostly teams working in React and Next.js. The paying base runs the full range. At one end, a solo developer or small startup pays per seat. At the other, large enterprises run production traffic on it: OpenAI, Perplexity, Under Armour, The Washington Post, Uber, and HashiCorp are named customers, and Next.js itself powers sites for Walmart and TikTok. The official audience and the paying audience line up cleanly here, both are React and Next.js teams. The growth story is the move upmarket: Vercel has hired a CTO of Security, earned automotive-industry security compliance, and partnered with the ad giant WPP to put its tools inside a global creative network. Enterprise is where the larger contracts are, and that's where the company is pushing.

Who shouldn't use it

A team that isn't on React or Next.js gets less of what makes Vercel special, since the tight framework-to-cloud integration is the point. A framework-agnostic host like Render or Railway, or raw cloud, may fit better. A media or video site that serves enormous amounts of traffic will run into Vercel's metered bandwidth and overage charges, where Cloudflare's unmetered model is cheaper at that scale. A team that's deeply committed to AWS and wants every resource living inside its own AWS account will prefer AWS Amplify, which keeps everything in-house. And a team that just wants the cheapest possible static hosting for a brochure site is paying for capabilities it won't use.

Sample customer stories

Real customer: Sonos. The audio company migrated to Vercel and Next.js and reported roughly 75% faster build times, the kind of improvement that changes how a team works day to day. Faster builds mean more preview deployments, and a preview for every code change lets developers see the real result before it ships instead of guessing. Vercel frames the win as developer efficiency and confidence, not just speed: the team ships more often because each change is easier to check.

Hypothetical example: a two-person startup building an AI chat product. They use v0 to generate the first version of the interface from a description, wire it to a model through the AI SDK so responses stream in as they're written, and deploy with one git push, no DevOps hire. When a cheaper, faster model comes out, they switch to it through the AI Gateway without rewriting the app. What used to need an infrastructure engineer now needs a weekend.

What only Vercel can claim

Vercel is built around a framework it owns. Because the same company makes Next.js and the cloud it runs on, the platform can read your code and provision the right infrastructure itself, something no other host can do, because no other host wrote the framework.

It owns the framework most of these apps are written in. Next.js was downloaded more than 500 million times in the past 12 months, more than from 2016 to 2024 combined, and it powers the frontends of Grok, Claude, and Cursor (Vercel). Why a buyer cares: the framework your developers write in every day is made by the same team that runs your deploys, so the defaults are tuned for it.

The framework and the cloud understand each other. Vercel calls this framework-defined infrastructure: one git push and the platform figures out what your app needs and optimizes for it. The payoff: there's no separate infrastructure to configure, so a small team ships like a big one.

There's a real flywheel. Next.js is free and open source, which drives enormous adoption; that adoption funnels deployments to Vercel; the revenue funds more framework development, which tightens the integration further. The result for a buyer: the tool keeps getting better at the exact thing you use it for, and it isn't going anywhere.

The AI tools share the same front door. The AI SDK reaches more than 60 models through one interface and crossed 3 million downloads a week, up from under half a million a year earlier, and the AI Gateway routes between providers with no token markup. The payoff: you can change which AI model you use without rewriting code or juggling API keys.

Breadth is optional, not a bundle. Next.js, v0, the AI SDK, Fluid Compute, Sandbox, and the security tools each stand on their own, so a team can start with hosting and add the AI pieces only when it needs them.

Why this is hard, and why it matters now

Durable structural shift: the web is moving from pages that people build to apps and agents that AI helps build and deploy. More than 30% of weekly deployments on Vercel are now started by AI coding agents, up more than 1,000% in six months. A team building an AI product needs infrastructure that does things the old static-hosting model never had to: stream long AI responses without timing out, run AI-generated code safely in isolation, and route across model providers that change every few months. That's a different job than serving a web page, and it rewards a platform built for it.

Current shift: AI models only got good enough to write and ship real software across 2024 and 2025. Vercel's own AI SDK went from under 450,000 downloads a week to more than 3 million in a year, and its v0 tool passed 4 million users. The category exists now because agents that build and deploy working software finally became viable, and the infrastructure to support them had to be built to match.

What people would use instead

Without Vercel, a React team falls back to assembling the same stack by hand on AWS, Google Cloud, or Azure: a content delivery network, serverless functions, a build pipeline, and server-side rendering, all configured and maintained by a platform or DevOps team. That's the painful old way, and it costs headcount. The lighter fallback is a competing host like Netlify or Cloudflare Pages, which gives a good deploy experience but without the Next.js-native depth that comes from owning the framework. For the AI side, a team would stitch together its own model SDK, gateway, and sandbox instead of getting them from one provider.

Competition

• Netlify, built around a smooth deploy workflow for static and content sites, also moving upmarket to enterprise.
• Cloudflare Pages and Workers, built around Cloudflare's huge global edge network and unmetered bandwidth, which is cheaper for traffic-heavy sites.
• AWS Amplify, built around keeping everything inside your own AWS account.
• Render and Railway, built around backend flexibility and more predictable pricing.
• The big clouds themselves, AWS, Google Cloud, and Azure, which keep improving their developer experience to make the abstraction layer matter less.
• On the AI-builder side, v0 competes with Bolt.new and Lovable, while coding agents like Cursor, GitHub Copilot, and Claude Code are increasingly the front door to deployment.

Cloudflare is built around its edge network; Netlify around its deploy workflow; AWS around its own ecosystem. Vercel is built around owning the framework most of these apps are written in, which is the one thing the others can't copy.

Website analysis

This is a buried-moat finding. The homepage leads with "Build and deploy on the AI Cloud." That's a category claim, and it's a contested one: Cloudflare, AWS, and the big clouds are all racing to own "AI" and "cloud," so the hero stakes ground several competitors are also fighting for. The single fact no competitor can claim, that Vercel makes Next.js and the cloud it deploys to, and that's why the platform can provision infrastructure straight from your code, is not in the hero. It surfaces lower down as "Framework-Defined Infrastructure," a phrase that names the mechanism but assumes the reader already connects it to framework ownership. The most defensible thing about the company is treated as a mid-page feature while the hero claims a crowded category. The fix is structural: promote the framework-and-cloud-ownership claim to the top, where the one un-copyable advantage belongs.

Website rewrite

This is a hero-and-subhead case: both are doing category work, not differentiation work.

• Current hero (verbatim): "Build and deploy on the AI Cloud."
• Current subhead (verbatim): "Vercel provides the developer tools and cloud infrastructure to build, scale, and secure a faster, more personalized web."
• Current CTA (verbatim): "Start Deploying"
• Rewritten hero: "We make the framework. We run the cloud it deploys to."
• Rewritten subhead: "Because Vercel builds Next.js and the cloud it runs on, one git push provisions and tunes your infrastructure. No DevOps required."
• Rewritten CTA: Keep as is. "Start Deploying" is the right low-commitment action for a self-serve developer audience.
• Reasoning: the original hero claims "AI Cloud," a category competitors are also claiming, so it doesn't separate Vercel from them; the rewrite names the one thing only Vercel can say. The original subhead is a generic infrastructure list ("build, scale, and secure"); the rewrite spends its words on the mechanism that ownership unlocks. This is a positioning hero, not an awareness one, the visitor already knows Vercel, so the page should claim the position rather than explain the company.

Messaging to consider

A gift to the team: three lines that are Vercel's alone to say.

1. We make the framework. We run the cloud it deploys to.
2. The cloud that reads your code.
3. One git push. We provision the rest.

Lead with the first. It passes ownability most cleanly: no competitor makes both a dominant framework and the cloud it runs on, so the line is literally true for Vercel and false for everyone else. The second re-categorizes Vercel out of the crowded "AI Cloud" space and onto the framework-ownership ground it owns alone.

Likely next questions a prospect would have

• How does pricing work at scale, and what do bandwidth and compute overages cost once I'm past the plan limits?
• After the April 2026 security incident, what's changed in how Vercel protects environment variables and internal access?
• How locked in am I to Next.js, and what happens to my apps if I want to leave?
• If my team isn't on React, how much of the framework-defined infrastructure benefit do I actually get?
• Does Vercel handle long-running background jobs and scheduled tasks now, or do I still need separate infrastructure for those?
• How does AI Gateway pricing compare to calling model providers directly, and is there really no token markup?

Sources

Company sources: https://vercel.com/home , https://vercel.com/blog/series-f , https://vercel.com/ai , https://vercel.com/fluid , https://vercel.com/frameworks/nextjs , https://vercel.com/blog/how-sonos-amplified-their-devex , https://vercel.com/kb/bulletin/vercel-april-2026-security-incident , https://vercel.com/about

Third-party sources: https://sacra.com/c/vercel/ , https://research.contrary.com/company/vercel , https://www.bleepingcomputer.com/news/security/vercel-confirms-breach-as-hackers-claim-to-be-selling-stolen-data/ , https://www.trendmicro.com/en_us/research/26/d/vercel-breach-oauth-supply-chain.html