Vercel Revenue: How Vercel Grew to a $500M Run Rate

Vercel grew by making it much easier for developers to put software on the internet.

It started with a command that turned code into a live URL. Then Vercel created Next.js, an open-source framework that brought millions of developers into its ecosystem. Preview deployments made Vercel useful to entire teams, not just developers. A free plan let people adopt it without talking to sales. Enterprise plans captured the companies that grew around those users.

Then AI made the same model even more powerful.

Vercel launched v0 and the AI SDK. Coding agents began creating and deploying software themselves. By early July 2026, Vercel said it had crossed a $500 million annualized revenue run rate.

The growth engine can be reduced to this:

open-source tools → developer adoption → Vercel deployment → team collaboration → paid company usage → more open-source investment

AI added another loop:

AI SDK or v0 → app gets created → app gets deployed → infrastructure usage → Vercel revenue

Vercel did not win by starting with enterprise sales.

It won developers first.

Vercel's growth at a glance

It started with one command

Before Vercel, Guillermo Rauch had already spent years building for the web.

He grew up around Linux and open-source communities in Argentina. As a teenager, he became involved with projects such as MooTools. He later worked on Socket.IO and Mongoose.

"Open source really was my path out of Argentina in the end," Rauch told First Round.

That experience mattered later.

But another idea came from Cloudup, the file-sharing startup Rauch started before Vercel.

Cloudup turned files into links.

A user could drag a file into the app and immediately get a URL they could send to someone else. Even a folder containing a website could become something another person could open in a browser.

Automattic acquired Cloudup in 2013.

A few years later, Rauch applied a similar idea to software deployment.

Instead of turning a file into a URL, why not turn an application into one?

Rauch founded ZEIT with Tony Kovanen and Naoyuki Kanezawa in November 2015.

Its first major product, Now, launched in 2016.

A developer could open a project directory and type:

now

The application would be built and Vercel would return a live HTTPS URL.

That sounds normal today.

At the time, deploying an application often meant dealing with servers, deployment scripts, DNS, SSL, networking, and other infrastructure.

Now hid much of that work.

In his 2016 review, Rauch said the service had already handled more than 40,000 deployments and two billion requests.

The first Vercel growth insight was simple:

reduce the number of steps between writing code and showing it to someone.

That idea would shape nearly everything Vercel built next.

Next.js became Vercel's biggest distribution channel

ZEIT needed a framework for building its own web applications.

React handled the interface, but teams still had to make decisions about routing, server rendering, code splitting, and production setup.

Rauch has compared React to an engine.

Developers still needed the rest of the car.

The ZEIT team built an internal framework called N4. After roughly six months, it became Next.js.

Rauch made the initial specification commit on October 5, 2016.

Then the company made an important choice.

Next.js was open source.

Developers did not need to become Vercel customers to use it. They could run Next.js somewhere else.

This gave Vercel much more distribution than a closed product could have.

Developers could discover Next.js through GitHub, npm, tutorials, other developers, conferences, or companies already using the framework.

By the time Next.js 4 launched in 2017, Vercel said it had almost 18,000 GitHub stars.

By April 2020, Vercel said Next.js powered more than 35,000 sites.

Next.js became the top of Vercel's funnel.

The framework solved the building problem.

Vercel solved the shipping problem.

And because Vercel controlled both, it could make deploying a Next.js application unusually easy.

This created one of Vercel's most important growth loops:

more developers use Next.js → more Next.js applications are created → more developers need hosting → Vercel is the easiest option → Vercel can invest more in Next.js

The framework was free.

The infrastructure around successful applications was not.

That is what made the open-source strategy work as a business.

Preview deployments turned hosting into collaboration

The first version of Vercel helped a developer put an application online.

The next breakthrough made the URL useful before the application was finished.

ZEIT connected deployments to GitHub and GitLab.

Every code change could produce its own deployment.

Instead of saying:

"I changed the checkout page."

A developer could send a URL.

Anyone could open it and try the new checkout page themselves.

By 2019, Vercel's Zero Config Deployments could detect the framework, import a repository, build it, and automatically create deployments for pushes and pull requests.

There was no need for every stakeholder to install the project locally.

Designers could review it.

Product managers could review it.

Marketing could review it.

A founder could review it from their phone.

Rauch later told Acquired that he had underestimated how powerful this would become.

The key insight was that the URL was no longer just the final destination.

It became part of how software was built.

That widened Vercel's value from:

deployment infrastructure

to:

a collaboration tool for everyone involved in a website or application.

And it gave Vercel another distribution loop.

A developer created the deployment.

Then they sent the URL to coworkers.

Those coworkers experienced Vercel without needing to discover Vercel themselves.

Every preview became a small piece of product distribution.

The free plan let developers bring Vercel into companies

In April 2020, ZEIT changed its name to Vercel.

The announcement said more than 300,000 frontend engineers and designers were already using its technology and Next.js powered more than 35,000 sites.

The company also made its position clearer:

Develop. Preview. Ship.

Vercel's free Hobby plan was important to that strategy.

An individual developer could connect a repository and start deploying without getting approval from a manager, talking to procurement, or booking a sales demo.

This matters for developer products.

The person who discovers the product is often not the person with authority to buy enterprise software.

Vercel removed that problem at the start.

The developer did not need approval.

They could just use it.

Then a project could spread.

One developer started using Vercel.

Then another developer joined.

Then designers and product managers started opening preview links.

Then the project became important.

Eventually, the company needed more collaboration, security, support, controls, or infrastructure.

That was the moment Vercel could charge.

The growth path looked like this:

individual developer → project → team → company → enterprise

This is product-led growth in its simplest form.

The product entered the company before the salesperson did.

Developers opened the door. Enterprise sales expanded the account.

Free adoption did not mean Vercel ignored enterprise sales.

It meant sales arrived later.

By then, someone inside the company often already wanted the product.

That is a much easier sales conversation.

In June 2021, Vercel announced a $102 million Series C.

The operating numbers were more interesting than the funding.

In the previous six months:

• traffic across the Vercel Edge Network had doubled
• weekly deployments had quadrupled
• Next.js npm downloads had grown from 4.1 million to 6.2 million
• Vercel's enterprise customer base had tripled

Customers included GitHub, IBM, McDonald's, Facebook, HashiCorp, Patreon, and Scale.

Vercel had found a strong combination:

bottom-up adoption + top-down expansion

Developers chose the product because it made their work easier.

Large companies paid for things such as security, support, governance, controls, and scale.

Vercel's first salesperson, Luke Miller, has said that selling developer experience itself could be difficult.

The value became easier to explain when the discussion moved closer to business outcomes such as faster sites, higher conversion rates, or teams shipping faster.

That is another reason the preview workflow mattered.

Vercel was not only saving infrastructure work.

It was helping companies shorten the time between an idea and something customers could use.

Next.js gave Vercel distribution without forcing lock-in

There is an important detail in Vercel's open-source strategy.

Next.js did not require Vercel.

It used the MIT License and applications could be hosted elsewhere.

That made adoption easier.

A developer did not have to commit to a commercial cloud vendor just to try the framework.

Vercel instead tried to make its cloud the most convenient place to run Next.js.

This is different from putting the useful parts of the framework behind a paywall.

The open-source project could spread widely.

Vercel monetized the infrastructure around it.

There has still been debate about how portable some Next.js features are in practice. Specialized caching and framework behavior historically made it easier for Vercel to offer the complete experience than other hosting providers.

Next.js 16.2 improved this in 2026 through a stable Adapter API, compatibility tests, and work with other infrastructure providers.

But the broader growth lesson remains useful:

Vercel did not need every Next.js user to pay.

It needed Next.js to become important enough that a meaningful share of successful projects eventually needed infrastructure.

Vercel repeated the same playbook with AI

The rise of generative AI could have been a threat to Vercel.

If AI changed how software was built, a platform designed around human developers might become less important.

Vercel went the other direction.

It started building tools for the new workflow.

In June 2023, Vercel released the open-source AI SDK.

The goal was to make it easier for JavaScript and TypeScript developers to work with AI models.

Once again, Vercel put a useful developer tool at the top of the funnel.

And once again, the tool naturally connected to infrastructure Vercel could sell.

The AI SDK grew quickly.

By June 2026, AI SDK 7 had reached more than 16 million weekly downloads.

The pattern looked familiar:

open-source SDK → developers build AI apps → apps need infrastructure → Vercel provides the infrastructure

Vercel had rebuilt the Next.js growth model for AI.

v0 brought non-developers into the funnel

Vercel also created a new product called v0.

The first version generated user interfaces from text prompts.

Instead of writing the interface by hand, someone could describe what they wanted.

v0 would generate it.

The product got attention immediately.

Its beta announcement said 100,000 people joined the waitlist in three weeks.

As AI models improved, v0 expanded from interface generation into a tool that could build much more complete applications.

The economics also improved quickly.

Rauch told Acquired that v0 took around ten months to reach $1 million ARR.

It then reached $2 million roughly 14 days later.

And $4 million another 20 days after that.

By February 2026, Vercel said more than 4 million people had used v0 since its general availability in 2024.

But the most interesting part is not v0's standalone revenue.

It is how well v0 fits the rest of Vercel.

A user describes an application.

v0 creates it.

The application needs somewhere to run.

Vercel is already there.

The new loop becomes:

prompt → application → preview → Vercel deployment → infrastructure usage

The interface for creating software changed.

The destination stayed the same.

Coding agents multiplied Vercel's core growth loop

AI did something even more important to Vercel's business.

It increased the amount of software being created.

A human developer can only write and deploy so much code.

A coding agent can work much faster and can run many tasks in parallel.

That means more code changes.

More previews.

More builds.

More deployments.

And more infrastructure usage.

Vercel was already built around a world where every code change could become a deployment.

Coding agents made that behavior happen much more often.

In early 2026, Vercel said deployments initiated by agents were rising quickly.

Its April 2026 post on agentic infrastructure said weekly deployments had doubled in three months and coding agents were responsible for more than 30% of deployments at the time.

Vercel later said agent-driven deployments had passed 50%.

The exact figures Vercel published during this period have varied, so they should not be treated as a clean historical series.

But the direction is clear.

Coding agents are creating a large and growing share of deployments.

That is unusually good for Vercel because deployments are already at the center of its business.

AI did not force Vercel to invent a completely different growth engine.

It increased the speed of the one Vercel already had.

When an AI agent picks the tool

This matters even if you don't sell to developers. More buying decisions now start with an AI assistant recommending a product, and it recommends what it has read about. Okara's GEO Agent shows whether you appear in those answers for your buyers' prompts, and the Coding Agent ships the technical fixes as GitHub pull requests.

The revenue followed the usage

Vercel's revenue grew as the developer workflow spread.

Rauch has put Vercel's revenue at roughly:

• $1 million in 2019
• $5 million in 2020
• $50 million by June 2023

Reuters reported that Vercel had passed $100 million in annual revenue around its May 2024 Series E.

In June 2025, CFO Marten Abrahamsen said Vercel had crossed $200 million ARR after roughly doubling revenue in 15 months.

Forbes later reported a roughly $340 million run-rate GAAP revenue pace at the end of February 2026.

Then came another jump.

On August 6, 2026, Vercel said it had crossed $500 million in annualized run-rate revenue in early July.

That wording is important.

It does not mean Vercel had already recorded $500 million of revenue during a completed year.

An annualized run rate takes the current revenue pace and projects it across twelve months.

Vercel has not published audited 2026 annual financial statements.

So the accurate description is:

Vercel reported a $500 million annualized revenue run rate in July 2026.

According to the company, growth at the time was being driven by more Vercel Pro sign-ups and enterprise adoption of its Agent Stack.

Vercel's growth engine

The individual tactics matter less than how they connect.

Vercel built several products that feed the same system.

1. Open source creates discovery

Next.js and later the AI SDK give developers useful products they can adopt without paying Vercel.

That creates a large top of funnel.

2. The product makes the next action easy

A Next.js developer needs to deploy the application.

An AI SDK developer needs infrastructure.

A v0 user needs somewhere to run the app they created.

Vercel is positioned as the natural next step.

3. Every deployment can spread inside a team

Preview URLs get shared with engineers, designers, product managers, founders, marketers, and clients.

One user's deployment exposes more people to the workflow.

4. Free usage removes procurement

Developers can start before their company has made a buying decision.

That lets Vercel enter companies from the bottom up.

5. Successful projects create the reason to pay

Once projects become commercial or important, customers need more usage, collaboration, security, governance, reliability, and support.

Vercel monetizes the projects that grow.

6. Enterprise sales expands what developers started

Sales does not always need to create demand from zero.

In many cases, developers have already introduced Vercel to the company.

Sales helps turn that internal adoption into a larger contract.

7. AI increases the amount of software being created

v0 makes software creation accessible to more people.

Coding agents increase how much code each person can ship.

Both lead to more deployments and infrastructure usage.

Put together, the loop looks like this:

useful developer tool → more apps → easier deployment → more preview links → more team adoption → paid infrastructure → more investment in developer tools

That loop has survived several changes in how software is built.

It started with developers typing commands.

Now agents are doing some of the typing.

Why Vercel's strategy worked

A lot of companies use open source.

A lot of companies offer free plans.

A lot of companies have developer communities.

Those things alone do not explain Vercel.

The important part is how close Vercel's distribution products are to its paid product.

Next.js does not attract a random audience.

It attracts people building web applications.

Those people eventually need deployment infrastructure.

The AI SDK attracts people building AI applications.

Those applications need inference, compute, observability, sandboxes, deployments, and other infrastructure.

v0 creates applications.

Those applications also need infrastructure.

Vercel repeatedly gives people something valuable before asking them to buy anything.

But the free product and paid product are not separate businesses.

Success with the free product naturally creates demand for the paid one.

That is the core of Vercel's growth.

What founders can learn from Vercel

The biggest lesson from Vercel is not "build an open-source project."

It is to build a free product or artifact that moves users closer to the thing you sell.

For Vercel, that artifact was often a URL.

A developer ran a command and received a URL.

A Git branch became a URL.

A pull request became a URL.

A v0 prompt became a working application with a URL.

An AI agent writes code and creates another deployment.

Each step moves the user closer to consuming Vercel's infrastructure.

That is much stronger than generating attention and hoping some of it turns into customers.

The distribution mechanism is part of the product.

More growth stories: how Linear, n8n and Zapier grew.

Frequently asked questions

How did Vercel grow?

Vercel grew by simplifying deployment, using Next.js as an open-source distribution channel, making every code change shareable through preview deployments, and letting developers start for free. Developers brought Vercel into companies from the bottom up, while paid Pro and Enterprise plans monetized commercial usage. More recently, v0, the AI SDK, and coding agents created another wave of growth.

What was Vercel called before Vercel?

Vercel was originally called ZEIT. The company changed its name to Vercel in April 2020 as it focused more clearly on helping teams develop, preview, and ship web applications.

Who founded Vercel?

Vercel was founded in 2015 by Guillermo Rauch, Tony Kovanen, and Naoyuki Kanezawa.

How did Next.js help Vercel grow?

Next.js gave Vercel a large open-source distribution channel. Developers could use the framework for free and host it anywhere, but Vercel built a hosting platform closely optimized for the framework. As Next.js adoption grew, the number of developers who could naturally become Vercel users grew with it.

How does Vercel make money from open source?

Vercel does not need to charge developers to use Next.js or the AI SDK. It makes money from the infrastructure, collaboration, security, support, and enterprise features needed to run applications at scale.

What is a Vercel Preview Deployment?

A Preview Deployment is a live version of an application created from a code change such as a Git branch or pull request. It gives the change its own URL so teammates can test and review it before it reaches production.

How did v0 help Vercel grow?

v0 lets users create applications from natural-language prompts. Those applications can then be deployed to Vercel. This creates a direct path from AI-generated software to Vercel's paid infrastructure.

How are coding agents helping Vercel?

Coding agents generate and change software much faster than humans working alone. Those changes need builds, previews, tests, sandboxes, and deployments. Vercel already provides that infrastructure, so more agent activity can lead directly to more platform usage.

What is Vercel's revenue?

Vercel said it crossed $500 million in annualized run-rate revenue in early July 2026. This represents its revenue pace projected over twelve months, not $500 million of completed-year recognized revenue.

Is Next.js locked to Vercel?

No. Next.js is released under the MIT License and can be self-hosted or deployed on other platforms. Some Next.js features have historically been easier to operate fully on Vercel because of specialized infrastructure. Next.js 16.2 improved portability through a stable Adapter API and shared compatibility work.

Sources

• Vercel's Path to Product-Market Fit, First Round
• Guillermo Rauch's 2016 review, Guillermo Rauch
• Next.js initial specification, GitHub
• ZEIT is now Vercel, Vercel
• Vercel founder interview on previews, revenue, and v0, Acquired
• Vercel Series C and operating metrics, Vercel
• Vercel AI SDK, Vercel
• v0 beta announcement, Vercel
• Vercel Series E, Reuters
• Vercel Series F, Reuters
• Next.js portability changes, Next.js
• Next.js Middleware bypass postmortem, Vercel
• Vercel's reported $500M annualized run rate, Business Wire