Vercel: The Red Hat of the frontend
How to make billions selling open source software to enterprises
I’ve never understood Vercel as a business. They are wildly successful (just raised at a $9b valuation) but I don’t know where that success came from. Multiple points have always been a mystery to me:
At their core, Vercel is a frontend hosting platform, but there are a bunch of similar ones like Netlify, DigitalOcean, AWS along with no-code tools like WordPress, Webflow, and Bubble. How does Vercel manage to stand out? They weren’t the first ones to make deploying and hosting web apps easy.
Even if they did stand out compared to the last generation of competitors, how do they stand out against the next generation? The AI app builders like Lovable, Bolt.new, and Replit?
The answer is probably Next.js which is cool, but it’s a big expense, a lot of work to evangelize and convince people to use, and there are plenty of other frontend frameworks that work well enough. Why did this beat React + your state management of choice, Vue, or any of the other options?
Also, they keep acquiring (or creating) open source projects like NuxtLabs, Tremor, and Turbopack but these seem to just increase expenses and only have marginal connection to their core product (hosting) that makes them money. Why?
As someone who uses Vercel every day, this was a lot of question marks. It wasn’t until recently that it finally clicked: Vercel is the Red Hat of frontend.
What is Red Hat?
Red Hat is a software company that builds and provides open source software. They are largely associated with Linux having created one of the first distributions in Red Hat Linux in 1993.
Since they couldn’t sell Linux licenses (it is free), their version takes thousands of packages, freezes the code, and creates an enterprise-focused version. This version provides the long-term stability, security, and compatibility enterprises wanted, along with a promise of support (for a subscription).
This enabled enterprises like banks, airlines, governments, and telecom companies to get the functionality of Linux with less risk.
Beyond its business model, Red Hat has built capabilities around developing and supporting open source communities. It is one of the top contributors to Linux and has catalyzed several other open source projects.
For example, Red Hat has acquired and developed open source projects for automation (Ansible) and orchestration (OpenShift). This was especially important as the world moved from on-prem to virtualization, cloud, and containers.
If you’re wondering if this path is a good one, yes. Red Hat was acquired by IBM in 2018 for $34 billion. It had annual revenue of $6.4b last year, employs 19,000 people, and is a core part of IBM’s business now.
The trouble with the Red Hat model
In 2014, a16z partner Peter Levine wrote that there would never be another Red Hat. In 2015, Dan Woods added to this saying:
“Will any company ever get to $1 billion or even $500 million in revenue from open source subscriptions and have a chance of being profitable? To this the answer is no. There will never be another Red Hat.”
Between the two of them, their argument was:
No other open source company could have a product as widely applicable.
The support-driven business model of Red Hat sucks and is unable to support adequate funding of ongoing investments. Anyone else trying to sell services for open source software won’t make it.
Linux is a unique piece of software that companies didn’t need to be convinced to buy and Red Hat was in a uniquely good position to sell because of their first-mover advantage and reputation.
Although these are all fair points, their argument reveals the problem with “saying never” about any situation. History may not repeat, but it does rhyme. In this case, Red Hat rhymes with Vercel.
How Vercel rhymes with Red Hat
Far from “another Red Hat never existing,” Vercel has seemingly followed a similar path intentionally or not. Both:
Built an enterprise version of a popular open source project that has strong adoption. For Vercel, Next.js is basically an enterprise version of React, which is the most popular frontend framework.
Monetized it by providing the best services around that open source product. Developed the expertise to provide services for that open source technology in the best possible way.
Bolstered their reputation and improved their offering by developing and acquiring other open source products.
Although Peter Levine was right that the “support” business model probably sucked, this doesn’t mean an open source product can’t be a successful one. If you adjust it slightly to be “adjacent services” like hosting, it seemingly works well. Vercel, with a reported $200M in ARR, is well on its way to the “impossible” $500M amount Red Hat hit.
Looking at Vercel through the lens of Red Hat cleared up a lot of my confusion about why it was successful.
None of the other frontend hosting platforms or AI app builders take enterprise as seriously as Vercel does. Intentionally or not, others appeal much more to vibe coders and hobbyists.
Although builders of all kinds use Next.js, it can be seen primarily as a product built for enterprises. They make this pretty obvious with their focus on performance and security on their homepage.
Next.js builds on React’s popularity and Vercel doesn’t need to carry the entire development load. They benefit a huge amount from all the investment in the React ecosystem (from Meta and others). Vercel, through Next.js, can capture a significant amount of benefits from this.
The acquisition and development of other open source projects is because it improves their reputation for caring about performance and developer experience while strengthening their core offering of “enterprise React.” For example, Turbopack is all about improving bundle speeds for large (AKA enterprise) apps.
A collision course in the age of AI?
The similarities don’t end in both companies’ histories. They both face similar circumstances in the age of AI. Simply, they both care a lot about it, and it shows with all the AI buzzwords they plaster all over their website.
Neither Vercel nor Red Hat are in a position to develop their own models (although Red Hat’s parent company, IBM, is) and doing this doesn’t fit with their enterprise open source model anyway. Instead:
Vercel aims to become the “AI Cloud.” They explain this on their site as helping companies “deploy AI in seconds. Access all major models through a single, unified interface and shared AI credit wallet.”
Red Hat aims to let enterprises “tune small models with enterprise-relevant data, and develop and deploy AI solutions across hybrid cloud environments.”
Basically, they both are aiming to be the middle layer between model providers and enterprise apps. This makes sense as enterprise apps are already hosted on their respective platforms, making it easier for them to move up the value chain.
What this also means is that they are on a collision course. This middle AI layer is going to end up somewhere and it’s going to be valuable. Vercel, Red Hat, and many other companies (like OpenRouter) are battling for it. Right now, there likely isn’t much overlap, but as the contracts get larger, there will be. It will be interesting to see how a potential battle between such similar companies plays out.