Vercel vs Netlify 2026: which to choose for your site

Vercel vs Netlify in 2026: DX, real pricing, serverless limits and hidden costs. My verdict after hosting SaaS Radar on Vercel.

TL;DR

Take Vercel if your project runs on Next.js and iteration speed matters more than your bill. Take Netlify if you want a framework-neutral platform with forms, identity, and functions included in the plan, plus a more permissive free tier for marketing sites. Both do the same core job: a git push, a build, a global CDN. And both get expensive at exactly the same moment, when traffic spikes.

Vercel and Netlify: two ways to sell the same deployment

Both platforms take your Git repository, build the site, serve it from a global CDN, and run dynamic code via serverless.

The workflow is identical on both sides. You connect GitHub, GitLab, or Bitbucket, push a branch, the platform builds and publishes a preview URL. You merge, it goes to production. This workflow hasn't changed in five years, and that's a good thing: it works.

The divergence is elsewhere, in the origins. Vercel was built around Next.js, which the company develops and funds. Netlify was born from Jamstack and claims neutrality: any generator, any framework, from SvelteKit to Hugo to a simple Vite build.

This difference in DNA explains almost everything else. At Vercel, the most advanced Next.js features arrive on the platform first, sometimes before being documented elsewhere. At Netlify, the effort goes into broad compatibility and adjacent building blocks that Vercel doesn't provide.

What real differences exist between Vercel and Netlify

On paper the two offerings look alike, but four gaps actually change the decision: framework, included extras, pricing, and lock-in.

The free plan point deserves emphasis because it catches a lot of people. Vercel's Hobby plan prohibits commercial use. A site that displays an affiliate link or sells anything must move to Pro, period. I switched for this reason, not a technical limit.

Vercel or Netlify for a Next.js project

If you're doing Next.js with App Router, ISR, and tag-based cache, Vercel has a real lead and you'll feel it in production.

It's not magic, it's integration. Incremental static regeneration, tag-based revalidation, server component streaming: all of it works on Vercel without you writing a line of config. Netlify caught up a lot of the gap with its Next.js adapter, but you stay a step behind on recent features, and debugging exotic cache behavior is lonelier.

The reverse is also true. If you're on Svelte, Astro, Nuxt, or a static site generated by an in-house script, Vercel's advantage evaporates. You're then paying for Next.js integration you're not using.

A word on lock-in, because the subject always comes up. Next.js is self-hostable, and the official docs describe the procedure. But the more advanced primitives you use, the more porting requires real work: cache adapter, image handling, revalidation. It's not a wall, it's a slope.

Vercel vs Netlify: what these two actually cost

Both show around $20 per user per month for the first paid tier, and in both cases that's just a floor.

The real cost is usage. Vercel charges for compute in active CPU units, network egress, and function invocations. Netlify mostly charges for bandwidth and build minutes. Two different models, same outcome: a low-traffic site pays the subscription and almost nothing else, a site that takes off sees the usage line explode before you understand why. Check current grids on Vercel's pricing page and Netlify's, they move often.

The practical difference is in the per-seat breakdown. Both count by team member. Solo, it's painless. At five people, you're paying one hundred dollars a month before serving a single request, on either platform.

My advice: don't choose based on advertised price. Estimate your monthly egress volume and function calls, then simulate. Both platforms have a simulator, and it lies less than the marketing pages.

There's one asymmetry that few comparisons mention. Vercel now charges for compute in active CPU time, not total request duration. A function waiting for a database response costs you almost nothing during the wait. Netlify still mainly reasons in invocations and bandwidth. Depending on whether your routes spend time waiting or calculating, the ranking flips.

The practical corollary: measure before migrating. A site serving cached HTML with few dynamic calls will cost about the same either way, and you'll have wasted two days for nothing.

Functions, builds, and limits: where it breaks

The real limits aren't in price, they're in execution duration, bundle size, and concurrent builds.

On functions, Vercel offers classic Node functions and a restricted but very fast Edge Runtime. Netlify offers the equivalent with its Edge Functions based on Deno. For typical use—lightweight API, webhook, on-demand rendering—both hold without issue. Vercel's documented limits are worth reading before you commit to an architecture.

Where it gets complicated is long-running work. Multi-minute processing, massive import, report generation: neither is built for this. Move that work out of the platform and into a queue, worker, or external cron. I've seen too many people try to fit a batch into a serverless function and end up with intermittent timeouts they can't reproduce.

On builds, Netlify counts minutes and Vercel counts concurrency. A monorepo with ten apps gets expensive fast on both. A well-configured build cache, or clean npm workspaces, changes your bill more than platform choice.

Third trap: function bundle size. A heavy dependency accidentally bundled into an API route fattens the deployed package, slows cold start, and eventually hits a limit. Check what you're importing on the server side, especially SDKs that pull half an ecosystem with them. It's the most common bug I see on these two platforms, and it never shows up locally.

Last concrete gap: environment variables. Both manage scopes per environment—production, preview, development. The classic mistake is checking only production, then watching all pull request builds fail without understanding why. Happened to me, twice.

What I measured hosting SaaS Radar on Vercel

SaaS Radar runs on Vercel in the Paris region from day one, with Cloudflare in front for edge cache, and I logged the numbers.

In August 2026, the bill broke down to twenty dollars for the Pro subscription and about two dollars fifty in actual usage. In other words, I'm paying eight times what I consume, solely because the free plan bans commercial use. This is the kind of detail no comparison tells you, and it decides things in practice.

Second lesson, more useful: the actual cost wasn't Vercel. It was the database running 24/7 next to it, and the network egress it generated. Putting a CDN in front of public HTML with a long TTL and tag-based purge did more for the bill than any hosting arbitrage. If you're hesitating between Vercel and Netlify to save twenty bucks, you're looking at the wrong line item.

Third point: DX. Vercel's branch previews and real-time logs saved me from several production mistakes. Netlify does the same thing, honestly just as well. On this specific terrain, I don't pick a winner.

What profile should choose what

The choice comes down to three very distinct profiles, and for each there's only one right answer.

You're doing Next.js in a product team, with ISR and permanent previews: take Vercel. The integration saves you hours per week, and those hours are worth more than the billing gap.

You publish client sites, static or lightly dynamic, with forms and sometimes a login area: take Netlify. Built-in forms and identity spare you a third-party integration on every project, and framework neutrality is real comfort in an agency.

You want minimal cost first on static content: neither. Look at Cloudflare Pages, whose pricing doesn't charge for bandwidth, or plain hosting if your stack allows it. For a site with no build, a well-tuned WordPress or Webflow costs less in time than a deployment platform used poorly: I detailed this arbitrage in my Webflow vs Framer comparison and in the best CMS guide.

Last case: stateful apps. If you need a database, auth, and storage in one product, Supabase or Firebase cover the backend, and you place the frontend on one of the two platforms. The no-code tools overview gives alternatives if you don't want to code at all.

Key takeaways

• Vercel wins on Next.js, Netlify wins on neutrality and included building blocks.
• Vercel's free plan bans commercial use: a monetized site must move to Pro.
• Both charge per seat, which matters more than people think after three people.
• The dominant cost item is rarely the host, it's the database and network egress.
• Next.js lock-in is a slope, not a wall: the more primitives you use, the more porting costs.