Small Next.js Apps Are Getting Priced Out of Vercel Plus Its Add-Ons

Vercel is a wonderful place to start a Next.js app and an increasingly expensive place to keep a small one. The free tier is genuinely generous — until the app grows the things real apps need: a database, scheduled jobs, more bandwidth, maybe a bit more compute. Each is its own line item or a paid plan, and the bill for a hobby project quietly crosses into "wait, I'm paying $40 a month for this?" That's the moment to ask whether a single small VPS does the same job for the price of a coffee.

Where the Vercel bill actually comes from

The platform itself is rarely the whole cost; the add-ons are:

• Database — a managed Postgres add-on with its own monthly price and usage metering.
• Cron / background jobs — scheduled functions, often gated behind a paid tier.
• Bandwidth & function invocations — fine at hobby scale, metered as you grow, with overages.
• Team/pro features you bumped into for one capability.

Stacked, a "free" project becomes a multi-service subscription. (This is the same dynamic Oracle free-tier refugees hit — convenience priced per add-on.)

What a VPS gives you for one flat price

A single small VPS runs the whole shape on one bill:

• The Next.js app itself (built standalone and run in Docker, or with a Node process manager).
• A Postgres right next to it — no per-add-on metering (just mind database sprawl if you add more apps).
• Cron via the OS, not a paid scheduler.
• Bandwidth within the VPS's generous included transfer, not metered per request.

One predictable monthly number, often less than the database add-on alone cost.

What you take on (be honest)

Self-hosting isn't free of cost — it moves cost to your time:

• Builds and deploys are now yours: a CI step or a pinned-image promotion flow instead of git-push-to-deploy.
• TLS, reverse proxy, and the edge are your job — the things Vercel handled for you in a PaaS-to-VPS migration.
• Uptime and patching are on you. For a small app that's minutes a month; it's not zero.

When to move and when to stay

• Move when the add-ons (database + cron + bandwidth) dominate the bill, the app is small enough to fit a $5–10 VPS, and you're comfortable owning deploys.
• Stay when git-push-deploy and zero-ops are worth the premium, traffic is genuinely spiky (serverless scale-to-zero earns its keep), or it's a throwaway you don't want to operate.

A middle path: keep the static/edge bits where they're cheap and move the stateful parts (database, cron) to the VPS — the hybrid logic of putting each workload where it's cheapest.

Migration sanity checks

• Build Next.js in standalone mode for a lean container.
• Move the database with a real dump, not a file copy.
• Reproduce cron jobs as OS cron / a small scheduler.
• Wire TLS + reverse proxy (and renewal that reloads).
• Benchmark the VPS so it actually handles your traffic.

Takeaway

Vercel earns its price for some apps and outgrows its value for small ones the moment you bolt on a database, cron, and bandwidth. If your app fits a cheap VPS and you'll own the deploys, one flat-rate box replaces a stack of metered add-ons — and the only real cost is the operations you take back. Decide by where your bill comes from, not by platform loyalty.