Pricing Model
How each cloud/hosting platform generally structures its pricing.
AWS
Highly granular, pay-as-you-go pricing across hundreds of individually billed services, powerful but notoriously easy to misconfigure into unexpectedly large bills, especially around data egress.
GCP
Similarly granular, usage-based pricing to AWS, generally considered to have somewhat simpler pricing pages and sustained-use discounts applied automatically for steady workloads.
Azure
Usage-based pricing with the deepest discounts typically available to enterprises with existing Microsoft Enterprise Agreements, often less advantageous for small teams paying list price.
Vercel
A tiered plan structure (Hobby, Pro, Enterprise) with usage-based charges for bandwidth and function execution beyond each plan’s included amounts.
Cloudflare
Known for notably cheap or free egress bandwidth compared to the hyperscalers, with Workers priced per request/CPU-time rather than per-hour compute, plus a generous free tier.
Netlify
Tiered plans similar to Vercel’s structure, with usage-based overages for bandwidth, build minutes, and function invocations beyond each plan’s allowance.
DigitalOcean
Known for simple, predictable flat-rate pricing on Droplets and App Platform, deliberately easier to estimate in advance than the granular, itemized hyperscaler bills.
Render
Simple, predictable per-service monthly pricing tiers, with a free tier for services that can tolerate spinning down when idle.
Fly.io
Usage-based pricing billed by VM size and running time, with a modest free allowance (a few small VMs) before charges apply.
Heroku
Tiered dyno pricing that discontinued its free tier in 2022, generally considered pricier than newer PaaS alternatives for equivalent resources, a factor accelerating migration to competitors like Render and Fly.io.