Why developers leave Vercel: bandwidth pricing punishes high-traffic content sites, function invocation pricing surprises at scale, and vendor lock-in to Next.js features (ISR, Image) makes migration off costly. Teams growing beyond ~1M MAU on Next.js routinely evaluate alternatives.
Cloudflare Workers + Pages run Next.js via OpenNext at meaningfully lower cost — especially on high-bandwidth content. Bandwidth functionally free. Cold starts under 5ms vs Vercel's ~150ms. Trade-off: less polish on Next.js features Vercel ships first. For non-app workloads Workers wins clearly.
Netlify for non-Next.js apps. Pricing flatter and more predictable. Forms + Identity built in. For Astro, SvelteKit, or framework-agnostic deploys Netlify's integration is on par with Vercel's. Wrong alternative for Next.js-coupled features.
Render for full-stack Heroku-style apps. Different shape than Vercel — long-running services + bundled Postgres + Redis. Right alternative if Vercel's function-only model was constraining and you actually wanted persistent backend services.
For most apps yes. Reviewers running production OpenNext on Cloudflare Workers and AWS Lambda report it works for App Router and Pages Router. The Next.js features that depend specifically on Vercel's edge platform (some ISR patterns, advanced middleware) need adaptation. Test thoroughly before migrating critical traffic.
Yes — many teams keep Next.js core on Vercel and proxy bandwidth-heavy routes (videos, large downloads, public assets) through Cloudflare. The hybrid stack reduces costs without a full migration. Image optimization can stay on Vercel; raw bandwidth moves.
1-3 sprints depending on Next.js feature usage. Heaviest cost is rewriting features that depend on Vercel's ISR or Image components. Static export is usually mechanical. Edge middleware needs the most adaptation. Many teams report the savings paid back the migration in <6 months at growth scale.
