Edge Runtime

• Cloudflare Workers: 300+ locations, V8 Isolates (not containers)
• Vercel Edge: Cloudflare backend, Next.js integration
• Deno Deploy: V8 + native Deno APIs
• Fastly Compute@Edge: WASM-based, any language
• Limits: 50ms CPU, 128 MB RAM, no native modules, limited fs

// Next.js Edge Function
export const runtime = 'edge';

export async function GET(request) {
  const { searchParams } = new URL(request.url);
  const city = searchParams.get('city');
  // ~10ms latency globally
  const weather = await fetch(`https://api.weather.gov/`);
  return Response.json(await weather.json());
}

• RSC
• Streaming SSR