Serverless Compute Cost Comparison

Scenario A (API-ish default): 15M requests — Workers 7 ms CPU → ~$8; Lambda 50 ms @ 512 MB → ~$2.80 (free-tier compute); Vercel Fluid iad1 10 ms active + 1 GB + 50 ms alive → ~$16.54. Scenario B (I/O-heavy): same 15M; Lambda 200 ms @ 1024 MB; Vercel 5 ms active / 200 ms alive — shows Workers' CPU-billing advantage when waiting on DB/API (Lambda wall-clock and Vercel provisioned memory keep charging). Narrative: Workers = request + CPU-ms (+ $5 floor); Lambda = request + GB-s wall time (+ free tier); Vercel Fluid = active CPU + provisioned memory + invocations. Static-asset heavy → Workers wins hard (assets free). Deep AWS event integrations → Lambda still fits. Next.js on Vercel → Fluid is the default path — compare before assuming “free on Hobby.” Soft-link: CDN pack and object-storage pack for the storage → CDN → functions funnel. Soft-link Workers AI calculator — inference tokens ≠ Workers request/CPU (different product).

Estimates only. Official list prices as of 2026-11-03 (Asia/Shanghai); provider pages prevail. Sources: Workers · Lambda · Vercel Fluid

Cloudflare Workers Cost Calculator · AWS Lambda Cost Calculator · Vercel Functions Cost Calculator · Serverless Postgres Cost Comparison · CDN Bandwidth Cost Comparison · Object Storage Cost Comparison · CI/CD Cost Comparison