Cloudflare Workers vs AWS Lambda@Edge vs CloudFront Functions (2026)

Edge computing puts your code closer to users — reducing latency for auth checks, redirects, personalization, and request manipulation. Three major options exist in 2026: Cloudflare Workers, AWS Lambda@Edge, and CloudFront Functions. They look similar but have very different tradeoffs.

Quick Comparison

Cloudflare Workers

Workers run at Cloudflare's edge network in 300+ cities. They use V8 isolates — not containers — which means true zero cold starts. A new isolate starts in microseconds.

Best For

• Auth/JWT validation at the edge
• Request/response transformation
• Geo-based routing
• A/B testing
• API proxying with caching
• Full edge apps (Workers + KV + D1 database + R2 storage)

Code Example

export default {
  async fetch(request, env, ctx) {
    const url = new URL(request.url)
    
    // Geo-based routing
    const country = request.cf?.country
    if (country === 'IN') {
      url.hostname = 'india.api.example.com'
      return fetch(url.toString(), request)
    }
    
    // JWT validation
    const token = request.headers.get('Authorization')?.replace('Bearer ', '')
    if (!token || !isValidJWT(token, env.JWT_SECRET)) {
      return new Response('Unauthorized', { status: 401 })
    }
    
    return fetch(request)
  }
}

Pricing

• Free tier: 100,000 requests/day, 10ms CPU time per request
• Paid: $5/month includes 10M requests, then $0.30/million

Strengths

• True zero cold starts
• Workers KV for edge caching (globally replicated key-value store)
• Durable Objects for stateful edge compute (real-time, coordination)
• Runs independent of your origin cloud provider
• R2 storage + D1 SQL database — full edge stack

Weaknesses

• Limited Node.js API compatibility (improving with compat mode)
• Workers KV is eventually consistent (not strongly consistent)
• Not ideal for CPU-heavy workloads (50ms CPU limit)

AWS Lambda@Edge

Lambda@Edge runs your Lambda functions at CloudFront edge locations. Unlike CloudFront Functions, it supports full Lambda runtime capabilities — Node.js or Python, larger packages, longer execution.

Best For

• Complex request manipulation requiring external calls
• Authentication that needs to call external identity providers
• Dynamic content assembly at the edge
• Image resizing/transformation (using Sharp or similar)

CloudFront Trigger Points

Viewer Request    → runs on every request before CloudFront cache
Origin Request    → runs only on cache miss, before going to origin
Origin Response   → runs after origin responds, before caching
Viewer Response   → runs on every response before sending to user

Code Example

// Viewer request — add auth header
exports.handler = async (event) => {
  const request = event.Records[0].cf.request
  const headers = request.headers
 
  // Add internal auth header
  headers['x-internal-auth'] = [{
    key: 'X-Internal-Auth',
    value: process.env.INTERNAL_TOKEN
  }]
 
  // Redirect www to non-www
  const host = headers.host[0].value
  if (host.startsWith('www.')) {
    return {
      status: '301',
      headers: {
        location: [{
          key: 'Location',
          value: `https://${host.slice(4)}${request.uri}`
        }]
      }
    }
  }
 
  return request
}

Pricing

• $0.60 per million requests + Lambda duration charges
• More expensive than CloudFront Functions for simple use cases

Strengths

• Full Lambda capabilities — install npm packages, use Python
• Longer execution time (5s for viewer events, 30s for origin)
• Larger package size (50MB zipped)
• Can make external network calls

Weaknesses

• Cold starts (100ms–1s for Node.js, worse for Python)
• Must be deployed to us-east-1, replicates globally (slow deployment)
• Tied to CloudFront — doesn't work without it
• More expensive than CloudFront Functions

CloudFront Functions

CloudFront Functions is AWS's answer to the "Lambda@Edge is too slow and expensive for simple redirects" problem. It runs in CloudFront's edge locations with sub-millisecond latency.

Best For

• URL rewrites and redirects
• Simple header manipulation
• Cache key normalization
• Basic request validation

Code Example

// URL normalization — remove trailing slash
function handler(event) {
  var request = event.request
  var uri = request.uri
  
  // Remove trailing slash
  if (uri.endsWith('/') && uri.length > 1) {
    return {
      statusCode: 301,
      headers: {
        location: { value: uri.slice(0, -1) }
      }
    }
  }
  
  // Add index.html for SPA routing
  if (!uri.includes('.')) {
    request.uri = '/index.html'
  }
  
  return request
}

Pricing

• $0.10 per million invocations (cheapest of the three)
• Free tier: 2M invocations/month

Strengths

• Cheapest option
• Lowest latency (~1ms)
• Simple deployment alongside CloudFront

Weaknesses

• Very limited: no network calls, no external packages, no async
• JavaScript ES5.1 only (no modern JS features)
• 1ms CPU time limit (seriously limited)
• 2MB package size limit

Performance Comparison

Cold starts are the killer for Lambda@Edge on low-traffic routes. Cloudflare Workers' isolate model eliminates this entirely.

Which One to Choose

Choose Cloudflare Workers if:

• You want the best performance (zero cold starts)
• You're not locked into AWS
• You want a full edge stack (KV, D1, R2, Durable Objects)
• You need to run on non-AWS infrastructure
• Cost predictability matters (flat pricing)

Choose Lambda@Edge if:

• You're fully on AWS and want deep integration
• You need Node.js packages (Sharp for images, etc.)
• You need longer execution time
• Your use case requires more than 1ms CPU but less than 5s

Choose CloudFront Functions if:

• You only need simple URL rewrites/redirects
• You're already on CloudFront and want the cheapest option
• Latency is critical and logic is minimal

Multi-cloud teams: Cloudflare Workers is the most portable — not tied to any cloud provider.

Related: AWS Lambda vs Google Cloud Run vs Azure Functions | AWS VPC Networking Guide | Cloudflare vs AWS CloudFront CDN

Affiliate note: Cloudflare Workers free tier covers 100K requests/day — enough for testing and small production workloads. AWS CloudFront free tier includes 1TB data transfer and 10M CloudFront Functions invocations/month.