Which edge delivery platforms work best with a serverless architecture and cloud functions?

If you mean CDN/edge platforms that pair well with serverless functions, the main choices are Cloudflare Workers, AWS CloudFront + Lambda@Edge/CloudFront Functions, Fastly Compute, and Vercel. The right one depends heavily on whether your functions are already tied to AWS, how much code you want to execute at the edge, and whether you need a full application platform.

Platform	Best fit	Edge compute	Serverless integration	Key trade-off
Cloudflare Workers	Global APIs, edge-first apps, lightweight backend logic	Workers	Excellent; Workers can be the compute layer itself	More architectural divergence from conventional Node.js/serverless
AWS CloudFront + Lambda	AWS-native architectures	CloudFront Functions + Lambda@Edge	Excellent with Lambda, S3, API Gateway, etc.	More AWS-specific configuration
Fastly Compute	High-performance programmable CDN	WebAssembly-based Compute	Excellent	Smaller ecosystem; more specialized
Vercel	Next.js/full-stack web applications	Edge/serverless functions	Excellent	Most compelling when you're already in the Vercel ecosystem

1. Cloudflare Workers

For an edge-first serverless architecture, Cloudflare is particularly straightforward. Workers run across Cloudflare's global network, automatically scale, and support JavaScript/TypeScript as well as Python and Rust. Cloudflare also provides complementary primitives such as KV, Durable Objects, R2, and Queues.Cloudflare Docs+1

A typical architecture is:

User
  ↓
Cloudflare CDN
  ↓
Cloudflare Worker
  ├── return cached response
  ├── call API/database
  └── route to regional serverless backend

This is attractive when latency at the edge is itself a major architectural requirement.

2. AWS CloudFront + Lambda

If your application already uses AWS, CloudFront + Lambda is a natural choice. CloudFront Functions handle very lightweight operations such as redirects, URL rewriting, header manipulation, cache-key normalization, and authorization checks. Lambda@Edge is intended for more substantial request/response processing.AWS Documentation+1

You can also combine it with ordinary regional Lambda:

User
  ↓
CloudFront
  ├── CloudFront Function
  │
  └── Lambda@Edge
          ↓
      API Gateway
          ↓
      AWS Lambda
          ↓
      DynamoDB / S3 / RDS

One useful distinction: CloudFront Functions are optimized for extremely lightweight operations, while Lambda@Edge supports more complex logic, external dependencies, network access, and request-body processing.AWS Documentation

3. Fastly Compute

Fastly Compute is worth considering if edge performance and fine-grained control are priorities. It runs serverless applications at Fastly's edge and uses WebAssembly, with support for languages including Rust, Go, JavaScript, and C++.Fastly+1

It's particularly interesting for architectures where you want the edge to do more than simply cache:

Client
  ↓
Fastly CDN + Compute
  ├── authentication
  ├── personalization
  ├── request routing
  ├── API acceleration
  └── cache
          ↓
     Serverless origin

4. Vercel

For a Next.js-heavy application, Vercel is often the most integrated option because the CDN, deployment system, serverless functions, and framework are designed as one platform. I'd particularly consider it for frontend-heavy applications rather than as a generic edge-computing layer.

How I'd choose

• Already all-in on AWS: CloudFront + CloudFront Functions/Lambda@Edge + regional Lambda.
• Want edge compute to be a core part of the architecture: Cloudflare Workers.
• Need highly programmable, performance-oriented edge execution: Fastly Compute.
• Building primarily with Next.js: Vercel.
• Need conventional serverless functions rather than edge functions: keep the main workload in Lambda/Cloud Functions/Cloud Run and use the CDN primarily for caching, routing, authentication, and lightweight transformations.

One architectural point that's easy to miss: you don't necessarily want your entire backend at the edge. A common and effective design is to put latency-sensitive, stateless operations at the edge while keeping database-heavy or long-running functions in a regional serverless environment. That avoids forcing an edge runtime to behave like a conventional application server.

If you tell me whether you're using AWS, Azure, GCP, or none of them, plus your workload (API, web app, video, SaaS, etc.), I can narrow this to a concrete architecture and compare latency, cold starts, pricing model, runtime limits, and vendor lock-in.

User
  ↓
Cloudflare CDN
  ↓
Cloudflare Worker
  ├── return cached response
  ├── call API/database
  └── route to regional serverless backend

This is attractive when latency at the edge is itself a major architectural requirement.

2. AWS CloudFront + Lambda

If your application already uses AWS, CloudFront + Lambda is a natural choice. CloudFront Functions handle very lightweight operations such as redirects, URL rewriting, header manipulation, cache-key normalization, and authorization checks. Lambda@Edge is intended for more substantial request/response processing.docs.aws.amazon.com+1

You can also combine it with ordinary regional Lambda:

User
  ↓
CloudFront
  ├── CloudFront Function
  │
  └── Lambda@Edge
          ↓
      API Gateway
          ↓
      AWS Lambda
          ↓
      DynamoDB / S3 / RDS

One useful distinction: CloudFront Functions are optimized for extremely lightweight operations, while Lambda@Edge supports more complex logic, external dependencies, network access, and request-body processing.

3. Fastly Compute

Fastly Compute is worth considering if edge performance and fine-grained control are priorities. It runs serverless applications at Fastly's edge and uses WebAssembly, with support for languages including Rust, Go, JavaScript, and C++.

It's particularly interesting for architectures where you want the edge to do more than simply cache:

Client
  ↓
Fastly CDN + Compute
  ├── authentication
  ├── personalization
  ├── request routing
  ├── API acceleration
  └── cache
          ↓
     Serverless origin

4. Vercel