Vercel Edge Functions reach general availability
Vercel has announced that Edge Functions are now generally available for all customers. The functions, which ran in public beta alongside the GA of Edge Middleware last summer, have recorded over 30 billion invocations on Vercel's Edge Network during that period.
Edge Functions are JavaScript, TypeScript, or WebAssembly functions built on a leaner runtime than traditional serverless compute. Deployed globally by default, they execute in the region closest to the incoming request to minimize latency. They are natively supported in Next.js, SvelteKit, Astro, and Nuxt, and can also be created as standalone functions via the Vercel CLI (version 28.9.0 or above supports the runtime: 'edge' configuration).
Unlike Routing Middleware, Edge Functions run after the cache, allowing them to both cache and return responses. This makes them suitable for data fetching and rewrites. For new projects, Vercel recommends Vercel Functions with the Node.js runtime and Fluid compute, while existing Edge Runtime code remains supported.
Performance and cost characteristics
The Edge Runtime is built on the V8 engine used by Chrome and does not run within a MicroVM, which generally makes Edge Functions faster and more cost-effective than traditional serverless functions. Vercel cites its own open-source OG Image Generation project as an example: moving to Edge Functions returned APIs almost 40 percent faster than a hot Serverless Function at a fraction of the cost. Generating a million images with OG Image Generation running in Edge Functions costs nearly 15 times less than the same workload in Vercel Functions.
Regional execution option
While global deployment is the default, Vercel now lets developers express a region preference for Edge Functions. By binding a function to a region close to a database or other data source, teams can avoid situations where global compute introduces more latency than it saves. This is particularly relevant for workloads that depend on external data retrieval rather than compute-only tasks.
One example cited is Keystone Education Group, which works with NoA Ignite to fetch data from Sanity, a headless CMS holding millions of files. The team shifted its API Routes incrementally from Serverless to Edge Functions, reducing costs and improving compute efficiency.
WebAssembly support
Because Edge Functions run on V8, Vercel added support for WebAssembly (Wasm). This lets developers compile functions written in languages like C or Rust for use directly in Edge Functions, often enabling existing libraries to be moved over by recompiling for Wasm. Computationally heavy workloads can run twice as fast in WebAssembly as in equivalent JavaScript.
By moving its OG Image Generation tool to Wasm and Edge Functions, Vercel reports that image generation became five times faster in P99 TTFB.
Larger functions and faster routing
Vercel has increased the size limit for Edge Functions to 2 MB for Pro customers and 4 MB for Enterprise customers, up from previous limits of 1 MB and 1 MB respectively. The company also states that routing infrastructure for Edge Functions has been improved, significantly reducing the time to start executing a function.
Pricing and billing model
Edge Functions are billed in execution units, each representing 50 ms of CPU time per invocation. CPU time counts computation only, not time spent waiting for data fetches or API calls. Hobby users receive 500,000 monthly execution units for free; Pro and Enterprise teams get 1 million monthly units included and can purchase additional usage. Usage metrics, including number of executions, execution units, and CPU time, are visible in the account dashboard.
Roadmap
Vercel's stated goal is to improve compatibility between Edge Functions and Vercel Functions, aiming for the Edge Runtime to be a proper subset of the Node.js API. This would let developers choose an execution environment based on performance and cost rather than API constraints.
The company is also working on better observability and error reporting, starting with Vercel Logs and Monitoring, and pursuing seamless interoperability across all its compute products, both locally and in production.
