Memory leak on async calls in dev server

Link to the code that reproduces this issue

https://github.com/pontasan/nextjs_case20251101_2.git

To Reproduce

1.Start the server with npm run dev.

2.Open the following link in your browser: http://localhost:3000/test

3.After a while, the server crashes while GC logs are being printed.

Current vs. Expected behavior

It appears that a memory leak occurs only in the development server when calling async functions from route.ts.
Running a large number of async functions in a loop results in a steady increase in heap usage over time.
In contrast, there is no such issue in the production build.

I noticed that when running the development server for an extended period, the heap usage never decreases, which led me to conduct this experiment.

This example is intentionally extreme to make the issue easier to reproduce, but even in development mode, I believe it should be able to handle asynchronous function calls inside a loop just as it does in production.

import v8 from 'v8'

export async function GET(request: Request) {
    v8.setFlagsFromString('--trace_gc')

    for (; ;) { // This is an intentionally extreme example to make the issue easier to reproduce.
        await testFunc()
    }

    return Response.json({})
}

async function testFunc() {
    // NOP
}

The following image was taken right after the development server crashed.

The issue does not reproduce in the production build or in the following Node.js code, and the heap remains stable.

const v8 = require('node:v8')
v8.setFlagsFromString('--trace_gc')

async function testFunc() {
    // NOP
}

(async () => {
    for (; ;) {
        await testFunc()
    }
})()

On the other hand, in the production build environment, the heap remains stable.

Running the same test in Node.js also shows stable behavior.

Provide environment information

Operating System:
  Platform: darwin
  Arch: arm64
  Version: Darwin Kernel Version 21.6.0: Mon Aug 22 20:19:52 PDT 2022; root:xnu-8020.140.49~2/RELEASE_ARM64_T6000
  Available memory (MB): 65536
  Available CPU cores: 10
Binaries:
  Node: 24.11.0
  npm: 11.6.1
  Yarn: N/A
  pnpm: N/A
Relevant Packages:
  next: 16.0.2-canary.3 // Latest available version is detected (16.0.2-canary.3).
  eslint-config-next: N/A
  react: 19.2.0
  react-dom: 19.2.0
  typescript: 5.9.3
Next.js Config:
  output: N/A

The same issue occurs not only in 16.0.2-canary.3, but also in 16.0.1.
It also occurs in Node.js v22.21.1.

Which area(s) are affected? (Select all that apply)

Route Handlers

Which stage(s) are affected? (Select all that apply)

next dev (local)

Additional context

Below is a video of the development server.
The GC logs show a steady increase in memory consumption.
https://drive.google.com/file/d/1_Z_2UWRbj1IWBtgIwCb7nQcfnZAdGwaw/view?usp=drive_link

Below is a video of the production build.
The heap usage remains stable.
https://drive.google.com/file/d/1byrJM7SMAw_vodNmji6zxkrpM6BrgG2d/view?usp=drive_link

Below is a video of the same process running in Node.js, where the heap also remains stable.
https://drive.google.com/file/d/1PKZdfE_0eCR2QMG8aANzc7RxglOHfPQM/view?usp=drive_link

The test code can be found at the following link:
https://github.com/pontasan/nextjs_case20251101_2.git

The Next.js example is located at: /app/test/route.ts
The Node.js example is located at: /test.js

Start with the reproduction repository and compare app/test/route.ts with test.js, running npm run dev and the production build to confirm the development-only behavior. Investigate the dev server's handling of the Route Handler's repeated async calls, using the supplied GC output as evidence; done means memory remains stable in development as it does in production and Node.js.