Unexpected long waiting time when testing components with Volt

Nobody has claimed this yet.

### Volt Version

1.7

### Laravel Version

12.2

### PHP Version

8.4.5

### Database Driver & Version

SQLite / MySQL 8.4.3

### Description

Here's the context:
- I installed [Laravel's Livewire Starter Kit](https://github.com/laravel/livewire-starter-kit) (+ Volt) via Herd.
- A test is missing to achieve 100% coverage with Pest.

The tests are fast enough overall, I was going to make a PR to their repo but I realised the following:
- Each time I run `pest`, the first test running and using `Volt::test` always takes more than 20s.
- Following tests (using `Volt::test` or not) are fast as expected.

Using a clean installation (filtering the tests so we have small examples to work with):


Indeed, [the test at the very top](https://github.com/laravel/livewire-starter-kit/blob/d1edb7934feae4f0cfda6977ffc969d78edf72b1/tests/Feature/Auth/AuthenticationTest.php#L37) uses Volt.

---

Now, changing the order a bit:


In this case, the first test in the list using `Volt::test` is **"reset password link can be requested"** (can be found [here](https://github.com/laravel/livewire-starter-kit/blob/d1edb7934feae4f0cfda6977ffc969d78edf72b1/tests/Feature/Auth/PasswordResetTest.php#L23)).
This time, the execution time was 27.32s instead of 0.22s (in the first screenshot).

Is it expected behavior?
Does Volt need a first long initialisation so the other tests can run lightning fast?

Love Volt so far btw.

### Steps To Reproduce

1. Install Laravel's Livewire Starter Kit (+ Volt) using Herd.
2. Run `pest` in the terminal once it's ready.
3. Expect an odd waiting time, if not I am crazy or it's time to get a new PC.

Start by running the Pest suite from the Laravel Livewire Starter Kit reproduction, then compare the first Volt::test call in tests/Feature/Auth/AuthenticationTest.php with the one in tests/Feature/Auth/PasswordResetTest.php. Measure the initialization delay and inspect the reported Volt test behavior; done means establishing whether the delay is expected or isolating a reproducible regression.

Written by the indexing model from the issue text.