PHP has been the bedrock of web development for decades. It was actually the first server-side language I learned how to program, 15 years ago. It has historically been one of the most dependable server-side programming languages, but the question is does it still hold that place in today's fast changing, ever growing market of programming languages?
PHP has a lot of things going for it, historical reliance, a large base of programmers who know and program it, server support in essentially every shared hosting plan available, as well as being the foundational language in two of the most popular Content Management Systems available (Wordpress and Drupal). Ease of use for newcomers is also key to its success. Once PHP is integrated with the web server it rarely if ever has to be managed or restarted and individual PHP scripts are run automatically by the PHP server service.
There is a lot of competition in today's programming market. Many of them have better tooling, are more exciting to work in and have great features that help with reliability such as typed variables, native multi-threading and speed of processing. But are these improvements worth the cost of the reliability and dependability of PHP? For me, that depends on the use case. If the application isn't going to have anyone experienced with managing it routinely, PHP is the best route to go because of the ability to easily deploy it on shared hosting environments by just uploading the scripts, while other languages such as Node.js, Swift, Go and others require an executable to be run and to ensure it remains up in the event of a crash, there needs to be a secondary server that restarts the application. If you need modern tooling with robust features such as typed variables and multi-threading it can help to use a more modern language rather than trying to use new libraries that make PHP more modern.
Ultimately I think it comes down most of the time to personal preference. There are likely few cases where any specific server-side language is going to drastically change things unless the application is frequently under a large load or has a lot of intensive data processing tasks.