To be honest, I don’t think it fair to judge an engine based on what the users are doing. I’ve seen horrible games made in Unity and Unreal that even made it on to Steam. But of course it does not explain the lack of commercial games made in GD.
I guess the bottom line is, most commercial games are made in other engine but not only because they are better but also because they do offer priority support. Free tools normally don’t provide that and not even many paid tools. When you make a commercial game and you do have a budget and a deadline to meet and not only make it for fun, you do want to get customer support so in case you have any game breaking bug that stand in your way to meet your deadline, you can request a fix and don’t need to wait for a public release so you can continue and meet your deadlines. Also if you need further explanation that the documentation does not provide you can ask for it or in case you simply need technical help in general with the problem you face you can request assistance over skype or messenger or even face to face with a technician. It is something that most paid commercial engines do offer. In case of open-source engine devs also need to consider how many people are familiar with the source code so how many people can potentially fix any problems they might have. So, when you expect to see studio quality games, you also need to consider this when you ask the question why people are using or not using X and Y for making commercial games. It is not only about pure features and capabilities but also support they can get.
So, only because people are not using GDevelop for commercial games not necessarily means limitations but only that, there are better solutions out there for commercial and big games.
For making HTML5 games that runs in the browser, personally I would use Phaser, then Construct 2 then either GDevelop or RPG Maker MV, so GDevelop is only the 3rd or in case you want to make an RPG, the 4th on my personal list for HTML5 games in the browser.
For HTML5 on mobile, personally I put Phaser first, RPG Maker MV 2nd because they do support atlas images and that is important for performance and their content management system is extremely well designed and fast, Phaser also offer a light weight physics engine specifically for mobile devices that is a lot faster on mobile then Box2D, Then I put Construct 2 3rd and GD is only the 4th again.
For making mobile games in general, personally I prefer AppGameKit, for many reasons it is simply the best and most simple solution out there in my opinion then I would put Godot and Unity 2nd and 3rd. But there are lot of other engines that people are using for mobile, like Cocos2D and Starling for example that I don’t really know as AGK does what I need perfectly and I did not look further.
For making games in general there is no best or worst engine. For the most part it is depends on what kind of game you want to make and what platforms you do want to support. Normally you need to choose engine for the game you want to make. For example, No Man’s Sky started in Unity then they decided to make their own engine. EVE Valkyre started in Unity then the devs decided to go with Unreal. Most of the time you can’t just pick a single engine and go with that one for a life, you need to be flexible and find the engine fit best the game you are making.
At the moment I would recommend GDevelop to complete beginners who do want to target desktop and the browser but in case you do want to target mobile, I would recommend to keep your ambitions and the number of devices that you going to support at the minimum with GD.