There are many studios with their own engines that rival or exceed UE5 - which seems overhyped, because at this point they caught up with graphics fidelity without terrible performance that dread a lot of UE5 titles.
Recent notable example is Crimson Desert, they spent years building their own engine for this game and IMO they raised the bar when it comes to creating a huge realistic world.
Others that come to my mind are Decima and RE Engine.
As a Korean freelancer, I’ve spoken with former developers from Pearl Abyss. They officially work 10 to 7, but the relentless crunch culture drives most people out.
While the company is extremely proud of its proprietary engine, I was told it causes severe internal politics. The studio is heavily biased toward the engineers who built the engine. Another huge downside is the lack of documentation—you can't just Google your bugs. (Granted, this was the situation two years ago).
The CEO is famously known in Korea for prioritizing developers while devaluing writers and planners. However, even within that developer-first environment, the proprietary engine has birthed a clear internal hierarchy among the programmers
The main point of using a 3rd-party engine like Unity or UE is not to buy technical excellence, but to get a 'good enough' asset pipeline, authoring tools and engine runtime cheaper than building and nurturing your own inhouse engine and tools team ... especially when the best programmers on those teams are then poached by Epic or Unity anyway ;)
I agree and that's why I'm happy not all studios not going this route. It usually means they game will be unique above average because one of the reasons would be needing the engine to work for something else than a "typical game template".
Crimson Desert's engine is heavily derived from Black Desert's engine (called BlackSpace Engine) so it wasn't really from the ground up, but your point still stands.
I've read somewhere they had several attempts until they could comfortably support the scale of the game, so maybe it was just rebuilding some major parts
In the last few years the pendulum has been swinging back from inhouse engine to Unreal Engine. There are a couple of holdouts, but my guess is that the majority of AAA games currently released are back on UE - at least it feels that way ;)
And Unity always ruled supreme for AA and mobile games.
A couple of holdouts? Hardly. Capcom uses the RE Engine. Released ~4 games with it in the past three years. In fact, SF6 went from UE to RE Engine. Death Stranding 2 is Decima. Id still uses IdTech, and that's never changing. Ubisoft still uses Anvil, and Outlaws was Snowdrop. EA still uses Frostbite. Bethesda continues to use their crappy engine, which I forget the name of. FromSoftware will likely never give up their engine, which has been used for Elden Ring and its successors. There's more that I can't be bothered to look up.
When the Xbox360 was released, a lot of game companies moved from inhouse to UE3. In the next generation the pendulum has swung back to inhouse, and now its swinging back to UE5. Seems to be a 10 year cycle or so...
I heard a great perspective on this from a Chinese developer. The reliance on outsourced engines like Unity or Unreal isn't just driven by the high costs of in-house development and yearly upgrades. The real issue is career mobility.
Knowledge of a proprietary engine is completely locked to that specific company. They pointed out that Unity and Unreal became industry standards simply because of the dynamics of changing jobs. I fully agree with that assessment.
It barely resembles id tech 3 at this point (I am an engine and pipeline programmer at Sledgehammer Games). There are very very few remnants of it in the codebase. I occasionally do some spelunking to see what’s left and these days it’s just the occasional comment and function names (although the content of the functions will have changed significantly).
I'm guessing its cheaper to iterate the old engine like this than to try and rebuild a new one from scratch one day, then move everyone onto that new engine and have them be as familiar as they were with the old one made of a couple decades of mostly duct tape it sounds like now.
Exactly! Armchair enthusiasts and gamers and such will talk about some game release and its "new engine" but throwing away the entire codebase is rarely a sensible choice financially speaking.
Recent notable example is Crimson Desert, they spent years building their own engine for this game and IMO they raised the bar when it comes to creating a huge realistic world.
Others that come to my mind are Decima and RE Engine.