Six Years, No Fix: Arrowhead Left Helldivers 2's Biggest Flaw Intact

Here is the short version: Helldivers 2 runs on an engine that officially flatlined in 2018. Yep. Arrowhead built a giant live-service shooter on Autodesk Stingray, which had no support for most of the game’s development. People yelling 'just switch engines' make it sound easy. It wasn’t. And honestly, even a pivot to Unreal 5 might not have saved democracy either.

What actually happened under the hood

About a year ago, a player laid out the uncomfortable timeline and warned the game’s code might start fraying. Thirteen months later, that anxiety looks less like doomposting and more like a decent weather report. We have performance crashes, frame rate nosedives, specific enemy encounters that hard-freeze the game, sound glitches where squids whirr like busted blenders, and a PC install that can balloon past 150 GB. Arrowhead’s been in nearly nonstop damage control.

'the code unraveling at the seams'

But pinning every headache on Stingray misses the real story: this studio has deep history with that tech, and walking away mid-production wasn’t a clean, push-button decision.

How Arrowhead ended up married to a 'dead' engine

• 2009: After Swedish studio Grin implodes, Niklas Frykholm and Tobias Persson build a lean engine called Bitsquid, designed for smaller European teams who couldn’t swing Unreal’s licensing back then.
• Early adopters: Fatshark gets onboard. Arrowhead follows, shipping The Showdown Effect in 2013, Gauntlet in 2014, and the first Helldivers in 2015 on Bitsquid.
• June 2014: Autodesk buys Bitsquid and rebrands it as Stingray, hoping to compete with Unity and Unreal.
• December 2017: Autodesk waves the white flag and announces Stingray’s shutdown. Sales end January 7, 2018.
• By then, Helldivers 2 had already been in development for roughly two years (Arrowhead kicked off around 2016). The game would ultimately cook for about eight years total, with about six of those spent on tech that had zero official support.
• Meanwhile, Arrowhead’s team had a decade-plus of muscle memory with this codebase. They knew every quirk and how to wring performance out of it. Switching midstream to Unreal or Unity would mean torching that expertise and spending at least a couple of years just clawing back to feature parity.

Why 'just switch engines' was not a real plan

This is where the inside baseball gets interesting. After six years of custom tools, custom code, and custom everything layered onto Stingray, Arrowhead was basically running a Ship of Theseus situation. Is it even Stingray anymore? The studio has been maintaining an abandoned engine longer than Autodesk ever did. Like it or not, that is impressive.

Also, ripping out your tech foundation halfway through a live-service game is a productivity black hole. Your senior engineers go from 'we know exactly how this weird renderer behaves' to 'how do we do basic lighting in Unreal' while deadlines stack up and Sony’s production folks start asking pointed questions. You lose months, maybe a year, just ramping. And that is if everything goes well.

Would Unreal Engine 5 have fixed the mess?

Short answer: probably not. Different engine, different headaches. The Unreal 5 list in 2024–2025 has included shader compilation stutter, Lumen kicking frame rates in the shins when fog shows up, and PSO caching issues that turn loading into a second job. Plenty of teams wrestled those problems into shape. Plenty didn’t.

So sure, Unreal has a massive community and far more documentation than a defunct engine. That can make certain fires easier to put out. But it isn’t a magic sprinkler system. You are swapping Stingray’s constraints for Unreal’s, and pretending there was a perfect window where Arrowhead could casually pivot ignores how development actually happens. Sometimes you stick with the busted-but-familiar truck because the brand-new one will take two years to build, and you still might stall out on the highway.

So, was staying on Stingray the wrong call?

Frustrating performance and ballooning file sizes make it feel that way. But the alternative was burning years to switch engines, losing all the hard-won knowledge of a team that lived and breathed this code, and gambling that Unreal’s own problems would be more manageable. It is a pick-your-poison scenario, not a simple fork in the road.

If nothing else, credit where it’s due: Arrowhead has kept a discontinued engine alive longer than its owner did, while shipping a massive, messy, wildly popular live-service game. Whether that was the best choice or just the least bad one is up for debate. Personally, I think democracy probably would have died in Unreal, too.

What do you think: take the painful engine switch and slip the schedule by years, or ride the zombie tech you know and fight fires forever?