Why browser games got good again in 2026

For about a decade, "browser game" meant a compromise. That stopped being true, and it happened faster than most of us expected.

Three things converged

WebGPU is now broadly available. Compute shaders in a tab mean particle systems, real-time global illumination and physics that previously required a native build. The performance gap against a native target is no longer the deciding factor for a mid-size studio.

WebAssembly got shared memory and threads. Engines can finally use more than one core without heroic workarounds. For anything simulation-heavy — pathfinding, destruction, large entity counts — this is the difference between shipping and not shipping.

Mobile stopped being the exception. iOS Safari now supports the same graphics and audio surface as desktop. Since roughly two thirds of our traffic is mobile, this single change altered what we are willing to publish.

What still hurts

Cold-start weight is the remaining hard limit. A player who taps a link expects to be playing in under three seconds, and a 40 MB bundle cannot honour that on a mobile network no matter how good your engine is. The studios doing this well ship a tiny playable core first and stream assets behind it.

Persistence is the other rough edge. Without accounts you are relying on browser storage, which users clear and browsers evict. Anything a player would be upset to lose needs an explicit export path.

Where this leaves portals

The old pitch for a games portal was convenience: no install. The new pitch has to be curation, because the technical excuse for a thin catalogue is gone. If a game can be as good in a tab as it is in a store, the only thing worth paying attention to is whether someone chose it carefully.