KDE developers made a quiet but telling move this week. They removed support for desktop OpenGL from the KWin compositor in the upcoming Plasma 6.8 release. The decision caught some observers off guard. It wasn’t a premature victory lap for the long-awaited Vulkan renderer. Instead, the team opted for OpenGL ES across the board as an intermediate step.
This shift simplifies the code. It reduces headaches from mismatched implementations. And it keeps older hardware viable longer than a strict desktop OpenGL path might allow. The change applies strictly inside KWin’s compositing layer. Users can still run full desktop OpenGL applications outside the compositor without restriction.
“The various incompatibilities between desktop GL and OpenGL ES cause problems again and again,” wrote KWin developer Xaver Hugl in the merge request. “Code written and tested for one might not work for the other, ranging from things being rendered wrong to falling back to a software renderer or even functionality being completely broken. Since we need to support OpenGL ES for some old hardware, going forward, KWin will only support OpenGL ES. This means everyone is running the same OpenGL code, and we can drop some practically redundant code as well.”
The Phoronix report on the development highlighted that OpenGL ES delivers every feature KWin currently requires. Some legacy GPUs actually perform better under the ES profile than under a restricted desktop OpenGL context. No performance penalty is expected from the switch. The team anticipates smoother maintenance and fewer edge-case bugs.
But this graphics tweak forms only one piece of a larger picture. Plasma 6.8 also marks the point where KDE goes all-in on Wayland. The X11 session vanishes from the login screen. X11 applications continue to function through Xwayland. Yet the native Plasma X11 experience ends.
“After nearly three decades of KDE desktop environments running on X11, the future KDE Plasma 6.8 release will be Wayland-exclusive,” explained the Plasma team in their blog post. Support for the X11 session will continue through bug-fix releases of Plasma 6.7 into early 2027. The split of KWin into separate X11 and Wayland binaries back in Plasma 6.4 already paved the way. It let the Wayland branch advance without legacy drag.
Statistics back the timing. Ninety-five percent of Plasma 6.6 users already run the Wayland session, according to data cited in a Phoronix follow-up article from early June. That majority has grown steadily. Developers now feel confident removing the training wheels.
The dual moves — GLES-only compositing and X11 deprecation — reflect the same philosophy. Drop duplicated paths. Focus engineering effort where most users live. Clean up code that has accumulated decades of conditional branches and workarounds.
Plasma 6.8 brings other compositor refinements too. Animation work improved the bouncing application launch icon. Notifications now slide in with a fresh effect. The release also fixed the most common crash observed in early 6.8 testing, tied to multi-monitor variable refresh rate handling. And it added support for Emulated Input protocol version 1.6.
These details appeared in the weekly This Week in Plasma summary. They show steady progress on polish while the bigger architectural decisions take center stage.
Hardware support considerations played a role in the OpenGL decision. Older GPUs that struggle with modern desktop OpenGL profiles often handle OpenGL ES more gracefully. By standardizing on ES, KWin avoids maintaining two rendering paths that frequently diverge in behavior. The Vulkan renderer remains under active development. Once mature, it will replace the ES backend and bring further performance gains, particularly on modern discrete graphics.
NVIDIA users stand to benefit indirectly. Triple buffering, already enabled by default for their hardware in recent KWin versions, pairs well with the streamlined OpenGL ES path. Wayland improvements for NVIDIA have accumulated over the past year. The removal of the native X11 session reduces one more variable in driver compatibility testing.
Gaming implications drew early discussion on forums. Steam and Proton rely heavily on X11 protocols in some titles, yet Xwayland handles the translation. Early tests suggest the transition introduces little friction for most users. A few edge cases around full-screen behavior or input latency may need attention during the 6.7 stabilization period.
Accessibility features, long a concern during the Wayland migration, have reached parity in most areas and surpass X11 in others. Screen readers, magnifiers, and global shortcuts now operate reliably on the Wayland session. Automation tools and remote desktop solutions have adapted as well.
The broader KDE community reaction has been measured. Long-time X11 holdouts expressed concern about niche setups. Most acknowledged the data: the vast majority already chose Wayland. Distro vendors running Plasma as default will likely ship 6.8 with Wayland only, offering older LTS branches for those who need the X11 session longer.
Developers plan a thorough code cleanup once the X11 session disappears. Thousands of lines of conditional compilation and legacy handling can vanish. That work should accelerate feature delivery in subsequent releases. HDR support, improved color management, and explicit synchronization all benefit from reduced maintenance overhead.
Intel, AMD, and Qualcomm hardware already run well under the current Wayland implementation. The GLES standardization should extend stability to older integrated graphics still found in many enterprise laptops. Embedded and single-board systems that depend on OpenGL ES will see consistent behavior too.
Looking ahead, the Vulkan transition forms the next major milestone. KWin’s Vulkan renderer has shown promising results in limited testing. Full integration will allow the compositor to take advantage of modern explicit APIs, lower CPU overhead, and better multi-GPU handling. The current GLES decision buys time to get that renderer production-ready without forcing users onto experimental code.
Plasma 6.7.2, released alongside these 6.8 developments, addressed the most frequent KWin crash reported in the wild. It also improved Chromium-based video playback under Wayland. These point releases keep the current stable branch healthy while 6.8 shapes up for a feature-focused debut later this year or early next.
The pattern is clear. KDE has spent years preparing this transition. Incremental improvements to Wayland, careful data collection on user behavior, and targeted removal of legacy code now converge in Plasma 6.8. The desktop environment that once defined the Linux experience on X11 now bets fully on the modern protocol. The compositor that powered both now speaks a single, streamlined graphics language internally.
Whether the moves pay off in noticeably smoother animations, fewer driver quirks, or faster feature velocity will become apparent once users run the first betas. For now, the technical rationale stands on solid ground. Fewer code paths. Broader hardware compatibility. A clearer road to future graphics improvements.
And that, for an open-source desktop project balancing millions of users and decades of history, counts as progress.
WebProNews is an iEntry Publication