What Ubuntu 26.04 gets right
Canonical's release announcement is not in my search results, so I'll lean on Wright's account for the basics. He says Canonical shipped "Resolute Raccoon" on April 23 with Linux kernel 7.0 and GNOME 50. Ubuntu's own release notes back both points. They say the GA generic kernel moved from 6.8 to 7.0. They also say the Hardware Enablement stack moved from 6.17 to 7.0.
That kernel caveat matters. Ubuntu 24.04 shipped with 6.8, but a 24.04 machine on the HWE stack may be running a newer kernel. The "6.8 versus 7.0" comparison therefore describes the original GA kernels, not every machine.
Two details in the release notes are directly relevant to gamers:
- NVIDIA drivers: Canonical documents
sudo ubuntu-drivers install --recommendedas the fastest route. It documentssudo ubuntu-drivers install nvidia:595as the supported route to the latest driver. It also says the only drivers Ubuntu supports officially are the ones installed those ways. That supports Wright's point that you no longer need to hunt for PPAs. It doesn't promise identical results on every card. - NTSYNC: The release notes mention an NTSYNC driver that emulates Windows NT sync primitives. Canonical says it offers better performance potential for Windows games on Wine and Proton. That's a gaming-relevant change Wright doesn't mention.
One date caveat applies. Ubuntu's release list shows 26.04.1 LTS arrived August 27, 2026. Ubuntu 26.10 "Stonking Stingray" arrived October 15, 2026. Wright's piece is about 26.04, so read it as a verdict on that LTS, not on the newest Ubuntu.
Where Windows still wins: anti-cheat
Wright calls anti-cheat the biggest roadblock, and the evidence supports the general point. Valve's Steamworks documentation says Proton supports some anti-cheat middleware, including Easy Anti-Cheat and BattlEye. Each needs publisher-side enabling. Valve recommends user-space anti-cheat and says kernel-space solutions are not currently supported.
That makes this mostly a title-by-title decision by publishers, not a blanket Linux ban. Wright's flat statements that certain games "probably never will" run on Linux are a prediction, not a documented fact. I couldn't verify the current Linux status of each game he names.
Battlefield 6 is the useful example, with a caveat. EA's August 28, 2026 anti-cheat update shows how much Javelin demands of PCs. It says EA is enforcing TPM 2.0 requirements it had relaxed. It says HVCI and VBS are required by the title but not yet enforced. Those are Windows security features. That fits Wright's conclusion that Battlefield 6 is a poor fit for Linux. But EA's post doesn't say Javelin is permanently incompatible with Linux, so cite that status to a game-specific source.
The NVIDIA DirectX 12 gap
Wright says DirectX 12 games run through vkd3d-proton, and that one test summary put the average NVIDIA loss near 18% versus Windows. He gives no methodology, and I couldn't trace the figure to a benchmark. Treat it as an unverified claim, not a measured fact.
The vkd3d-proton release notes do show active work on NVIDIA-related issues. They include fixes submitted by NVIDIA and a performance workaround for NVIDIA's kernel when waiting on fences. They also note that VK_EXT_descriptor_heap support wasn't enabled by default in the 3.1 release because of driver quality concerns. So the picture is "improving, game-specific, not settled." It isn't a single universal penalty.
The Steam Snap problem is fixable
Wright says Ubuntu's default Steam is Canonical's Snap and that Valve recommends the .deb instead. Valve's Steam-for-Linux repository documents installing steam_latest.deb. You can double-click it or run gdebi steam_latest.deb after installing gdebi-core. Valve's README also lists the latest Ubuntu or Ubuntu LTS as the supported stable-client platform.
The claim about rising bug reports tied to the Snap isn't confirmed in my sources, so I won't repeat it as fact. If you switch packages, plan how you'll handle an existing Steam library. Installing the .deb is a packaging choice, not a limit of Linux itself.
What the Steam survey really says
Wright's market-share argument is that Linux is too small for publishers to prioritize. The October 2026 numbers check out. Windows stood at 93.70% and Linux at 3.95%.
Context from the survey coverage:
- The +0.90 headline gain compared October with September's originally published 3.05%. Valve later revised September to 3.90%, so the real gain is about 0.05 points.
- Linux reached an all-time high of about 5.33% in March 2026 and stood at 4.52% in April. Wright says it "cracked 5%" in March, and that holds up. His claim that the spike "didn't stick" is fair given later months, though April was still above 4.5%.
- Simplified Chinese respondents fell from 31.39% of the sample to 21.98% in October. Because those respondents are overwhelmingly on Windows, Linux's share rose. Month-to-month swings partly reflect who answered.
- The survey is voluntary and anonymous, so it is a snapshot of participating Steam users, not the whole PC gaming market.
One more detail undercuts the idea of a Linux gaming audience that's all in on Ubuntu. CachyOS led Linux Steam users at 16.39%, SteamOS Holo followed at 13.89%, and Ubuntu 26.04.1 LTS sat at 3.63%. Ubuntu is a real option, but it isn't what most Linux gamers on Steam run.
The claim that a small share explains publishers' anti-cheat decisions is a plausible inference, not something the survey shows.
Windows-only extras and the dual-boot answer
Wright says Xbox PC Game Pass isn't available on Linux, and that Razer Synapse and Logitech G Hub are Windows-only. I didn't verify current platform support for those products, so check each vendor's support page. Browser access, community tools and cloud streaming are not the same as a local Windows app. Wright notes GeForce Now can stream Battlefield 6, but he dislikes the subscription, latency and image compression.
His practical conclusion is dual-booting, with Windows on a separate SSD for the games that need it. That works, but you then maintain two operating systems. For a player who wants to leave Windows entirely, it defeats the purpose.
Who should switch
Based on the evidence above, here is a rough way to decide:
- Good candidates: mostly single-player Steam libraries, an AMD GPU, and no reliance on kernel-level anti-cheat.
- Riskier: NVIDIA owners who play DirectX 12 titles and care about every frame. Competitive multiplayer fans. Anyone who depends on Windows-only peripheral software or Game Pass.
- Before you commit: list the games you actually play and check each one's current Linux status. Then decide whether a dual-boot setup is acceptable.
Wright's headline question is fair, and the answer is more nuanced than "Linux gaming isn't ready." Single-player and Steam-centric libraries are in good shape on Ubuntu 26.04. The remaining obstacles are mostly publisher decisions, driver-specific performance gaps, and Windows-only software. If your game list avoids those, Ubuntu is a credible primary OS. If it doesn't, keeping Windows around is the pragmatic call.