Alex Tan

PC modding is one of the largest reasons people still own gaming PCs. Cloud gaming inherits the parent platform's mod support — but the boundary is messier than either side admits.

Marketing keeps promising the latency gap is closing. Speed-of-light math says there's a floor neither encoder nor data centre will get past — and we're already close to it.

Every cloud gaming conference includes a VR streaming demo. Almost none of them ship as products. The reason isn't the technology you'd guess.

NVIDIA's data centre footprint, its bring-your-own-library structural advantage, and the fact that it owns the encoder pipeline end-to-end. Plus the three things that could topple it.

Many users who could play a title on cloud explicitly wait for the local port instead. The pattern is rational, growing, and a meaningful headwind for the cloud gaming category.

Cloud services advertise 4K. Almost none of them ship a bitrate sufficient to make 4K materially better than 1440p. Here's the math, and why we'd take 1440p at high bitrate over 4K at low bitrate every time.

Cloud gaming on transit was supposed to be the killer mobile use case. In 2026 it works in specific countries on specific networks and not at all elsewhere. The detail matters.

Modern anti-cheat is kernel-level and detects virtualisation. Cloud gaming, by definition, runs games in a virtualised environment. The collision is uglier than the industry admits.

We asked 47 tournament-level players across five competitive shooters how cloud gaming fits into their practice. The answers are sharper and more nuanced than the standard 'cloud is unfit for competitive' line.

In-flight WiFi has been a punchline for two decades. Starlink's aviation rollout in 2024-2025 is changing what's possible at 35,000 feet — including cloud gaming.

Every major cloud gaming service runs Windows-on-bare-metal in the data center. The architecture is incidental rather than deliberate. The 2027-2028 question is whether the industry should keep doing it.

Fortnite is the most-played game on Earth. It's broken or unavailable on most cloud services. The reasons are a fight about platform fees, anti-cheat, and a publisher that doesn't want cloud.

Tencent's Start, NetEase's cloud platforms, Huawei Cloud Gaming. Three of the world's largest cloud gaming services run inside China and almost no Western analyst covers them. Here's what we know.

Cloud gaming services charge different prices in different markets. The variance is bigger than streaming video pricing and the reasons are more interesting than 'purchasing power parity'.

Chinese carriers bundle cloud gaming with mobile data. South Korean telcos bundle it with fibre plans. In the West, none of the major cloud services have done the same deal at scale. Why?

If 2027–2028 brings a real GPU oversupply like the post-crypto crash of 2023, cloud gaming's value proposition shifts. Three scenarios and what we'd watch for.

In 2022 cloud gaming services marketed RTX 3080-class hardware. By 2027 those same GPUs are mid-tier. How the cloud datacentre depreciation cycle actually works.

Standard cloud gaming advice: don't use a VPN. The standard advice is mostly right but the cases where it's wrong are interesting and the cloud services don't tell you about them.

Conventional wisdom: cloud is unfit for FPS because of latency. Reality: aim-assist on cloud-played console shooters is doing more work than it does locally — and that has interesting consequences.

Wi-Fi 7 has been shipping in flagship routers since 2024. The cloud gaming benefits are real but more constrained than the router marketing implies.

Every major cloud gaming service advertises HDR. Almost none of them deliver it correctly end-to-end. The reasons are technical, the impact is visible, and the fix isn't on any roadmap.

Every major carrier marketed 5G as the network technology that would enable cloud gaming everywhere. In 2026 the 5G cloud gaming reality is narrower than the marketing implied. What went wrong.

When some players in a competitive match are on cloud and others on local hardware, the matchmaking has to handle latency variance across the lobby. The current solutions are awkward and the effects on ranked play are real.

Xbox controllers, DualSense, 8BitDo, GameSir, Backbone, the Razer Kishi family. Cloud gaming services handle each of them differently. The fragmentation is worse than the marketing implies.

Cloud gaming services don't currently offer meaningful uptime guarantees. As the audience matures, SLA marketing is going to become a competitive dimension. Forecast for what's coming.

Chromebooks should be the perfect cloud gaming client. They aren't, and the gap between the potential and the reality is structural rather than incidental.

Streaming a cloud gaming session to Twitch is technically possible. Doing it well is harder than either Twitch or the cloud services admit. The friction explains why so few streamers play primarily on cloud.

AI frame generation doubles frame rates on local hardware. On cloud the integration is messier than the marketing suggests, and the user-visible benefit varies by tier.

AI upscaling — DLSS, FSR, XeSS — was built for local hardware. Running it on a cloud GPU and streaming the result raises questions the marketing doesn't address. The answer is more favourable than skeptics expected.

Capturing cloud gameplay for clips, montages, or content is a workflow with hidden complexity. Where the tools work, where they break, and what to know before you record.

Cloud gaming gives publishers a low-friction way to run open betas, public playtests, and demo access. Most publishers haven't figured this out yet.

Local gaming PCs handle multi-monitor setups natively. Cloud gaming services don't. The architectural reasons are interesting and the workarounds are worse than the problem.

League of Legends, Dota 2, Smite. MOBAs are the genre where cloud gaming should fit best — they're stable, latency-tolerant for casual play, audience is global. Instead they're underrepresented on cloud. Why.

Cloud gaming users on wireless headphones experience additional latency the cloud services don't disclose. The chain adds up in ways that matter for competitive play.

Gyro aiming has quietly become one of the best controller-based input methods for shooters. On cloud gaming it's almost-but-not-quite usable. Where the chain breaks.

Linux desktop gaming has been 'almost ready' for two decades. Cloud streaming did something Valve's Proton couldn't fully do — make every AAA accessible on Linux machines without the compatibility tax.

Most GTA V is great on cloud. GTA Online is broken on it in specific, frustrating ways. The reasons are technical, not incidental — and they tell you something about how cloud architecture handles certain genres.

Denuvo Anti-Tamper is hated by many players for affecting local performance. On cloud sessions the equation flips — Denuvo's performance overhead is invisible. The cloud as Denuvo's accidental rehabilitation.

Apple Silicon Macs are great computers and terrible local-gaming machines. Cloud gaming closed that gap, but the way it happened is not what the Mac community predicted.