Cloud Gaming Is Technically Ready — So Why Hasn't It Gone Mainstream?

Cloud gaming's technical feasibility was proven long ago, but it's still a long way from becoming the future of the gaming industry. The Verge podcast host David Pierce and senior editor Sean Hollister dive deep into why cloud gaming has yet to deliver on its promise — from network latency to copyright battles to the business model dilemma.

The Technical Foundation of Cloud Gaming: It Works, But Not Reliably Enough

Sean Hollister has been covering cloud gaming since 2010, having witnessed the rise and fall of Gaikai and OnLive. His assessment is blunt: the technology has "sort of worked" under good network conditions since the 2010s, but there's a chasm called reliability between "works" and "works well."

Cloud gaming works similarly to Netflix streaming — compressed frames are transmitted quickly over the network — but the key difference is two-way interaction. When a user presses a button, the command must travel to a cloud server, and the rendered result must be sent back to the screen. This end-to-end latency must be low enough to ensure playability. Sean himself beat Genichiro in Sekiro using GeForce Now, proving the experience can be excellent — but only if the network doesn't falter.

The problem is that any fluctuation in the home network — a roommate suddenly downloading large files, an unstable wireless signal, suboptimal peering between the ISP and the cloud provider — can instantly shatter the experience. Netflix can buffer ahead, but games can't, because every frame depends on the user's real-time input.

End-to-end latency and frame synchronization: Cloud gaming is far more latency-sensitive than video streaming. From the moment a user presses a button to the on-screen response, the entire chain includes input transmission, cloud rendering, encoding and compression, network transmission, decoding and display, and more. The industry generally holds that for most action games, end-to-end latency must be kept under 100 milliseconds to ensure basic playability, while competitive games demand even lower. Additionally, cloud gaming must handle frame synchronization — every frame rendered in the cloud must correspond strictly to user input, and any network jitter can cause dropped frames or out-of-order delivery, resulting in perceptible stuttering. This is fundamentally different from Netflix's buffering mechanism: video can preload several seconds of content, but games must respond in real time.

The Overlooked Hidden Barrier: Copyright and Licensing

Sean admits he doesn't play cloud games often, and the real reason isn't latency — it's friction and ownership anxiety. He'd rather play on his own hardware without wondering, "Is this game still available on the cloud service? Will the license expire in a few months?"

This concern isn't unfounded. During GeForce Now's beta, Nvidia allowed users to install any game from their Steam library, but game publishers subsequently demanded that "streaming rights" be licensed separately. Even if a user simply wants to stream a game they purchased to a rented remote PC, the publisher's approval is required. Valve later included streaming authorization by default in the Steam distribution agreement, but that very fact proves the legal obstacles are real.

Streaming Rights: In the game industry, publishers typically license a game's "public performance rights" or "streaming rights" separately from the ordinary personal-use license. A cloud gaming service that wants to stream a game to users must obtain explicit authorization from the publisher, even if the user has already purchased the game. This is because cloud gaming involves transmitting game content beyond the user's device, which may be considered a form of "redistribution." Nvidia's GeForce Now beta allowed users to install any Steam game, but publishers subsequently demanded separate licensing, leading to a large number of games being pulled. Valve's default inclusion of streaming authorization in the Steam distribution agreement was precisely intended to simplify this process, but it also highlights how widespread the legal barriers are.

The Business Model Deadlock: GPUs Don't Get Cheaper by Magic

The economics of cloud gaming have never added up. Running a game requires GPU compute, and in the cloud someone still has to put that GPU in a server. GPUs are expensive in themselves, and AI data centers are currently buying up these chips en masse, leaving compute resources for cloud gaming both costly and scarce.

Sean points out that Nvidia is one of the few companies that can sustain long-term investment because it is itself a GPU manufacturer; Google's Stadia has shut down; Amazon Luna has repositioned itself multiple times, starting with lower quality and "faking it until making it." Sony acquired the patents of Gaikai and OnLive and integrated them into PlayStation Now, but did not spark a revolution.

One possible turning point: when AI data centers retire older GPUs, that hardware might be repurposed for cloud gaming. After all, the vast majority of games — the long tail that the Steam Deck can run — don't need the latest and greatest chips. If cloud services could offer cheap streaming of older games, they might win over some hardware-price-sensitive players.

GPU compute costs and the cloud gaming economic model: Cloud gaming service providers must allocate GPU resources for each concurrent user, and the procurement, maintenance, and electricity costs of GPUs are high. Unlike video streaming, cloud gaming's compute demand grows linearly with the number of users, making it difficult to significantly amortize costs through economies of scale. The current scramble for GPUs by AI data centers has further driven up market prices, putting cloud gaming providers at a disadvantage in compute procurement. Additionally, cloud gaming must consider server utilization — user playtime is unevenly distributed, leading to resource strain during peak hours and idle capacity during off-peak hours, further reducing economic efficiency. As a result, cloud gaming subscription prices often fail to cover costs, as exemplified by the shutdown of Google Stadia.

Microsoft's Predicament: It Has All the Puzzle Pieces but Can't Put Them Together

In theory, Microsoft has everything needed for cloud gaming success: Azure cloud infrastructure, the Xbox brand, and the Activision Blizzard content library. But the reality is that Microsoft has not been able to use these to disrupt the gaming industry.

Sean analyzes two layers of reasons. First, Microsoft's cloud infrastructure wasn't optimized for gaming before the AI boom — the buildout of game servers came later. Second, the legal commitments from the Activision Blizzard acquisition require Call of Duty to be available on all platforms, but the vision of "the phone is an Xbox" is blocked by Apple's and Google's app store rules — the 30% cut and closed ecosystems leave Microsoft unable to circumvent them.

The deeper issue is that every party in the value chain has its own interests to protect. Publishers want to control licensing, platforms want their cut, and hardware makers want to keep selling consoles. If even one link doesn't cooperate, the ideal picture of "playing all games anytime, anywhere" cannot be realized.

App store commissions and closed ecosystems: Apple's App Store and Google Play take a 30% commission on digital content transactions and prohibit in-app links directing users to external payment channels. This directly obstructs Microsoft's vision of "the phone is an Xbox" — if Xbox games are streamed through a mobile app, Microsoft would have to pay hefty commissions to Apple or Google and could not directly handle subscriptions and purchases. Microsoft has tried to bypass store restrictions through web versions, but browser limitations on iOS still affect the experience. Additionally, Apple and Google have separate rules for cloud gaming apps, requiring each game to be submitted as a standalone app for review, further increasing operational complexity. These platform rules constitute a key obstacle to cloud gaming's mobile deployment.

Can GTA 6 Be the Turning Point?

The interview ends with a discussion of GTA 6. Sean believes it could create a brief window: when a blockbuster game launches, players who can't get a console will seriously consider cloud gaming. With game hardware prices high, the PS5 Pro scalped far above MSRP, and budget handhelds hard to find, cloud services might gain a wave of curious early adopters.

But whether this pull is lasting, Sean is skeptical. Once the hype fades, whether players are willing to keep subscribing remains unknown. He puts it bluntly: "If GTA 6 can't push the world toward cloud gaming at this moment, I don't know what can."

Cloud gaming technology is already good enough — good enough to overlook the occasional stutter. But it has never crossed that threshold — not bandwidth, not latency, but systemic friction woven from reliability, copyright, business models, and industry power struggles. Perhaps when AI data centers retire their old GPUs, or when the next phenomenon-level game arrives, the situation will change. But until then, cloud gaming remains that technology that "works, but hasn't won."

One-line Summary

Cloud gaming technology is already playable, but reliability, copyright, cost, and industry power struggles have kept it from delivering on its promise.

Paragraph Summary

Cloud gaming's technical feasibility was proven long ago, but under the multiple obstacles of reliability, licensing, business models, and industry power struggles, it has never become the future of the gaming industry. The Verge podcast host David Pierce and senior editor Sean Hollister dive deep into this dilemma. Technically, cloud gaming demands extremely low end-to-end latency and high network stability, and any fluctuation in the home network can ruin the experience. On copyright, publishers require streaming rights to be licensed separately, meaning even streaming a game you purchased requires permission. On business models, GPU compute costs are high, and AI data centers' scramble further drives up prices, leading to Stadia's shutdown and Luna's multiple repositionings. Despite owning Azure, Xbox, and Activision Blizzard, Microsoft is constrained by app store rules and legal commitments, failing to realize the vision of "the phone is an Xbox." GTA 6 may open a brief window, but whether it translates into long-term subscriptions remains doubtful. Cloud gaming is still a technology that "works, but hasn't won."