Steeam

Steam Machine's reveal has generated unprecedented excitement, but looking more closely at Valve's disclosures, perhaps a more forward-looking strategic play has been overlooked. In producing a console-like box, Valve is looking to take the Steam audience beyond the PC desktop, but the fact that the ARM-powered Steam Frame can run x86 Windows games points to a much more ambitious strategy. Microsoft likes to say that everything can be an Xbox, but Valve's strategy is clearly very similar - and potentially more wide-ranging. The technologies are in place for everything to be a Steam Machine: phones, tablets, consoles and PCs.

Let's consider what Valve has already achieved and what happens next. The original Steam Machines launched in 2015 with the stated aim of allowing all Steam titles to run on Linux. Game developers did not support the endeavour with native Linux ports, ensuring that the initial hardware rollout was a failure, with SteamOS failing to gain traction. However, Valve didn't give up. If developers wouldn't step up, another solution was required. Over the following seven years, Valve shepherded an open source environment, helping to bring Windows games to Linux, with the Proton compatibility layer taking centrestage. Steam Deck was the first hardware launch built to exploit the technology, but Steam Machine expands Valve's reach even further. You no longer need Windows to play thousands of Steam games.

Now, with Steam Frame, Valve now feels comfortable to release a device that allows x86 games to run on an ARM-based system. The FEX compatibility layer works in concert with Proton to free your Steam library not just from Windows, but from x86 PCs entirely. In our conversations with Valve, FEX has an overhead of around 10 to 20 percent of CPU time - a slice that will only diminish as ARM processors become more powerful.

In the short term, ARM devices running Linux will benefit. So, this is good news in particular for Nvidia, quietly working on its N1 and N1X ARM processors for laptops, which contain potent RTX graphics cores. Previously reliant on Microsoft's Windows ARM translation layer, Nvidia now has options - assuming its Linux drivers are up to snuff. We can also assume that Qualcomm and others who don't hold x86 licenses will be very interested in this new technology too.

The proliferation of ARM elsewhere, combined with the dismantling of walled garden mobile stores, also puts Valve in a good position to bring Steam to mobile to phones and tablets. Of course, the limited horsepower of these devices means that less ambitious or older titles are better suited to this class of machine - but that's fine. The Steam back-catalogue stretches back decades. Already, classic PC games are running on phones and even Meta Quest headsets using emulators like Winlator, but Valve has the opportunity here to produce a more user-friendly experience to download and run those games. And presumably, it rates FEX over existing x86 to ARM translation layers like Box.

To cut to the chase, Valve has positioned Steam Frame as a streaming-first device, with standalone play via FEX taking a secondary role - but the point is that it has shipped the first iteration of an ARM-compatible Steam, and from there - assuming the technology is competent and performant - the potential is enormous.

Comparisons with Microsoft's "This Is An Xbox" strategy are inevitable - but perhaps more localised in terms of target devices. Having effectively lost the console war against Sony, the direction of travel with the next-gen Xbox console is to hybridise with a Windows-based PC, while porting its enviable catalogue of games to as many systems as possible. And for devices where this is not possible, the strategy is to use cloud streaming as a fallback.

Valve's alternative strategy is arguably more basic: get the Steam library running on as many different devices as possible via its translation layers, without needing to build enormously expensive, expansive cloud server infrastructure. Local play across more devices and streaming from home hardware may be the smarter move: Microsoft has Azure, Valve does not.

Of course, Microsoft has ARM aspirations of its own and has already deployed a wave of laptops based on the Qualcomm Snapdragon X Elite and its variants, though I found launch gaming performance to be extraordinarily bad and virtually unusable to the point where I returned my unit. I'm told game performance and compatibility has improved, which is essential bearing in mind the upcoming wave of new ARM-based Windows laptops, based on the Snapdragon X2 Elite. Qualcomm is making some big, big performance claims for the Snapdragon X2 Elite specifically - but hardware performance counts for nothing if the drivers and the necessary translation layers fall short.

However, I can't help but think that focusing on ARM for Windows or indeed ARM for laptops isn't the most ambitious use of the translation layers now available. Of course, handhelds could benefit from a wider range of available processors - and I'll be interested to see if Nvidia's upcoming chips end up in portable devices beyond laptops. Bearing in mind how frighteningly efficient Switch 2 is at lower wattages on an older architecture pared with an outdated process node, you can't help but feel that the sky's the limit with chips based on the latest technologies.

But it's the idea of Steam moving beyond the PC ecosystem entirely and towards phones and tablets that is intriguing, so what's the next step? First of all, whether it's Windows or SteamOS, I'm itching to discover just how performant the relevant translation layers are. Beyond that, I'd imagine that once the mobile walled garden ecosystems are fully opened up, we should look for the debut of Android and iOS Steam clients. Let's see how that pans out in 2026...