If you click on a link and make a purchase we may receive a small commission. Read our editorial policy.

AyaNeo Air 1S/ AyaNeo 2S review - faster than Steam Deck but what about ROG Ally?

AMD's Ryzen 7840U delivers a new wave of powerful handheld PCs.

One area where AMD is simply untouchable is in its APU business. Having dominated the console space with custom x86/Radeon processors for consoles, the firm is doing it again in the PC handheld space. Steam Deck made the impossible real by allowing triple-A titles to run on a totally viable handheld system, and while Valve is facing a number of competitors in this space now, the bottom line is that it's inevitably an AMD processor at the heart of each and every one of these systems. AyaNeo has established itself as a manufacturer of quality PC handhelds and its latest generation of devices - AyaNeo Air 1S and AyaNeo 2S - are excellent, overall.

The AyaNeo 2S is very easy to describe. It's essentially identical to the AyaNeo 2 and AyaNeo Geek we've already reviewed, with the older generation Ryzen 7 6800U swapped out for the new 7840U. We're looking at a large, high-quality handheld here with a top-tier seven-inch LCD screen, Hall Effect sticks, responsive buttons and excellent I/O - you can't really argue with three USB-C ports (two top, one bottom). The AyaNeo Air 1S is based on the same processor, but it's an altogether different device, with a 5.5-inch AMOLED screen, a tiny form factor (think of it as a thicker Switch Lite) and twin USB-Cs - one top, one bottom. The screen is simply beautiful, with phenomenal contrast.

Integrated into these handhelds is the Ryzen 7 7840U, the top-end processor in AMD's latest line of processors, codenamed Phoenix. On the face of it, we're looking at a slightly tweaked version of the AMD Z1 Extreme chip found in the Asus ROG Ally. Both have CPU clusters based on AMD's latest Zen 4 architecture, while the same 12 RDNA 3 compute units make up their integrated graphics component. Even memory bandwidth is much the same, though AyaNeo says that 7500MT/s memory may be an option in its BIOS - but it recommends 6400MT/s for lower power modes.

Checking out the AyaNeo Air 1S and AyaNeo 2S in a 30m+ DF deep dive.
AyaNeo Air 1S/2S Asus ROG Ally AyaNeo Air Plus Steam Deck
Main Processor AMD Ryzen 7 7840U AMD Z1 Extreme AMD Ryzen 7 6800U Custom AMD 'Van Gogh'
CPU Zen4, Eight Cores, 16 Threads, Max 5.1GHz Zen4, Eight Cores, 16 Threads, Max 5.0GHz Zen3+, Eight Cores, 16 Threads, Max 4.7GHz Zen2, Four Cores, Eight Threads, Max 3.6GHz
GPU RDNA 3, 12 Compute Units, Max 2.7GHz RDNA 3, 12 Compute Units, Max 2.6GHz RDNA 2, 12 Compute Units, Max 2.2GHz RDNA 2, 8 Compute Units, Max 1.6GHz
Memory 32GB LPDDR5 6400MT/s 16GB LPDDR5 6400MT/s 16GB LPDDR5 6400MT/s 16GB LPDDR5 5500MT/s
Display 1920x1200 - 60Hz AMOLED/IPS 1920x1080 - 120Hz IPS with VRR 1920x1080 - 60Hz IPS 1200x800 - 60Hz IPS
Battery 50.25WHr/38WHr 40WHr 46.2WHr 40WHr

Clock speeds based on released specifications also look very, very similar between the Z1 Extreme and the Ryzen 7 7840U to the point where I believe that essentially, they are the same processor. Performance differences between them will come down to the surrounding hardware. During my testing, I found benchmarking the 7840U very difficult, as performance could vary quite dramatically from run to run, even in identical conditions.

In the benchmarks on these pages, I took multiple samples across multiple days and chose individual runs that seemed to be representative of standard performance, but even on max cooling there could be a six to seven percent delta. Even so, across these pages you may see instances where the AyaNeo Air 1S outperforms the meatier AyaNeo 2S, even though it's the same chip running on the same games at the same settings in identical conditions.

Then there's the software. AMD recently released a unified driver package for 7840U devices, but I found that this caused big issues for the AyaNeo Air 1S I installed it on. Performance dropped, and I had issues with the video outputs essential for Digital Foundry work. AyaNeo's suggestion was to stick to their verified driver, which was unfortunately somewhat old. So, while we've tried our best with the benchmarks, I'd be looking at them as indicative only of general output, though the power scaling tests should prove more illuminating.

We're kicking off with the 15W shoot-out here, where we should expect the Steam Deck to perform better than the 6800U and the 7840U relatively speaking. Larger chips typically require more juice to get the most out of them and the later AMD chips are designed more for laptops with a 28W power budget. The Deck's processor, on the other hand, is designed for 15W.

15W is important as it's where you still get more reasonable battery life, but it's safe to say that it's at 20-22W where we find the performance sweet spot of the 7840U. We'll be checking out power scaling on the next page. For this selection of data, I'm showing a selection of the benchmarks I took - there are a couple of games tested in the video where the results seem entirely beholden to either massive thermal issues or driver problems, and I'm betting it's the latter. One final note before we set off into the numbers: all machines are tested on Windows 11, except for Steam Deck, where we've benched from within SteamOS.

Forza Horizon 5 at 15W

For all the benches here, we have 720p and 1080p metrics, to get an idea of how performance scales when resolution increases at the same settings with the same power. We've also got new ROG Ally benchmarks - we borrowed a unit to get performance numbers from the most recent firmware as Asus took back our sample and failed to return it.

Results live up to expectations here, in that the latest AMD processor offers up a 31 to 34 percent performance increase over the Steam Deck, while the older-generation Ryzen 7 6800U sits in the middle. There is a small difference between the Z1 Extreme and 7840U results, but it's within the run-to-run variances I saw on all the 4nm devices. The Asus ROG Ally seems to drop back a touch at 1080p, but only by a couple of points. Interestingly, the stutter issue I saw at launch with the ROG Ally is still there, but isn't on the 7840U AyaNeo units. I saw a similar situation in the Cyberpunk 2077 benchmark too.

Forza Horizon 5, High Settings, 4x MSAA, 15W

Cyberpunk 2077 at 15W

In this benchmark, we're looking at Cyberpunk 2077 at medium settings with FSR2 reconstruction disabled, even though the game automatically engages it. Expect low frame-rates from handheld hardware then, but as you'll see in the video, careful tweaking can still deliver an excellent experience - even at a reconstructed 1080p. You can see how the scaling works here across the various AMD chips. Steam Deck is slowest, as you might imagine, but the difference vs the more potent Ryzen 7 6800U is insignificant, while the gains from the 7840U fail to match the impressive spec boost.

The situation is similar at 1080p, though with more of an emphasis on GPU performance, the 4nm parts gain proportionately more performance compared to their outputs at 720p. There's a problem with the Asus ROG Ally in that intrusive stuttering in the benchmark drags average frame-rate down and is impactful on those one percent lows. We saw this in the original ROG Ally review and it's still there on the present firmware.

Cyberpunk 2077, Medium Settings, TAA, 15W

Hitman 3 at 15W

There's no ROG Ally 720p score here as the device defaulted to 120Hz mode for some reason, nullifying results in our video-based capture workflow (and the device was returned to its owner before I could re-test). There are also oddities across the board here, in that the AyaNeo Air Plus's 6800U delivered stronger performance at 15W than the 7840U. You'll note that in the 1080p scores, the Z1 Extreme in the ROG Ally pushes ahead of the 6800U while the 7840U still lags.

This benchmark is indicative of the kind of frustrations I encounted in testing these devices. Performance at 15W seems variable to say the least, and while there can be big gains over the Steam Deck, it's safe to say that there can also be disappointments too. To get the most out of the latest chips, they really need more power.

Hitman 3, Default, TAA, 15W

Far Cry 2 at 10W

In the Asus ROG Ally review, we noted that the Steam Deck wiped the floor with the Ally testing at 9W. Subsequence firmware revisions have put that right to a certain degree, as you can see in this re-tested, much longer Far Cry 2 segment. Again, there was a capture issue at 720p with Ally here, but at 1080p you can see that the Z1 Extreme slots into place with the other 4nm chips. Even so, it is a touch disappointing to see that at full HD, the Air Plus's 6800U is up there with latest processors.

At 720p, there's clear distance between the Deck's Van Gogh chip and the brace of 7840U machines, with the 6800U closer to the Valve handheld than the 7840U devices. It should be stated that frame-times on all devices are highly variable - especially so on Steam Deck, which manifests here as absurdly high top one percent scores. Frame-rate limits are a good idea, even on legacy games.

Far Cry 2, Very High, No AA, 10W

AyaNeo Air 1S/AyaNeo 2S Analysis

From Assassin's Creed to Zoo Tycoon, we welcome all gamers

Eurogamer welcomes videogamers of all types, so sign in and join our community!

Find out how we conduct our reviews by reading our review policy.

Related topics
About the Author
Richard Leadbetter avatar

Richard Leadbetter

Technology Editor, Digital Foundry

Rich has been a games journalist since the days of 16-bit and specialises in technical analysis. He's commonly known around Eurogamer as the Blacksmith of the Future.

Comments