It was the E3 where we half-expected Sony to break cover with its plans for the next generation of PlayStation hardware, but come the hour, it was actually Microsoft that confirmed that its hardware engineers are hard at work architecting what Phil Spencer called "the next Xbox consoles". Implying that more than one new Xbox is in development right now is an interesting - and dare we suggest, deliberate - choice of words.
This was swiftly followed up by a report from Microsoft/Windows-focused website thurrott.com, reporting on internal MS roadmaps describing a 'family of devices' currently in development under the 'Scarlett' codename, set to arrive in 2020. Thurrott's insider stories along these lines are typically well-sourced and the timelines tie-in with the arrival of the technological building blocks that will enable next-gen hardware. But with Spencer talking about new hardware in the plural and the Scarlett project reported as more than just a single console release, we have to wonder what form the next generation is going to take, and what separates each of these machines from a technological perspective. After all, in the here and now, Microsoft often refers to its Xbox One 'family' - similar devices in many ways, but with radically different levels of rendering power.
There are further strands to weave into the discussion here, with Spencer revealing more in his annual E3 interview with Giant Bomb. It's well worth a watch, with the Xbox boss seeing next-gen as an opportunity to focus on higher frame-rates and to leverage the firm's existing work with variable refresh rate and 120Hz display technology. Spencer also sees the arrival of new hardware as the means by which CPU and GPU power can be rebalanced more in line with what we see on today's PCs, as opposed to the situation we have now where even Xbox One X - the most powerful home console - is pairing a six teraflop GPU with low-power x86 CPU cores originally designed for tablets.
At the same time, Microsoft also announced that Xbox would be moving into the cloud with the arrival of a streaming service, presumably the 'Netflix for games' model that crashed and burned so spectacularly with OnLive, reinvigorated with better tech and faster infrastructure. Again, in the Giant Bomb interview, Spencer talks about making Xbox games available to people who simply won't buy a console and who may not even own a TV. Microsoft isn't alone in this - EA announced an Origin streaming service for its titles during E3, and we're aware of at least two more industry giants who have yet to announce, but will be looking to move into this space before next-gen consoles arrive. Spencer's talk of an emphasis on high frame-rates is fortuitous here - despite retaining a healthy scepticism, we've actually had some good streaming gameplay experiences, but running the game server-side at 60fps or better is the only way to get anything like a low enough latency response to pass for a local experience.
So, with all of this information in mind, let's consider what could possibly constitute a prospective family of next-gen Xbox devices. There is some 'out there' blue skies thinking here, but we've never had a mid-generation console refresh as profound as PS4 Pro and Xbox One X, and actually delivering a worthwhile leap in graphics power at a console-friendly price-point is going be challenging to say the least - and that certainly informs some of the speculation here.
Option #1: A next-gen console and streaming box
Based on the information revealed so far, a very simple explanation for what Microsoft is up to could simply come down to a full-blown next-gen console, accompanied by a secondary family member that compromises of a cheap streaming receiver box with an Xbox controller.
To deliver a truly appreciable upgrade over Xbox One X (certainly in GPU terms), any next-gen machine will be rather expensive - we're talking about a cutting-edge 7nm processor, a lot of memory and a fast storage solution. If you can play the same games with a small latency and image quality hit and save a massive amount of cash via inexpensive streaming hardware, that could be a compelling alternative for a mainstream audience.
In terms of disadvantages, creating a streaming box does seem at odds with the intended market indicated by Phil Spencer, and the concept of a mainstream piece of streaming hardware may still be too early when the infrastructure required for a good experience remains patchy in many countries.
Option #2: A next-gen console with Xbox One X as the new base machine
The emphasis on higher frame-rates talked about by Phil Spencer opens up a potentially interesting situation, bearing in mind how capable the existing Xbox One X's GPU is. Perhaps Xbox One X could replace the S as the new base console via a hardware redesign - with a 7nm shrink of the X's Scorpio Engine and a possible move to a more efficient 256-bit GDDR6 memory arrangement. We would still have the issue of the under-performing AMD Jaguar CPU cores, but they would only need to run next-gen games at half the frame-rate. GPU-wise, doubling per-frame render time and perhaps running at reduced resolutions should allow the graphics side to keep up. Another key advantage would be that existing owners of the Xbox One X would see the lifespan of their console extended and their library would still transition seamlessly to the full next-gen box as and when they upgrade.
Disadvantages? First of all, AMD's Ryzen CPU technology is far, far faster than a mere 2x performance boost over the Jaguar cores in the current-gen machines. It's a true generational leap, and my concern would be that next-gen's most profound improvement in terms of core spec would not be fully tapped into if support for last-gen consoles persists for much longer than the transitional 'cross-gen' year we usually see. Secondly, for this strategy to gain momentum from developers, I suspect it would require Sony following the same principle, with the PS4 Pro becoming the new base machine. It's questionable whether the raw horsepower is there to make this happen, while Sony - in the form of Mark Cerny himself - has told us about the firm's commitment to the traditional console life cycle, suggesting that PS5 will offer a clean break with the past.
There's also the question of whether what is effectively a 'Slim' version of an existing Xbox One console would fit into the existing definition of a console 'family'. Xbox One S and Xbox One X definitely are of the same family - they run the same games, the same OS, and operate in the same ecosystem. I suspect that this would not be the case between a next-gen Xbox and the X, with the new machine offering a lot of new features current-gen hardware can't deliver. Despite the negative points, don't rule this one out though. Remember, Phil Spencer has only talked about "the next Xbox consoles" and a revised Xbox One X would fit the description in the same way that the Xbox One S would have back in the day.
Option #3: Two - or more - next-gen machines
Bear with me on this one. When Xbox One X launched at $499/£450, there was a lot of controversy about the console simply costing too much. My response? Prepare yourself for a next-gen console equally as expensive - if not more so. Moving from a 16nm to a 7nm process will make a more powerful machine possible, but the cost per processor rises - especially so in the short term. A faster GPU also requires more bandwidth and more memory, at a time when RAM prices are skyrocketing. On top of that, the existing storage standard - the 2.5-inch laptop hard drive - will likely be too slow to meaningfully feed a console with 16GB or 24GB of memory. Solid state solutions would do the job, but the cost is high there as well.
However, as Phil Spencer (and indeed AMD's product line-up) has indicated, the biggest increase to power delivered by the next-gen consoles will be via a much more powerful CPU - and the good news is that this component is not at all expensive. In terms of silicon area, we'd estimate that an eight-core, 16-thread Ryzen set-up at 7nm occupies a similar amount of space as the existing AMD Jaguar clusters at 16nm in Xbox One X. This opens the door to a very interesting scenario: two next-gen Xboxes of the same family, separated only by their GPU power, which would inevitably translate into varying in-game resolutions per box, and perhaps other graphical differences along the lines seen between Xbox One S and Xbox One X.
A cheaper Xbox that retains a fully enabled Ryzen set-up would run the same games as the full-power box, and in theory could do so for the entire duration of the next console generation - a role that Xbox One X could not deliver. This cheaper box could have a smaller amount of memory (meaning lower texture quality in-game), and could conceivably still use the standard 2.5-inch mechanical hard drive for storage.
The processor for the less powerful console could either be a unique, smaller design to cut costs or else it could be based on salvage parts from the full-fat machine - partly defective chips from the production line with GPU compute units disabled in order to make them viable (this process is used in virtually all graphics cards). Disadvantages? Any variances between the models in memory allocation could potentially restrict developers and hold back the fully enabled box from being everything it could be.
Option #4: The Xbox One X dev kit model
As we've already explained, not every processor that rolls off the production line is perfect. To increase the yield of useable chips, Xbox One X has 40 active GPU compute units when in actual fact, there are 44 in total on the silicon. Indeed, every current-gen console has either two or four compute units disabled, whether you own a Sony or Microsoft machine. In the case of Xbox One X's Scorpio Engine, the fully enabled design is actually put to use in development hardware, which also features 24GB of RAM up against the standard console's 12GB. It also has a 1TB SSD, which obviously offers a speed bump compared to a mechanical laptop-style drive.
This set-up potentially opens the door to a new Xbox 'Elite' console, with a performance boost and much faster loading times - either through caching to the extra RAM, or else through the inclusion of a higher specification storage system. If pursuing high frame-rates and accommodating high refresh monitors is an objective of the new Xbox, it stands to reason that there'd be a minority of users who'd be looking to get as fast a box as possible - and would be willing to pay a premium for the privilege. Toss in a second-gen Elite controller and I suspect that it would find a sizeable audience to make the enterprise worthwhile.
The disadvantages are self-evident, however. Hardware that qualifies as a genuine generational leap over the machines of today may well command a price premium anyway, and any kind of Elite console would be even pricier and performance increases may be limited - or non-existent - depending on the title.
Option #5: New boxes with staggered release intervals
Up until this point, we've been assuming that the Thurrott report is entirely on the money and that more than one 'Nextbox' will appear in the mooted 2020 window, but intervals between releases is obviously an option for Microsoft. When I met up with Microsoft corporate vice-president of the Xbox and Windows platform, Mike Ybarra, he outlined a vision of an evolving family of devices based on the way that people upgrade their smartphones. Microsoft believes that users want the best hardware as soon as possible, and this doesn't fit the existing notion of a console generation lasting six or seven years.
In terms of how this could apply to the next generation of Xbox devices, it's pretty straightforward: AMD's big leap in CPU power has now been achieved, and future Ryzen processors - for the next few years at least - are likely to be more iterative advances; the Intel model, in effect. However, graphics will continue to evolve and scale and a substantial leap in GPU power will be achievable with every new process shrink. It's 7nm technology that makes next-gen machines possible in 2019/2020 but the next process node beyond that should offer something in the order of a 2x increase to GPU performance. Crucially, this would still keep the 2019/2020 box in contention - it would just operate with a reduced visual feature set.
New console boxes with each process shrink are likely to happen anyway, but is this what Phil Spencer really meant during his E3 2018 media briefing? In this scenario, there'd be no real need to specifically float the idea of another Xbox model due in 2022/2023 when the one due in 2019/2020 is the only one really worth talking about.
Next-gen: the challenges and the opportunities
We need not worry about the CPU side of the next-gen machines. AMD's Ryzen offers phenomenal performance for the amount of silicon area it uses, it's power efficient and it's broadly competitive with the best on the market - Intel's Core architecture. As we've discussed in the past, the potential here for deeper, richer, more complex games - or indeed more titles running at 60fps - is mouth-watering.
The issue is that while AMD can offer a generational leap over the graphics performance of the base PlayStation 4 (effectively the main target current-gen target platform), delivering the same increase in performance over PS4 Pro and especially Xbox One X is far more challenging. Doubling X's GPU power is achievable, but if the aim is also to double frame-rate, the ability to deliver richer graphics is therefore more limited. If the objective is to lock to native 4K and deliver hugely improved visuals, again, 2x Xbox One X performance isn't enough.
My best guess? The so-called 'FauxK' upscaling techniques seen today will be back, refined and improved for next-gen - be it through developers' ingenuity or via dedicated hardware. A 6x to 8x leap in GPU performance over the base PlayStation 4 can be delivered, but servicing a 4x leap in resolution doesn't leave a huge amount of remaining overhead for radically improved visuals. The jump from OG Xbox to Xbox 360 required a bleeding-edge GPU to deliver a 3x increase in pixel density with enough headroom for much improved graphics. Meanwhile, the leap from PS3 to PS4 saw only a 2.25x increase in pixel-count. In this respect, expecting native 4K across the board from next-gen doesn't seem likely.
More probable is innovative use of custom hardware. There are already rumours of Sony collaborating directly with AMD on its Navi architecture, and PS4 Pro offered some fascinating technology in allowing a console GPU to punch well above its weight in supporting 4K screens, even though third-party developer buy-in was relatively limited. Next-gen is a crossroads for game technology and this time around, we may well see alternative visions from Sony and Microsoft etched directly into the silicon, with each taking strategic bets on the future of graphics via the integration of custom hardware. If the make-up of the current-gen machines was defined by two vendors using very similar AMD technology, next-gen - by necessity - may be a little more interesting, and dare we say it, a little more exotic? And the idea of a new, more tightly focused series of consoles could add further spice.
Digital Foundry specialises in technical analysis of gaming hardware and software, using state-of-the-art capture systems and bespoke software to show you how well games and hardware run, visualising precisely what they're capable of. In order to show you what 4K gaming actually looks like we needed to build our own platform to supply high quality 4K video for offline viewing. So we did.
Our videos are multi-gigabyte files and we've chosen a high quality provider to ensure fast downloads. However, that bandwidth isn't free and so we charge a small monthly subscription fee of £4.50. We think it's a small price to pay for unlimited access to top-tier quality encodes of our content. Thank you.Support Digital Foundry