It’s been generally expected that AMD would continue to provide silicon solutions for both Sony and Microsoft in their next generation of consoles. We’ve seen numerous leaks and rumors that pointed in this direction and Sony has confirmed that the PlayStation 5 is expected in 2020. What we haven’t heard is formal confirmation of the fact from AMD — until now.

Speaking on CNBC (Overclock3D), AMD CEO Lisa Su confirmed that her company is “working with both Sony and Microsoft on consoles” and expected to provide the “secret sauce” for both designs. Right now, what we’ve heard is that the CPUs in both consoles are likely to be Zen derivatives, while the GPU for the PS5 is said to be based on AMD’s Navi — which may itself have been heavily influenced by Sony’s PlayStation 5 plans.

Assuming these plans are accurate, we can assume one thing for certain about the two upcoming consoles — they’ll pack significantly more CPU performance. With the PS4 and Xbox One, Sony and MS both pitched an argument that implied CPU performance was less important because the integrated GPU was far more programmable and could take over many tasks formerly handled by the CPU. The success of these platforms, including recent hits like the highly rated Spider-Man for PS4, is proof that the consoles can deliver great gaming experiences — the Xbox One X is a better gaming system than any you can build from scratch. This was particularly true during the cryptocurrency boom when GPU prices exploded and effectively broke the DIY market. But even allowing this point, it’s still true that the Jaguar core was designed to serve as a mobile CPU for low-power netbooks intended to outperform Intel’s Atom, but not to deliver the same performance as a traditional ‘big core’ x86 CPU. Some of the Xbox One and PS4 performance peculiarities, like the huge L2 latency when reading from one CPU cluster to the other, are an artifact of design decisions made for the low-power mobile market.

I once down-clocked a Kaveri and dropped it to a single memory channel to perform the closest clock-for-clock comparison against Kabini (Jaguar) that I could, even though the half-speed L2 cache on Jaguar made it impossible to compare the chips in an apples-to-apples configuration. The two CPUs were surprisingly competitive in terms of overall performance, though Kaveri was faster overall. But Ryzen is much faster than Kaveri — and that gives AMD some room to maneuver when it comes to picking CPU characteristics.

If Sony and Microsoft want a true eight-core SoC (as opposed to two distinct quad-core CPU clusters), they can have one with . Its vastly increased efficiency relative to Bulldozer or Jaguar means that both companies will have the option to keep clocks (and therefore power consumption) low, allocating more power for the GPU cluster if desired. If we assume that the gap between Kaveri and Kabini was just 1.15x clock-for-clock and the gap between Ryzen and Kaveri IPC corresponds to the 1.52x AMD claimed at the Zen launch, then AMD has the opportunity to deliver a huge efficiency improvement without touchingthe clock speed. This last point matters because keeping CPU clocks low is a critical way to tamp down on power consumption, allowing more horsepower to be diverted to the GPU when necessary.

Without knowing more about Navi, it’s hard to speculate on what kind of GPU capabilities might be baked into the platform, but there’s a chance that the gains here could be significant. AMD has been using the same fundamental GCN architecture since 2012. The company has made the decision to align its product roadmaps with the console manufacturer timelines and needs. Given the desperate financial straits, AMD was in five years ago, this decision made a great deal of long-term strategic sense and may have even saved the company from bankruptcy. But it also means that this PS5 / Xbox Next technology ramp is probably our best chance to see a major shift in graphics technology.

Sony’s smartphone division continues to turn in unimpressive performance, but the company remains the market leader in image sensors. Its latest design, the IMX586, promises a leap in image quality by dramatically increasing the resolution to 48 effective megapixels (8000 x 6000), which Sony says is the highest pixel count in the industry.

Image quality isn’t simply a matter of adding more megapixels — that can be counterproductive, with smaller pixels leading to noisy photos in low light. The 0.8-micron pixels used in this sensor will be the smallest on the market, in fact. But Sony says it’ll get around this by using a quad Bayer color filter array and allowing each pixel to use signals from the four adjacent pixels, which supposedly raises light sensitivity to the equivalent of a 12-megapixel image captured with 1.6-micron pixels, the Verge reported.

Phones from Nokia’s 808 PureView in 2012 to this year’s Huawei P20 Pro have experimented with similar pixel-binning techniques on sensors with 40 megapixels or more, but Sony’s IMX586 is likely to be a more mainstream solution. Sony is keeping the size down to 8mm diagonal, meaning there won’t be the need for a huge camera bump, though the lens in front of the sensor will of course play a big part in the camera’s ability to resolve an image. The focus is also on producing usable 48-megapixel images rather than downsampling by default — this might not be incredibly useful for everyday snaps, but it should at least allow for much better digital zoom.

You can expect to see the IMX586 on smartphones next year; Sony is planning to start shipping samples this September, with a unit price of 3,000 yen ($27) each.