The handheld PC gaming space is about to get a lot more interesting. For years, AMD has had that market essentially to itself, powering everything from the Steam Deck to the Asus ROG Ally with its custom APU silicon. That monopoly looks like it has an expiration date: early 2028.
A leak from tech journalist Erdi Özüağ points to Q1 2028 as the target window for Intel's first processor featuring Nvidia graphics chiplets, with CES 2028 floated as the potential launch stage. Intel and Nvidia have already confirmed publicly that they're working on combining Intel CPU architecture with Nvidia GPU chiplets, so this isn't speculation about whether it happens. The question has always been when.
Now there's an answer, and the timing is loaded with implications.
Pay less for your games.
Get discounts up to 80% off
Why 2028 is shaping up as the year APUs actually matter
Here's the thing: 2028 isn't just when the Intel-Nvidia chip is supposed to land. It's shaping up as the year the entire APU market gets a serious upgrade across the board.
Nvidia has already confirmed a Rubin-based version of its own RTX Spark CPU-GPU superchip is coming in 2028. RTX Spark is Nvidia's play at building its own complete SoC for handheld and portable devices, and the second-gen Rubin version would bring it in line with whatever desktop GPU architecture is current by that point. Rubin desktop cards, potentially branded as the RTX 60 Series, are currently rumored for late 2027 or early 2028, which lines up neatly.
AMD isn't sitting still either. Chips codenamed Grimlock Point and Grimlock Halo are reportedly targeting 2028, representing the generation after the upcoming Medusa Point and Medusa Halo APUs due later this year or in 2027. So AMD would be entering 2028 with its third consecutive APU generation designed specifically with handheld gaming in mind.
Three competing APU platforms in a single year. That's not a normal hardware cycle.
What the silicon roadmap actually means for performance
The manufacturing side of this matters as much as the chip designs themselves. By 2028, TSMC's N2 node will be well into production, and its A14 process could even be in the mix. Intel's 14A node is also expected to reach full production by that point.
What most players miss in these chip announcements is that efficiency matters more than raw performance in a handheld context. A chip that can push 60fps in demanding titles at 15W is more useful in a Steam Deck-sized device than one that hits 80fps at 30W. The move to next-generation nodes should help on both fronts simultaneously.
For PC performance optimization in the meantime, our Once Human PC performance guide covers how to squeeze the most out of current-gen hardware while the next wave arrives.
The Steam Deck 2 angle nobody wants to ignore
Valve has been transparent about its approach to a Steam Deck successor: it won't release one until there's a genuinely significant leap in both performance and efficiency. Not an incremental refresh. A real jump.
2028 looks like exactly that moment. With multiple competing APU platforms arriving simultaneously, Valve would have options for custom silicon that didn't exist before. The original Steam Deck used a custom AMD APU, and there's no reason Valve couldn't negotiate something similar with whichever vendor offers the best thermal and performance profile for a portable form factor.
The one variable that could complicate all of this is memory pricing. The current DRAM market has been rough, and if costs haven't normalized by 2028, even the most impressive new APUs could end up in devices priced out of reach for most players. That's a real concern worth watching as these platforms get closer to launch.
For now, the gaming guides hub has everything you need to get the most out of whatever hardware you're running today. And if you want to see how modern PC graphics settings scale across GPU tiers right now, the 007 First Light PC settings guide is a solid reference point for understanding what next-gen APUs will eventually need to match.
