The giant chip makers’ big pact isn’t just a technical twist—it could mean some major changes for the PC market. Here’s what we still don’t know.
September 18’s big announcement of a surprise partnership between Nvidia and Intel could represent a sea change in the x86 computing landscape. There’s a lot to digest about it—and a lot of unanswered questions. From the just-announced Intel x86 RTX system-on-a chip (SoC) effort for PCs (bringing Nvidia graphics right onto Intel CPUs, in one combined, fast-communicating unit), to the $5 billion, around-5%-investment in Intel by Nvidia, this deal will reverberate across the coming years.
While the partnership has huge implications for the Nvidia-powered AI and data center landscape (bringing Intel CPUs into that fold) and for Intel’s general economic fortunes, I’m going to focus here on the consumer-centric announcement—the RTX SoCs. These products look poised to show up, eventually, in millions of PCs, fusing Intel’s high-performing CPU processing cores with an Nvidia RTX graphics “chiplet” on one chip.
This has the potential to shake up several segments of the PC market. Following the announcements, these are the five biggest questions that came to my mind.1. Is This a Revolution for Integrated Graphics?
In short: yes, though I would add a few asterisks. First, do these chips even count as integrated graphics?
An RTX graphics chiplet will be (literally) “integrated” into the same system-on-chip (SoC) as Intel’s CPU, fusing the two into one unit. This is a discernable difference from running a discrete Nvidia graphics chip, with its own memory and on its own module or section of a laptop motherboard, alongside a separate Intel processor. But that alone is not traditionally what we mean when talking about integrated graphics.
Intel and AMD have sold processors with, relatively speaking, weak integrated graphics for many years, dubbed things like Intel Graphics, Intel Xe Graphics, Intel Arc Graphics, and AMD Radeon Graphics. They’ve gotten better in recent years, but these solutions still handle graphics workloads at a level far lower than discrete GPU chips can (that is, the powerful graphics chips sold almost exclusively by Nvidia and AMD). This is the usual setup for most PCs that don’t require real graphics horsepower, which is generally reserved for specialist professional PCs and gaming systems. Traditional integrated graphics usually power entry-level, general-use, and business laptops.
However, that definition is changing. These Intel x86 RTX SoCs join a developing space that we don’t currently have one name for (see my colleague Brian Westover’s take, which ponders a new name for this kind of chip), akin to Apple’s M-series silicon and AMD’s Ryzen AI Max. They are not discrete graphics, but they aren’t the usual integrated graphics, either. In these new solutions, one SoC powers the processing and graphics workloads, offering higher performance than traditional integrated systems. Part of that power is down to the system as a whole and the graphics sharing a large pool of fast “unified” memory; it looks like the RTX SoCs will utilize unified memory, too.
All told, even if these chips go against our more classic use of the term, they are a kind of integrated graphics. Nvidia CEO Jensen Huang referred to them as such during the joint press conference with Intel CEO Lip-Bu Tan that followed the announcement, too.
Does this announcement herald a revolution in integrated graphics, then? I’d say yes, but again, a caveat applies. As mentioned, Nvidia and Intel are not first movers in this area, but once the fruits of this partnership bear out, they will likely be the biggest ones. They may not have opened the door, but this looks like a massive step in pushing the PC industry in this direction. If the new “integrated graphics” umbrella now covers these multipurpose SoCs, integrated graphics are evolving already. In the future, expect the term to be associated with higher performance than we were used to for decades.2. What Type of Laptops Are Going to Change?
This question is a natural follow-on to the integrated graphics discussion, because certain types of laptops are better suited to integrated versus discrete graphics.
