The announcement of Ryzen 7000 mobile and the GTX 40 series for laptops brings us into a new chip generation, but also at the end of an era where naming schemes and nomenclature for laptop processors were accurate, intuitive, and relatively simple to understand. We won't sugarcoat it: the names AMD, Nvidia, and Intel have come up with for their new chips are terrible. We find ourselves double-checking what a processor actually is because its name no longer says anything useful or informative. We thought we were seeing improvements after years of horrible naming conventions, but we're going backwards. So, what happened?
The end of an era of openness
Back in the old days, laptop CPUs and GPU naming schemes were terrible. You might think a GeForce GTX 980M was just a mobile version of the GeForce GTX 980, but in fact, it had just 75% of the cores. Meanwhile, Intel and AMD CPUs had a totally different nomenclature compared to their desktop chips. This prevented assumptions about the relationships between similarly named desktop and mobile parts, but it meant you had to learn two different naming schemes for each brand.
Things started to change in 2016. With its GTX 10 series, Nvidia began to name laptop GPUs similarly to how it named its desktop GPUs. A GTX 1080 was a GTX 1080, whether it was in your desktop or your laptop, with the same core count and memory amount. When Intel launched its Coffee Lake CPUs in 2017, the company also introduced eighth-generation H-class CPUs that mostly followed the same naming scheme the desktop eighth-generation chips did. AMD never completely unified the nomenclature of its desktop and laptop parts but did make significant progress in developing a decent naming scheme.
A GTX 1080 was a GTX 1080, whether it was in your desktop or your laptop.
Of course, this system wasn't perfect. A GTX 1080 in a laptop is always slower than a GTX 1080 in a desktop because laptops have stricter power and heat limits. That means clock speeds have to come down, leading to lower performance. Different laptops have different power and cooling capabilities, so one laptop's 1080 wasn't always the same as another's, for example. This was a key complaint about mobile components when companies did away with M prefixes and wholly unique nomenclature.
I strongly disagree with those complaints because naming schemes aren't just about performance but what you're buying. You can and should expect to pay lots of money for a high-end RTX 3080 GPU, whether it's in a desktop or a laptop, and you can also expect it will be a higher-end GPU relative to other options. The naming was both clear and honest, and it worked.
However, that's now dead with the announcement of next-generation CPUs and GPUs. The RTX 4090 mobile is probably one of the most high-profile examples of this new nomenclature. It's just an RTX 4080! Intel hasn't gotten much flak for its situation with 13th Gen, but that's probably because only higher-end CPUs like the 13900H (which has 10 fewer cores than the Core i9-13900K and 13900HX) have notably bad names. I'm particularly worried about AMD's naming scheme catching on, as obfuscating architectural differences is especially misleading. AMD even made a whole generation exclusively for its APUs just to avoid confusion about architecture and series names, and now, most Ryzen 7000 mobile CPUs don't use Zen 4, instead using the older Zen 2 and 3 architectures. Accuracy has gone out the window, and as a user, names are no longer intuitive.
Companies want to eat their cake and have it too
There are a couple of reasons why Nvidia, AMD, and Intel are all moving away from good nomenclature at the same time. Mainly, it's getting harder and harder to squeeze efficiency gains out of each generation of new hardware. Laptops live and die on how efficient they are because, as previously discussed, there are power and heat limits. It's why new desktop CPUs and GPUs are increasing in power, which is the only way to meet expectations of continuously growing performance on the higher end.
It's just not economical to put a full 4090 into a laptop because its clock speed has to go way down to meet power and heat limitations. When power is limited, high-end processors can't stretch their legs, and they end up being about as fast as lower-end models. This has always been an issue, but now it's worse than ever before, which becomes a huge problem when you're naming desktop and laptop parts similarly.
So why doesn't Nvidia just limit itself to the 4080 and leave the 4090 off the table? It wouldn't be unusual — there wasn't a 1080 Ti, 2080 Ti, or 3090 for laptops. In the same vein, Intel didn't need to name a 14-core CPU the same as one associated with a 24-core processor, and AMD doesn't need to use the Ryzen 7000 branding for CPUs that belonged to other series.
These companies want to use the high-end branding they've been using for years, and they're not going to let the technical specifications get in the way.
Ultimately, it just comes down to branding. These companies want to use the high-end branding they've been using for years on desktop and laptop components, and they're not going to let the technical specifications get in the way of that. They can even use higher-end branding more freely for components that wouldn't otherwise fit the requirements. It was probably inevitable it would end up this way, no matter how great or small the efficiency gains were, but steadily decreasing generational improvements certainly sped things up.
The result is the worst naming system we've ever seen. It's similar to what we had in the pre-2016 era, except without the prefixes that tell you mobile parts aren't the same as desktop parts. Instead, seemingly minor number and character differences now denote completely distinct processors in the most important ways. Ultimately, we're all left wondering if what we're paying for is actually worth the money.
Either add an M or change the names
I get it, naming a processor is difficult. It's hard to express all the little intricacies in a name with four or five numbers and a letter or two. However, that's not a license to act as if names don't mean anything. At the very least, I expect product names to convey the value of the part in question. The cost of producing an RTX 4080 is much lower than that of an RTX 4090, but I'm sure an RTX 4090 laptop GPU (which is actually an RTX 4080) will be priced based on its name, not its specifications.
If Nvidia, AMD, and Intel want to use these high-end and current-gen names, they need to add a clear marker that signifies what's what. I think we'd all be fine with the RTX 4090M, even if it still costs an arm and a leg. Right now, we're getting the worst of both worlds, and my biggest fear is that this scheme will be so profitable that it's permanent. On the other hand, lots of terrible decisions have been reversed or corrected with enough criticism, and I'm hoping next generation, we'll see a course change.