Because XT60 has current rating, but no voltage standardization.

EPS 12V is better. It’s good up to 300W easily, and two of them would work for any GPU and it’s no bigger than an 8-pin PCIE

The fun fact is that it actually costs more to do this for the licence at least ISA licence is more expensive than the core net list licence, which is more expensive than the implemented core design licence.

If that translates to the royalties, then Apple would pay more per chip than Samsung, and they are paying more than someone using a stock core.

Why? Because they can.

Because anyone whom wants an ISA licence to design their own core from scratch has a ton of money to spend and thus arm charges them more.

Too many pins makes too expensive.

Higher speed pins too support pcie and memory speeds that don’t even exist yet are too expensive.

There’s are reason that all the HEDT sockets are all super expensive.

And this problem is so important that AMD even started splitting up their server lineup into SP5 and SP6 to bring socket costs down.

god I love that both of these companies are basically tied in performance and value.

Socket lifespan is actually only relevant if you plan on upgrading within that lifespan.

If you buy into the 7800X3D right now, you will basically 100% have Zen5 to upgrade too, with a high likelihood of having Zen6.
Past Zen6 however it would be a mystery, but I wouldn’t count on it.
Which means that if you see yourself upgrading in 2 generations (having one generation gap) then the path makes sense.

But your current CPU is a 9700K and you are going 5 generations in 1 swoop. Even assuming you discount 10th and 11th gen due to them being lacklustre over 9th, you still are jumping 3 gens.

If it’s cheaper then go for it, but I’d say to stay away from the 7900X3D and 7950X3D as the multi-die X3D scheduling isn’t always correct.
AM5 also has a lot more weirdness with DDR5 and memory overall, but the nice thing about the X3D parts is that they are less sensitive to memory performance because of the massive cache.

Being able to get your hands on a future Zen6 part also sounds pretty good too.

Yes because what Nvidia is doing isn’t super special. Of course AMD will have an equivalent or better solution, so the question really should be “how many years behind” will AMD be.

They closed the gap significantly in raster perf. Power efficiency is pretty close, and so is area efficiency. AI mostly a software problem and AMD aren’t blind to this, and are very clearly investing a ton more into software to close this gap. (They just bought Node AI and absorbed all their talent)

The hardware is arguably better in many aspects. MI200 and MI250 are HPC monsters and MI300 is a chiplet packaging masterpiece that has has HPC performance on lockdown.

There’s a reason that no new HPC super computers are announced with Nvidia GPUs.

Nvidia has lead in AI, AMD has lead in HPC. Nvidia has lead in area efficiency, AMD has lead in packaging expertise (which means they can throw a ton more area at the problem with the same cost of Nvidia)