Nvidia was also shipping its 5090 and 5080 cards back then.

But yes, it's hard to point the exact reason why the market had such big swings in so little time.

nVidia and Apple seem to gobble up most of the wafer supplies currently, so AMD tries to make do with what it can with whatever is left over.
Case in point, nVidia managed to get TSMC to dedicate supply and fab capacity and even a "special" optimized node for nVidia... That doesn't happen without guaranteed volume.
AMD had to resort to using a more common and less optimized node and stockpile chips.

Even DRAM nVidia seems to be scoring most of the contracted supplies at the moment due to Server/A.I markets which is driving up prices of GDDR memory as DRAM is a commodity so it's priced based on supply/demand.

I guess the bright points for AMD have actually been on the integrated graphics side, their integrated graphics in the Steamdeck, ROG Ally and Consoles have been a resounding success... And that is likely what keeps AMD investing in GPU technology to keep those design wins happening in the future which brings in 100's of millions of dollarydoo's.

Let's see what they can cook up with GFX13/UDNA/Radeon 10070.... Which is their 16th graphics architecture, but I don't expect a big turn around in terms of marketshare for many years yet.

...And that probably puts Intel's GPU efforts into an even more tumultuous predicament if they can't get marketshare either... But Intel's long term plans is probably the A.I market.

I think I heard it from HArdware Unboxed that the supply of 9070 cards won't increase notably until the second half of the year, which gives us an idea of how busy is TSMC if AMD has to wait this much to get access to more waffers.

At least supply for the 9060XTs seem ok, as there are still cards available at MSRP over here. But maybe the demand isn't as high as one could expect. Hard to know.

As for Intel... the recent talks that they won't make products that don't guarantee at least a 50% profit margin put their GPU business at risk, once again.

There are rumors that Samsung is trying to woo Sony and AMD away from TSMC for next-generation PS platforms (console + handheld), offering one of their 2SF nodes as an alternative to TSMC's N3X. Maybe that will pan out for GPUs, too.

Let's see. That's another area of the industry where more competition is not only welcome, but necessary.

Laptops need a shakeup... Massively. Was looking at upgrading my couple year old RTX 4060 laptop with Intel 12400H CPU and claim it on tax... I paid $900 AUD for that... That same notebook today but with an upgraded 13420H CPU is now $1500.
May need to wait a few more years to get a notebook with better than an 8GB GPU at a sensible price.

Budget gaming laptops are still plagued with RTX 2050 4GB cards... But paired up with modern CPU's.
............You may actually be better off with Ryzen integrated graphics in some instances over the RTX 2050 4GB.

PC Gaming hardware is just borked.

I wonder what price range of GPUs people who used to buy $200 or $300 graphics cards (e.g. GTX 1060 or RX 480) are using today.

That was probably my range, although my previous GPU I didn't buy used was the GTX 770 (I later upgraded to a used GTX 1650 Super after a brief adventure with a used GTX 970). In my current PC, I have an RTX 4060. It was terrible value, but that was really the price range I was aiming for, and I really wanted ray tracing capabilities, even if they're quite limited. Turns out going Nvidia was the right call, because the main demanding game I play is Cities: Skylines 2 anyway, and it's been much more pleasant to look at after I discovered DLAA. I probably won't be upgrading to the latest gen of GPUs, but the one after I might, but it'll probably be an xx60 series GPU again.

I think most of us have dropped down to the value segment.

Some will have increased their budget to $350-400, with a few managing to go up to $500, but I think most will have stayed in the same budget and just get a less powerful card.