Worldwide chip pricing and it's effects on the overall economy are fascinating. I expected by now somebody with a gen -2 class VLSI plant would swing into action and make <thing> for 2/3 the cost of the majors, and clean up in volume as the market absorbs the price shock. But no, instead it suits everyone in the pipeline to whine about it, but mark up prices instead.

I am guessing the other side of this, the price drops will happen but slowly, and just like gas pricing, the profit is in rapid reaction to shortage and slow reaction to competition returning.

Chip pricing a sawtooth would make a LOT of money for somebody.

Well, you are working from all sorts of misguided assumptions about how you convert, the yield, the efficiency, the shared capital, and history.

Old gen logic barely exists, and won't convert. The tools are wrong so it wouldn't yield. And there isnt the expertise (on tool recipes and integration) to do it.

Historically, memory is a sawtooth business. But history isn't a great guide here: 10 years ago, a new plant added meaningful capacity, as it came with a shrink of about 30%. Today....it doesn't. So it take huge capital to add a very small amount of capacity.

You can dream that things like 3d dram and 4f cells will help, but they are unlikely to offer enough with demand.

And finally, everyone running a dram plant has lived through these capacity boom bust cycles and the consequently layoffs and pay cuts. I suspect they are happy to take money this time. I would be in their place.

I do hope we see a new dram or dram-like designs. And that there are more efficient dedicated ai processors. And that models themselves become more memory efficient. But I don't really believe any of these come soon.

Even on previous gen nodes it takes a year or two to bring a chip to full production.

Unfortunately from what I understood a production line that makes "normal" digital chips generally can't make memory cells with concrete physical differences in the process pipeline

The memory crunch started October 2025. DDR6 will come out in 2027. At that point you're investing new capacity into an obsolete technology.

If I was a memory company, I would try to bring DDR6 to the market as soon as possible. DDR6 allows high end CPU based server platforms to reach the memory bandwidth of an A100 or half a H100 without the costly HBM.

I think the DRAM manufacturers understand this is a bubble, and the massive orders won't continue long enough to make spinning up matching new production worth it. All they can do is sell everything they can produce and let the market stabilize to that level.