Yup. More and more manufacturers are actively showing loyal end consumers the middle finger by producing even less to drive up prices even more. Meanwhile other manufacturers are going bankrupt because who is buying cooling systems or new cases when no one can afford a new build? This is an active push against ownership and for tech feudalism.
It makes sense that they are reluctant to make investments. Building new data centers will presumably require more new RAM units then maintaining those that are already in existence. The large influx of new infrastructure is a temporary condition.
Regardless, a quick Google search indicates that about a half a trillion dollars has recently been spent on building new foundries in the United States. Since the release of ChatGPT in late 2022 coincides with the spike in demand, we are only just at the point where these investments could come into fruition.
Careful, “foundries” usually means CPU and maybe GPU. Anything memory related at the moment is just “announced” or “planned”. You could say Micron spent $20B, but that’s global not just US.
Besides Micron no one is planning to build memory fabs in the US.
It’s RAM, they can ramp it up in 2-3 years tops.
They just don’t want to. When they did it last time demand plummeted and we got cheap RAM.
For my 64 GB 6000 CL30 I paid 200€ last year. They don’t want a repeat of that, especially with the Chinese now catching up.
Yup. More and more manufacturers are actively showing loyal end consumers the middle finger by producing even less to drive up prices even more. Meanwhile other manufacturers are going bankrupt because who is buying cooling systems or new cases when no one can afford a new build? This is an active push against ownership and for tech feudalism.
It makes sense that they are reluctant to make investments. Building new data centers will presumably require more new RAM units then maintaining those that are already in existence. The large influx of new infrastructure is a temporary condition.
Regardless, a quick Google search indicates that about a half a trillion dollars has recently been spent on building new foundries in the United States. Since the release of ChatGPT in late 2022 coincides with the spike in demand, we are only just at the point where these investments could come into fruition.
Careful, “foundries” usually means CPU and maybe GPU. Anything memory related at the moment is just “announced” or “planned”. You could say Micron spent $20B, but that’s global not just US.
Besides Micron no one is planning to build memory fabs in the US.