for me RAM is a perfected technology, new buses will come, more speed, but it will fundamentally be the same manufacturing process, same materials. The prospect is that LLMs will keep getting larger, more RAM will be required, and the prices will keep getting higher, or along the curve, while the demand will keep up with it because everything has RAM in it. Do you see a point in the future where the industry forks out of this, and there’s an alternative where the end user is not affected as much from the demand of this resource?
That’s just a more handwavy version of 1. Unless you are saying that new technology will come along that will make existing fabrication facilities capable of higher throughput without substantial upgrades/expansions to those facilities? It seems unlikely to happen any sooner than 1 does either way, unless there is a high readiness level technology that’s going to massively disrupt the established field of industrial engineering in the next year that I haven’t heard anything about
Then the current companies buying it will just buy more. So far there isn’t significant movement in the space to try and do more with less, and the current strategy is to just keep throwing more resources at it.
I’m not sure there’s a point where production could eclipse corporate orders with the current path things appear to be on.