LOADING
LOADING
Jalapeño is not a Nvidia obituary. It is OpenAI's attempt to shift high-volume inference from a supplier-priced bottleneck into a stack it can optimize and bargain around.
In Brief My conclusion: this is not a normal memory cycle anymore. The market is still right to discount generic DRAM cyclicality, but it is probably underpricing the durability of scarce AIlinked mem…