“The demand is off the rails in the hyperscalers,” says Patrick Moorhead on $NVDA's results. “That demand is now going to the data centers that aren’t necessarily connected to the hyperscalers… so this increased insatiable demand to work on generative AI models continues.”