Pat's forecasts for PCs were way wrong and that was a market he was in... so, why would we think his AI forecast is right?
For example, he says lower chip prices are good because demand goes up... but, not only would demand need to go up, it would have to go up by MORE then the reduction in price. Worse yet, the margin is lower demand has to go up by much more then the price cut to maintain margins.
So I for one, will not be adding to positions in the AI market esp., when the current valuations are about as bad as CISCO during the internet bubble... and the stock today is STILL lower then the peaks in 2000. Live and learn. Don't count on forecasters esp. when the track record is bad.