All good points. I will just add this to @Gregmal's excellent point, it is very very difficult to really pin point on which valuations are out of whack to the upside etc. Barring the really crazy time of the spac and tech. mania from 2021 it has been hard throughout the last decade in doing this. In late 2022, NVDA touched below $120 or so for around a month - that is some 15 months ago. Estimates are for it to print $20 EPS in 2024 - buyside expect at least 10-20% more. Back in Oct. 2022, 2024 EPS estimates were $5.5 or so. So the earnings power has gone up almost 4 times in last 15 months with stock up close to 5 times. Until a month or 2 ago, stock was also up around 4x from that Oct. 2022 bottom. All it says its very hard. For the analysts who follow it, for most of the company management involved as well - it is very hard. I am not arguing whether this situation for NVDA sustains or shoots higher or lower. Given its stronghold, everyone invested could lay out several reasons it could (CUDA, 95% market share, etc.). While many could say otherwise (GM's very high, pricing will go down as AMD enters, inference foothold will not be as strong as training, etc.). My take has been it is too hard for most investors to forecast. Best is if someone has it to keep holding it or the next best is to not be bothered about it or the next best depending on one's ability is to trade around it - that is what most of the big pod shops do - but one needs focus around this trading strategy imo. Anyways, for individual investor one cannot really have staunch views on such stocks where things keep on changing so rapidly - rather the views need to be updated constantly.
@Spekulatius on comparing AAPL with KO in 1998. KO was around 45-50 times forward PE. Since then sales, EPS has grown around ~3,5% resp. Total price CAGR has been 4-5% or so including 2% div. yield. While SP500 has total return CAGR'ed some 7-8% during this time. While AAPL at 28 times is high and all and is not value. If one looks at how iPhone has behaved last few years (with units constant and MSD price increases), and services growing HSD/LDD, overall GM's increasing significantly as service margins are double the product margins, and 3-5% buyback. There are risks to this as well with services regulation, China competition, no real upgrade value of iPhone. Keeping these aside for a bit, if one assumes overall sales increase 4-5%, buyback add say 3-4% with 0.5% from div. yield, one can get around 8-10% and then assume a multiple degradation from say 28 to 20 if the moat still holds over a reasonable period of time. It still could be ~7% total CAGR with not much tax consequence esp. for someone of the size of Buffett - esp. for the size and time of life in Berkshire balance sheet right now vs. back then.