Yeah, these whims... They can also happen quite short-term. Like today with Nvidia, at 09:40 AM it was 819.72, and at 11:10 AM it was 779.70, and less than an hour later at 802.58.
Savings plans are way more relaxed in such regard. Like if set up for GOOG in late 2021, one would have been continously buying for 2 years for less than what it is now. Still risky, but not necessarily as crazy as putting in $3.000 at once, hoping for the best, and then looking at -50%. Something, which may not be much an issue for (institutional) trader with a spread portfolio, but quite an issue for some of us, who don't have more than these $3,000 to work with.
Anyhow, I survived this week, my first week at the stock exchange. I had some AMD, and went in at the beginning of the Nvidia Wednesday rally. My gains were something around +8% with AMD and +20% with Nvidia - or put otherwise, with about 3,000 Euro I gained about +400 Euro, and a few lessons.
Part of the gain already went into regular payments for ETF savings plans, part of which are Nvidia and AMD. So not really devaluing them, but at the same time also redistributing the wealth a bit through these ETFs, for other companies, including in U.S., Japan, Taiwan, Korea, and Europe.
Have we actually accomplished something with value from AI though? All I see is AI generated videos, pictures etc... I am still not really seeing anything that has actual value, and I am sure the power bills are starting to count...
We were almost better off with all the power going to crypto mining, because at least you once in a while heard of someone making some good money with their investment in mining hardware.
There arguably is quite some potential in regard to improved data processing, as such to begin with. The Magnificent Seven companies are all with products and services for the broad market. So their sales pitch turns around how awesome it is that an AI can find a fitting cooking receipt within a second, among ten-thousands of cooking receipts. The possible applications seem to go beyond that though. In example, when there is 16-bit sort-of-infrared 12K photo of a landscape, and when there are a number of such pictures, that adds up quickly to terabytes of data. And the possibility to process that pretty much at once, to determine mean ground temperatures, without needing to save the photos to a harddrive first, that may not sound much interesting, but improved hardware can help with it (albeit in some use cases perhaps not before next gen).
In any case, in crypto there is no added value of something provided. Like which GPU can crypto actually offer, or does it bring more jobs than just for a cleaner in a mansion on a tropical island? None, and not really, does it? Meanwhile companies usually bring something at least some people can make use of in their lives and or business (even if there currently may not be much more than to have a virtual librarian for cooking receipts), and such companies have some value, determined by their assets and revenue (and profit). That seems to make it less volatile than when missing out on the cute video promoting some other crypto, and 90% jumping ship to that one, leaving one easily behind without care that someone very trusting ended up losing 10+ years of their savings.