So they got all that money from Uncle Sam’s CHIPS Act only to lay off 10,000 employees and make themselves “lean”. Govt funded unemployment.
So they got all that money from Uncle Sam’s CHIPS Act only to lay off 10,000 employees and make themselves “lean”. Govt funded unemployment.
ARM and Qualcomm aren’t really involved with AI, and AI only makes up 15-20% of AMD’s revenue. Nvidia the one to watch out for, an entire 85% of their revenue is just AI and Mellanox. The Nvidia pump has been insane.
The elephant in the room is the semiconductor ETF trade. People are actively trading ETFs more than actual tickers (SOXL is routinely one of the highest volume tickers day in and day out, and it’s been like this for a couple years now). Used to be that NVDA and AMD correlated with crypto because people would flip them on that basis, but not really the case anymore due to all that but I digress - if NVDA tanks, it’s really overweight in all these ETFs, ETFs tank, managers dump every holding in the ETF, and a bunch of non-AI semi stocks will wind up getting walloped about as hard
That’s true, Nvidia’s fall will probably crash the whole market. SOXL tickers don’t trade in unison though, see how much Intel fell in the premarket without affecting other stocks. Of course, Intel is a tenth of the size of Nvidia.
Yeah, the tail could definitely wag the dog. NVDA is now in the top 3 holdings of both SPY and QQQ, so if something like 2 Metas or 1 Microsoft started paring back and skipped on a product cycle, Nvidia’s forward metrics would suddenly be garbage, and between those couple companies, suddenly >15% of market is tanking and dragging entire sectors along for a wild ride
Nvidia’s AI gambit is at least diversified to different kinds of AI. Even if (or probably when) LLM AI taps out, Nvidia will likely also be behind the AI tech that takes its place.
I honestly don’t think investors can tell the difference between LLMs+SD and more useful neural networks like audio and video filtering tools. All the money is in LLMs, anyways, I don’t think anybody is buying $1B of datacenter CPUs for the more useful kinds of AI.