I agree that CE will be affected for years, but for a different reason than the implied shortage. The continuation of the AI boom is dependent on continued financing and so far a lot of the money is flowing in a loop (or not flowing at all). Nvidia invests in an AI company and then the AI company agrees to buy Nvidia GPUs, but as you say, there is no power to run the beast. They hype causes gullible investors to throw money at the AI company and by association at Nvidia, which has a completely ridiculous market cap. The timeline is too long for investor patience to last, even if there is some magic application out there that could eventually produce enough revenue to fund the monster. When the house of cards collapses, the dot com bust will look like child's play and the ensuing recession will have an effect on CE for years to come. Take pictures and invest wisely.
I wasn't exactly intending to enumerate all the causes of the likely years-long implosion of consumer electronics.
I agree with the description of the AI "boom" as a venture capital scam, see here:
https://pivot-to-ai.com/2025/08/23/...e-think-the-ai-bubble-keeps-going-until-2027/
And I think it will indeed keep going into 2027. If it was based on investment fundamentals in any way, it would already be over. It is not gullible investors exactly, but self-interested venture capital scammers (that's their only "innovation": endless fraud) AND plenty of idiots AND people just jumping on the bandwagon because the markets and institutional investors function on FOMO.
The memory and storage supply implosion right now is caused by AI "hyperscalers" (people hoarding chips for datacenters that aren't built and for which there is no profitable use) buying out the entire supply of memory manufacturers - there are only 3 meaningful manufacturers on earth for RAM, which was already a very corrupt cartel. When AI collapses, the likes of Micron will still have no consumer supply, so prices will stay elevated for some time (and they will try to keep them up even when supply increases as there is no competition).
"Crucial" is Micron's consumer business, which they've ended because they're making so much money on datacenter HBM, which is not usable for anything else (similar to crypto mining chips, which were sold to many of the exact same people pushing AI).
Smaller electronics retailers and product lines will be crushed because with DDR5 RAM at 5x price vs. a year ago and fast SSD storage at something like 3x, a lot of consumers simply stop buying certain things entirely. As I said the other day, I'd have bought a 2nd 4TB external SSD the other day, but it has doubled in price. It's something I could use, but will get around until I have absolutely no choice.
Anyways yes, the recession from the AI crash will also further implode consumer demand and buying power, added on top of the tariff nonsense and the monumentally stupid imperial ambitions in the middle east. Everything will be worse for consumers for a long time.
Reference material/comments:
Here's a more recent article on the bubble, the timeline, the years required to rebuild supply lines, and "hyperscalers" buying more chips than they can use purposefully:
https://pivot-to-ai.com/2026/03/27/the-ai-bubble-and-the-global-polycrisis/. Ed Zitron has dug into the numbers a lot, and only 1/3rd of claimed data centers are actually being built:
https://www.wheresyoured.at/the-ai-industry-is-lying-to-you/.
OpenAI shut down its Sora video slop model the other week; of course, video models are the pinnacle of setting money on fire. They still only make worthless memes, but they cost astronomically more to run than text or still image generators. This is one of the first real cracks in the glass.
I think the "Subprime AI crisis" analogy is a good one (Ed first brought it up in 2024, but his most recent article breaks down a ton of numbers here:
https://www.wheresyoured.at/the-subprime-ai-crisis-is-here/).
