bitzu101 wrote on 2026-07-10, 17:09:
ElectroSoldier wrote on 2026-07-10, 15:25:The AI bubble will burst as soon as the supply catches up with demand.
Thats why the companies are not racing to build new manuf […]
Show full quote
The AI bubble will burst as soon as the supply catches up with demand.
Thats why the companies are not racing to build new manufacturing equipment to increase output.
Companies are dumping a lot of investment into tech that just isnt returning any money for them.
They "know it will" but for now its just a hope, in much the same way as the early .coms did in the 90s.
The earliest it can happen is late 2026, but more likely to be 2028.
where is this demand? think there is oversupply now as it is.
Who:
Microsoft, Google, Amazon, Meta, Apple and Tesla.
Why:
Massive, gigantic training clusters, Inference farms, AI cloud farms, recommendation engines. This is 70-80% of global demand.
What:
GPUs. H100, H200, B200, MI300 and GH200.
RAM. DDR5 RDIMMs and HBM3E
NVME. U.2/U.3 enterprise SSDs
They smaller players are the AI startups. Think Mistral AI, xAI, Cohere, AI21 Labs, Aleph Alpha and Stability AI.
Theyre at risk because they are competing with the big corp players in the game. There are many others, Leonardo.ai, Higgsfield AI... they are compute resouce hungry but the monitisation is fuzzy.
Then there are smaller companies who want to get in on the act, spent a lot of cash and already crashed out due to the high costs and no return in the time required.
This bubble will hit hard because so many are involved. Even companies that should otherwise be safe will be hit as they are starting to use AI as a cost but no sustaiability for what they are doing.
When the second hand tech hits the market you wont want because its packaged in enterprise level tech, like U.2 and GPUs you never heard of. it will take a while for the companies SK Hynix to name only one to get back to their bread n butter sales.