The headline tells you many mouths need feeding. Not enough food exists. The logical next step is cannibalism.
AI hyperscalers built data centers the size of small nations. They promised profits would arrive any quarter now. Just keep buying the chips. Just keep stacking the GPUs. Revenue will materialize like manna from heaven.
It didn't.
So now we have Google and Microsoft and Amazon and Meta all eating from the same trough. Each spent tens of billions on infrastructure. Each needs customers who will pay enough to justify the spending. Those customers don't exist yet. They might never exist.
The bear case writes itself. Too much capacity. Not enough demand. Basic economics. A child could see it.
Retail traders cannot see it. They bought the dip seventeen times. They averaged down until their cost basis looked like a phone number. They read the headline about hyperscalers being the epicenter of a bear case and thought "epicenter sounds important so I should buy more."
The word "hyperscaler" does a lot of work here. It sounds futuristic. It sounds like disruption. It distracts from the actual problem which is that these companies spent a fortune building restaurants in a town where nobody is hungry.
Many mouths to feed. Not enough to eat. The metaphor falls apart when you remember the mouths belong to trillion-dollar companies who could simply stop building mouths. But they won't. Because stopping means admitting they built too many mouths.
So they'll keep spending. Keep building. Keep insisting the food is coming. And retail will keep buying because a man on YouTube told them hyperscalers go up forever.
Turns out the bear case was just math wearing a scary costume.
Photo by Thomas Bonometti on Unsplash
Leave a Comment