U.S. stocks are set to skyrocket driven by AI narratives; the last time such a boom happened was in 2000. Will the AI bubble meet the same fate and cause a crash in the U.S. stock market?

The real cause of the 2000 dot-com bubble wasn't greed—
It was that 90% of fiber optics had never been lit up~
From 1995 to 2000, internet traffic increased 500 times. The market saw this number and collectively lost its mind: downstream, midstream, upstream equipment vendors, each layer was expanding wildly with the logic that "traffic is never enough," raising funds recklessly, burning cash like crazy~
But there was a fatal cognitive blind spot:
The market measured infrastructure capacity by "lit fiber optics"—thinking supply was tight, thinking broadband would always be insufficient.
What’s the reality? Lit fiber optics only account for 5% to 11% of the total laid fiber—90% of the fiber is lying underground, unpowered, unused, ready to be activated at any time~
Supply isn’t lacking; it’s an enormous surplus hidden away~
Once capital starts to retreat, silent supply suddenly floods out—equipment orders plummet, ISP financing dries up, ICP burns through cash, dominoes fall from upstream to downstream, one after another~
Now replace fiber optics with GPUs, ISP with cloud providers, ICP with AI applications—
NVIDIA’s capacity is climbing, Microsoft, Google, Amazon data centers are frantically stacking computing power, but the monetization logic for AI applications hasn’t yet taken off~
Isn’t the story on the supply side a bit familiar?
This time, the "dark fiber" might be called idle computing power~
The bubble structure has never changed; it’s just dressed in a new outfit for its comeback~
History doesn’t repeat exactly, but it rhymes, and it rhymes perfectly—word for word~
Do you think this wave of "big infrastructure" computing power bubble will burst with a whistle from the People’s Bank of China in June, or will it last until a crazier moment in 2027?
#AI泡沫 # Internet Bubble #供给过剩 # U.S. Stock Risks