Key insights
- The author posits that a market crash could benefit hyperscalers with large cash reserves by cheapening or bankrupting assets needed for AI integration. The timing is crucial, requiring sufficient infrastructure build-out to identify clear winners. Premature crashes could lead to stranded assets. The current infrastructure build-out lags behind the AI narrative, suggesting a crash would be more beneficial in a few years.
It’s beneficial in ways for the hyperscalers or those with significant cash reserves to have a market crash as it cheapens /bankrupts what’s needed for deeper AI societal integration long term. A crash timing becomes useful only though once there is sufficient build out for clear take all winners. Left too long the result is “dark fibre” especially in data centres that are not grid connected and with GPUs that age out.
Crashes following railroads, radio, housing, internet all went far in the infrastructure build out but here the narrative enthusiasm probably would require being sustained for a number of more years to enable for the underlying infrastructure demands to catch up to the vision.
Given where we are at in the infrastructure versus narrative build out what do you imagine would be a helpful crash time for ultimate benefit to a few and later then leading to integration?