A Portfolio Risk Thought Experiment

Thoughts on the Market by Andy Krieger

July 27, 2026

Prepared as a hypothetical scenario analysis for portfolio risk framing purposes. Not investment advice — see disclaimers at the end. Market-cap and capex figures reflect data available as of late July 2026 and these figures will invariably shift.

1. Executive Summary

The starting premise of this report is a historical pattern, not a prediction about AI's long-term value. Nearly every major technological innovation in modern financial history has followed the same arc: genuine transformation, a burst of market enthusiasm, a speculative bubble, and then a sharp — sometimes severe — stock market sell-off. This happened with railroads in both Britain (1840s) and the US (Panic of 1873). It happened with the internet (the 2000 dot-com crash). Automobiles are a partial case: the sector didn't have its own standalone crash, but automobile stocks were among the most speculative winners of the 1920s boom and were wiped out in the same 1929 crash that took down the broader market (full detail in Section 4). This report treats AI as a candidate for the same pattern -- not because AI lacks genuine long-term value.  None of the previous technologies lacked genuine long-term value either, but their equity holders got wiped out anyway. Railroads, the internet, and (per most economic historians) AI all represent real, durable improvements to how the economy functions; that fact and "the stock market sell-off could still be severe" are not in tension — they're the same story told at different times and in tremendously different contexts.  Let me emphasize again, this is to be seen as a thought experiment about near-term market dynamics, not a comment on whether AI will ultimately be transformative.