In 1901, a group of investors gathered in New York to debate whether electricity was overhyped.
1901, coal tycoons, steel magnates, and bankers gather under sputtering gas chandeliers in lower Manhattan to debate whether electricity was overhyped.
These were no fools. Kerosene worked, and so did steam. It was a profitable system, and these were practical people. Why rip it up for expensive wires and unproven dynamos?
Within a generation, the question wasn’t whether electricity was a bubble but why anyone ever doubted it.
Power became abundant, cheap, and ubiquitous. In turn, it completely reorganized the economy around it.
This is the same dynamic investors are facing right now with artificial intelligence.
The risk isn’t that AI fails to live up to the hype, but that it works exactly as advertised… That it strips friction out of research, coding, legal work, customer service, and analysis — not for $10,000 a seat, but for the cost of compute… or $20 a month.
You can already see it in the tape. The indices look calm, but underneath, it’s a split screen: On one side there’s capital pouring into the physical AI supply chain while, on the other side, entire clusters of software and “friction businesses” get marked down in days.
When AI disintermediates one layer of spending, the money migrates, concentrating in the infrastructure that makes the system run — and that infrastructure must be built, expanded, powered, and refreshed repeatedly.
That’s why capital keeps attaching itself to areas like the semiconductor complex — with exchange-traded funds (ETFs) such as VanEck Semiconductor ETF (SMH) and iShares Semiconductor ETF (SOXX) — and to chipmakers such as Taiwan Semiconductor Manufacturing Co. Ltd. (TSM), Broadcom Inc. (AVGO), Micron Technology Inc. (MU), ASML Holding NV (ASML), and Advanced Micro Devices Inc. (AMD)… even as many software stocks struggle to hold gains.
So the real question isn’t whether to “buy the dip” in yesterday’s winners… It’s where the dollars flow when AI gets better every quarter.
In this week’s Being Exponential, we break down the widening divide, explain why some software rebounds may be little more than reflex rallies, and outline how to position on the right side of this migration before the crowd fully recognizes what’s happening.
Take a look and let us know what you think in the comments:
