On the Madness of Spending Fortunes Before the Meter Runs

I have received, by means I cannot explain and do not trust, a peculiar dispatch claiming to describe the state of electrical — no, they call it 'computational' — enterprise in the year 2026. I shall treat it as I treat every rumor that crosses my desk at Menlo Park: with suspicion first, and calculation second. The report speaks of companies — Meta, Microsoft, Alphabet, Amazon — names unknown to me, yet described as belonging to a 'trillion-dollar club.' A trillion! The entire national wealth of these United States would blush at such a figure. These concerns are said to be pouring historic sums into something called 'AI infrastructure,' which I gather is a species of thinking machine requiring enormous engines of computation. They build and build, yet the returns, the dispatch warns, remain deferred. Wall Street grows impatient. Now, I know something about spending money before the meter runs. When I built the Pearl Street Station, when I laid mains under the streets of lower Manhattan, when I tested six thousand filament materials before finding one that would glow for a thousand hours — I heard the same grumbling from bankers. 'Where is the revenue, Edison? When does the lamp pay for the dynamo?' My answer was always the same: the system pays, not the piece. You cannot sell light without the wire, the generator, the meter, and the customer all working in concert. But you must build the whole thing before the first bill is collected. So I understand these men of 2026, if they are real. They are laying mains for a new kind of current. The question I would put to them is the question I put to myself every morning: How many hours of test have you run? Have you measured the loss in every joint and junction? Because spending is not invention. Capital is not cleverness. I have seen men pour fortunes into schemes that wanted not more money but more thought. Theory without application is philosophy. Spending without measurement is vanity. If these companies are testing, iterating, failing cheaply and adjusting — then Wall Street should hold its tongue. But if they are merely building cathedrals to an idea they have not yet proved at the bench, then they deserve every dollar they lose. I measure progress in patent filings and hours of filament glow. Let them show me their equivalent, and I will judge whether this 'super-cycle' is revolution or ruin.