> ## Content Index
> Fetch the complete content index at: https://www.dailydebasement.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Opinion: $700 Billion in AI Spending Is Actually Very Affordable If You Divide It by the Number of Humans We Hope to Replace
- URL: https://www.dailydebasement.com/opinion-700-billion-in-ai-spending-is-actually-very-affordable-if-you-divide-it-by-the-number-of-humans-we-hope-to-replace/
- Published: 2026-08-16T01:53:27.000Z
- Updated: 2026-09-01T14:10:49.000Z
- Author: The Daily Debasement
- Tags: Opinions & Commentary, Tech, Business

There has been a lot of hand-wringing lately over the astonishing amount of money Big Tech is spending on artificial intelligence infrastructure.

More than $700 billion this year, by some estimates.

To the untrained eye, that may sound like an alarming amount of capital to spend on warehouses full of GPUs, power contracts, cooling systems, transmission lines, and software designed primarily to summarize emails nobody wanted to receive in the first place.

But this criticism ignores the most important variable in the AI investment thesis: **people are extremely expensive.**

There are roughly 3.6 billion employed people on Earth. Suddenly $700 billion doesn’t sound so crazy, does it?

That’s barely $194 per worker.

For less than the price of a nice pair of noise-canceling headphones, Silicon Valley is attempting to determine whether your entire profession was a temporary technological inconvenience.

Frankly, that sounds like value.

Critics argue that AI revenues remain small relative to infrastructure spending. But this misses the point. Nobody builds $80 billion data centers because customers currently need them. You build them because one day customers may need them, preferably after you have eliminated enough customer-service representatives to make the economics work.

Consider an accountant earning $90,000 annually. Replace that accountant with a server rack costing $12 million and suddenly you have what Wall Street calls “operating leverage.”

Or take a graphic designer. A human designer requires salary, healthcare, vacation days, encouragement, and occasionally asks whether the company has considered “work-life balance.”

A GPU requires electricity.

Lots of electricity, admittedly.

Potentially the output of Belgium.

But it never asks for Friday off.

The important thing is to think long term. Today, AI companies are spending hundreds of billions to generate modest revenue. Tomorrow, after replacing lawyers, programmers, analysts, teachers, artists, accountants, drivers, consultants, and everyone in marketing, the addressable market becomes virtually unlimited.

Unless, of course, those people were also supposed to buy things.

But that sounds like a macroeconomic problem, and fortunately tech companies don’t report those on earnings calls.

---

Inspired by:

[Reuters - "Wall St Week Ahead: With Fed mum on next move, investors look to earnings to keep stocks afloat" (August 14, 2026)](https://www.reuters.com/legal/transactional/wall-st-week-ahead-with-fed-mum-next-move-investors-look-earnings-keep-stocks-2026-08-14/?ref=dailydebasement.com)