Where this was said
The Business Case: Promising But Flaky
At 8:50 · chapter starts 7:30
Spencer Jakab is not dismissive of the technology itself: ChatGPT and Claude have genuinely amazed users, advancing so rapidly that he compares AI's progress to a child jumping from kindergarten to high school [1] — Spencer Jakab "The products are impressive — ChatGPT and Claude have genuinely amazed users. But the underlying economics are fragile: these companies spe…" 07:00 . But impressive products do not automatically make for profitable businesses. The fundamental problem is that AI infrastructure costs far more to build and run than these companies currently charge customers — and neither OpenAI nor Anthropic is profitable. Jakab uses the word 'flaky' deliberately, and it lands hard. The scale of the investment commitment makes the risk even more acute: when all promised AI infrastructure spend is tallied, Jakab estimates it exceeds the internet, railroads, and even the Apollo moon programme, approaching the economic scale of World War II [2] — Spencer Jakab "When all the AI infrastructure commitments are counted up, the total surpasses every major technology buildout in history — the internet, r…" 08:00 . For that level of spending to pay off, revenue must grow at a pace no company in history has ever managed.
When all the AI infrastructure commitments are counted up, the total surpasses every major technology buildout in history — the internet, railroads, even the Apollo program. Spencer Jakab's jaw-dropping benchmark: we're approaching the economic scale of World War II.
Total committed AI investment is projected to exceed the cost of the internet, railroads, and the Apollo moon program — approaching the scale of World War II spending.
Anthropic and OpenAI are both expected to go public at valuations around $1 trillion or more later in 2026.
