Anytime you hear someone claim that something would bring about the benefit of something being "too cheap to meter", stop believing them right there. These companies are not charities, they are not there to give you something so insignificant in value that they cannot be bothered to measure its worth. That's their entire reason for being: to find new ways to measure value that's otherwise unmeasured and charge for access to that value.
Right obviously. So then what happens with 10x the data centers? How do you saturate that in any reasonable way? The options are 10x usage, 1/10 cost.
Probably 10x usage.
Think about it this way: when OpenAI was founded, it was a public benefit corporation. It had to take into account the "public good" - which is nebulous, but whatever - and capped returns for shareholders at 100x initial investment. It has since been converted into a for-profit corporation and I recall reading that the earnings cap has been done away with, but that should give you at least some idea of the stakes.
According to these guys [0], there has been a total of $180 billion invested in OpenAI since its founding. Assuming there's not been a change in what the shares were sold for to each round - which is a fairly large assumption, I know - for them to reach the 100x threshold that was apparently too low to be followed, the company would have to reach a valuation of $18 trillion dollars.
For what it's worth, the year-on-year GDP of the world's largest economy, the United States, was $32.56 trillion in Q2 2026. [1]
You don't reach a little over half of the economic output of the US in valuation by cutting costs.
[0] https://tracxn.com/d/companies/openai/__kElhSG7uVGeFk1i71Co9...
[1] https://fred.stlouisfed.org/series/GDP/