David Jackson

Interesting short article from Bain. At the demand predictions for use of AI, there is simply not enough capital to keep up.

“Bain’s analysis of sustainable ratios of capex to revenue for cloud service providers suggests that $500 billion of annual capex corresponds to $2 trillion in annual revenue. What could fund this $2 trillion every year? If companies shifted all of their on-premise IT budgets to cloud and also reinvested the savings anticipated from applying AI in sales, marketing, customer support, and R&D (estimated at about 20% of those budgets) into capital spending on new data centers, the amount would still fall $800 billion short of the revenue needed to fund the full investment.“

How Can We Meet AI’s Insatiable Demand for Compute Power? | Bain & Company

So how will this play out?

I don’t think AI gently plateaus. I think it will look more like a denial-of-service attach on a website, or trying to get Taylor Swift tickets from Ticketek. The system will groan under the weight of demand, degrade, and then stop.

In my own life I am using AI about 3 times more than I did 3 months ago. That is 3x more demand. Is that 3x the compute power? If I multiply out my own experience by billions of people, I smell a demand and supply gap of GALACTIC proportions.

$500 billion in annual spending on new data centres is an extraordinary number. For the AI business model to work, it doesn’t simply have to provide value to people. It needs to provide value IN EXCESS of what it costs to provide that value.

Unless there is a major break through in the cost of computing power, or the existing computing power is focused very narrowly on those that can pay, I think this whole game has to stop, or at least slow down considerably while we work out how to power it.