Found via a Reddit post about a WSJ article quoting a Sequioa presentation
In a presentation earlier this month, the venture-capital firm Sequoia estimated that the AI industry spent $50 billion on the Nvidia chips used to train advanced AI models last year, but brought in only $3 billion in revenue.
This 17x number is just for chips – Nvidia chips alone, I think – so the actual cost-to-revenue multiplier is much higher in reality.
So the hardware it’s installed in and the actual CPUs are extra. Research is extra. The army of freelancers used for RLHF training are extra. Electricity cost is extra.
And chips depreciate in value pretty rapidly. Especially since every chip vendor on the planet has more specialised ML chips in the pipeline that are more effective at the task. This investment will be worthless pretty quickly.
The numbers are very far from lining up.
Remember those news items that said these services were running at a loss?
Well, those calculations were based on pure compute costs, which DON’T include capital expenses like chips or hardware and it didn’t include labour costs which are, ironically, quite high for generative models.
All of which means the economics for “AI” is much much worse than most people think.
Oh, and remember that the software industry is structured entirely on incredibly high margins (software being a non-rival, non-exclusive good and all). Even if the industry did manage to reach break even on “AI”, the industry would still collapse because it’s too dysfunctional to function on regular margins.