My colleague Mike Boland and I have remarked many times in our discussions on Localogy’s This Week in Local podcast that one of the notable features of AI is how incredibly expensive it is. AI is a massive data hog. And that makes it an energy hog as well.
Open AI, founded in 2015, has raised more than $11 billion to date. ChatCPT has not been backed by investors with bats flying out of their wallets.
A recent report in The Information, cited by multiple news outlets in recent days, suggests that $11 billion isn’t going to cut it. Open AI may lose $5 billion this year. And it may go belly up within a year if it doesn’t get a massive infusion of cash.
An estimate from Appfigures suggests OpenAI generates between $3.5 billion and $4.5 billion annually from subscriptions and access fees.
But the burn rate at OpenAI, which has an $80 billion valuation, is staggering.
One Expensive Beast
OpenAI is moving fast into the future. And it turns out that it is really expensive.
News reports last week said the company is running near capacity, with 290,000 of 350,000 servers dedicated to ChatGPT.
It’s no wonder that AI is overtaking crypto as every climate scold’s favorite target.
Apparently it costs $700,000 a day to keep ChatGPT chugging along. And OpenAI has reportedly invested close to $7 billion on training ChatGPT and $1.5 billion on staffing. AI talent doesn’t come cheap.
So what happens next? Will ChatGPT go belly up? It seems hard to imagine that would happen. Stranger things of course, but maybe OpenAI is too good to fail?
This reporting does seem to bolster a common critique of the company. And that is that, so far at least, OpenAI lacks a business model to match its ground-breaking core product.
The OpenAI team is apparently uncowed by reports of its financial turmoil and is remaining focused on achieving AGI, which may come as soon as 2027.
