© Andrew Neel/Unsplash
$10 billion, then $6.6 billion… OpenAI, the parent company of ChatGPT, is raising funds at lightning speed and its valuation now exceeds $150 billion. But this is never enough and other cash injections will soon be necessary. The race for generative AI is currently costing huge sums without bringing in any money. And it is not about to improve. Explanations.
Very expensive AI
In a fascinating article, our colleagues at the New York Times explored this issue. OpenAI estimates that by 2029, it will cost $37.5 billion annually to train and run its various AI products, compared to $5.4 billion today.
The reasons for these exorbitant costs are not far to look for. Indeed, these technologies do not fall from the sky. To function, language models must have computing power from gigantic data centers.
These spaces house crazy quantities of computers equipped with thousands of state-of-the-art electronic chips, each of which can cost more than $30,000. And their price is all the higher as demand rages on this market, while the tech giants seek to gain a head start on the competition.
200% Deposit Bonus up to €3,000 180% First Deposit Bonus up to $20,000Added to this problem is the high energy cost of these data centers, which further increases the bill in addition to posing concrete sustainability issues. The American daily rightly points out that these chips spend months performing calculations that allow ChatGPT to identify patterns in all this data. As a result, the price of each training cycle can reach hundreds of millions of dollars!
The worst is yet to come
It goes without saying that free users of OpenAI's AIs cost it quite a lot, because they are not yet monetized. That said, the company says it is considering adding advertising, although it has not yet made a decision. However, while the company charges its customers 20 euros per month for the premium subscription, this still does not manage to make the initial investment profitable.
Things are not likely to improve as OpenAI is eyeing models such as o1. Launched at the end of the year, the latter “thinks” before responding. It therefore explores the very many possibilities before generating content that is supposedly more reliable than what it would produce without this “thinking” time. This functionality requires even more computing power. Enough to further inflate the company's expenses.
📍 To not miss any Presse-citron news, follow us on Google News and WhatsApp.
[ ]