OpenAI, the San Francisco start-up that launched generative AI with ChatGPT at the end of 2022, is now worth $157 billion, after raising $6.6 billion in funding.
Still unknown to the general public two years ago, it thus strengthens its status as an essential champion of technology, despite the recent departure of several managers, scandals, fierce competition and, above all, no immediate prospect of profitability.
This is not enough to dampen the enthusiasm of analysts who see AI as a real “revolution”.
“A fourth industrial revolution is arriving in semiconductors, software, infrastructure, internet and smartphones over the next 12 to 18 months,” assure Dan Ives and his colleagues at Wedbush.
“ChatGPT was to generative AI what the iPhone was to smartphones,” they continue. OpenAI has enabled the success of AI and its adoption by Nvidia, Microsoft, Google and the entire tech world. » They estimate that all these companies will spend a trillion dollars on AI over the next three years.
Because this technology is very expensive: to produce texts, images and other content on a simple query in everyday language, AI models like GPT-4 need to be “fed” with massive quantities of data.
Customers not yet ready
Training and operating models therefore requires new IT infrastructure, a lot of energy, highly qualified engineers, etc.
Microsoft has pumped $13 billion into OpenAI in total, and announced more than $15 billion in overseas AI investments this year, from Germany to Indonesia.
Google released more than 13 billion in “properties and equipment” in the second quarter, almost double over a year, a budget which reflects its needs for new data centers and ultra-sophisticated chips.
“But this spending has barely yielded results so far, other than efficiency gains for developers,” noted the Goldman Sachs bank last June in a report.
“We are currently in a bubble where all the suppliers are shouting that we must deploy generative AI in all directions,” observes Rob Enderle, while customers “are not yet ready”. According to this independent analyst, “more than 80% of first deployments are failures”.
But the big tech players are not yet really looking to make profits, explains Grace Harmon, analyst at Emarketer. Above all, they fear “underinvesting in AI and losing out (…) even if the returns on investment are not guaranteed,” she emphasizes.
Bubble or no bubble, most observers nevertheless perceive the medium or long term potential of the new technology.
An economic model
“Are there very high valuations? Yes. But is there value behind it? Yes,” observes Michael Mansard, director of the Subscribed Institute think tank. According to this sector expert, the challenge for companies is to find an economic model that will ensure their survival after a potential explosion of the bubble. “For a software manufacturer, having 200 clients instead of 100 does not cost twice as much to run the infrastructure. But for a generative AI provider, it is possible. »
Despite its global success and paid subscriptions for individuals and businesses, OpenAI is expected to lose $5 billion this year, for $3.7 billion in revenue, according to the New York Times.
Carolina Milanesi, independent analyst, has no doubt that generative AI will soon become part of everyday life. But to stay in the race and impose their AI assistants and other tools, companies are forced to invest. “The business model is difficult to determine, because the more you use the model, the more you spend,” she notes.
“It’s a bit of a Darwinian situation: only the strongest will survive. One or two brands will control the market. »
Julie JAMMOT with
Glenn CHAPMAN/AFP
OpenAI, the San Francisco start-up that launched generative AI with ChatGPT at the end of 2022, is now worth $157 billion, after raising $6.6 billion. Still unknown to the general public two years ago, it thus strengthens its status as an essential champion of technology, despite the recent departure of several managers, scandals, fierce competition…
- -