Sam Altman of OpenAi says the AI market is in a balloon

Openai’s CEO Sam Altman believes that the artificial intelligence market is in a balloon, according to a report By Verge published on Friday.

“When bubbles happen, smart people are exaggerating to the nucleus of truth,” Altman told a small group of reporters last week.

“Are we in a phase where investors are generally excessive about AI? My opinion is yes. AI is the most important thing that has to happen in a very long time? My opinion is also to,” he quoted.

Altman seems to compare this dynamic to the infamous DOT-COM bubble, a collapse on the stock exchange focused on internet-based companies that led to a large-scale enthusiasm for investors in the late 1990s. Between March 2000 and October 2002, NASDAQ lost nearly 80% of their value after many of these companies failed to generate revenue or profits.

His comments add to the increasing concern among experts and analysts that investment in AI is moving too fast. Alibaba co -founder Joe Zai, Ray Dalio of Bridgewater Associates and Apollo Global Management Torsten Slok’s chief economist all raised such warnings.

Last month Slok stated in a report The fact that he believes that today the AI balloon is bigger than the internet balloon, with the top 10 companies in the S&P 500 more than in the 90s.

In an email to CNBC on Monday, Ray Wang, CEO of Silicon Valley -based Constellation Research, told CNBC that it believes that Altman’s comments carry some validity, but the risks are dependent on the company.

“From the point of view of broader investment in AI and semiconductors … I don’t see it as a balloon. The foundations in the supply chain remain strong and the long-term trajectory of AI trend maintains an ongoing investment,” he said.

However, he added that there is an increasing size of speculative companies to pursue capital with more non -delicious foundations and only perceived potentials that could create pockets of overestimation.

Many fears of AI bubble had hit a fever early this year, when the Chinese launch Depepeek launched a competitive model for reasoning. The company claims that a version of its modern large language models has been trained for less than $ 6 million, as part of the billions spent by US leaders in the AI market as Openai, although these allegations have been met with some skepticism.

Earlier this month, Altman told CNBC that Openai’s annual repetitive revenue is about to pass $ 20 billion this year, but nevertheless, it remains unprofitable.

The release of OpenAi’s last GPT-5 AI model earlier this month was also rocky, with some critics complaint that there is a less intuitive feelingS This has led to the fact that the company is restoring access to Legacy GPT-4 customer payment models.

After the model is released, Altman also signals more caution for some of the more of the AI industry’s forecasts.

Speaking to CNBC, he said he believes the term artificial general intelligence or ‘Agi’ loses significance., Asked whether the GPT-5 model moves the world closer to AGI.

AGI refers to the concept of an artificial intelligence form that can accomplish any intellectual task that a person can, to which Openai has been working for years, and what Altman said earlier can be achieved in “reasonably the near future.“

Nevertheless, the belief in Openai by investors remains strong this year. CNBC confirmed Friday That the company is preparing to sell about $ 6 billion in shares as part of a secondary sale that will estimate it at about $ 500 billion.

In March he announced a $ 40 billion in funding Round in estimated $ 300 billion, far the largest amount ever raised by a private technology company.

In Verge’s article on Friday, Openai CEO also discussed the expansion of Openai in the hardware of users, brain computer interfaces and social media.

Altman also said he expects Openai to spend trillions dollars on building its data centers in “Not too distant future” and signals that the company will be interested in buying Chrome if the US government has to impose Google To sell it.

Asked if he would be Openai CEO in a few years, he was quoted, saying, “I want to say that maybe AI is in three years. It’s a long time.”