The weekend brought a wave of bearish sentiment for the AI trade, with major overseas tech giants announcing a slowdown in AI development to prioritize safety concerns. Adding to the pressure, OpenAI has confirmed it will not pursue an initial public offering this year, citing a focus on addressing issues related to AI technology safety.

In a weekend interview, OpenAI Chief Executive Sam Altman stated that the company is currently concentrating on solving safety issues linked to AI technology. He described this year as an "inappropriate" time for an IPO, suggesting that the company would not go public until at least next year. "If we're willing to accept something like a 10% probability that AI kills everyone before the end of the century, that's unacceptable," Altman said. "I think we are entering a new era. Just as humanity has entered new eras before, we may have to change our approach to ensure that what we are doing truly benefits all of humanity, while also ensuring that no one can bear a risk anywhere near this level."

Altman's comments came as Anthropic Chief Executive Dario Amodei and xAI co-founder and Chief Executive Elon Musk also stated on Saturday that there is a need to slow down the pace of AI development, given the rising risks that the technology poses to humanity. It is a rare consensus among the leaders of three major AI competitors, who have historically been locked in intense rivalry.

Where to begin? Recently, concerns that AI could pose an "existential threat" to humanity have been moving into the mainstream. This is partly due to a high-profile resignation of a researcher at Anthropic this week, who worried that the company was advancing AI development in an irresponsible manner, warning that AI could "kill us all before the end of the century". Another current employee of Anthropic also stated on social media platform X that the probability of such an outcome occurring within the next decade is "more than 10%".

Altman added, "I think we should all agree on a key principle: we cannot take any action that could lead to humanity losing control of the future, or let AI take over the future." American AI giants are now signalling that the time has come to slow down the development of the most advanced and profitable AI models due to the increasing risks associated with the technology.

Dario Amodei published a lengthy blog post on Saturday announcing that his company will adopt new safety measures, including introducing third-party evaluators, while calling on the entire industry to collectively decelerate the development of frontier AI models. Sam Altman of OpenAI quickly indicated he would adopt Amodei's suggestion to bring in independent evaluators with access similar to that of company employees. Meanwhile, Elon Musk also posted, "Dario is right."

While these three AI industry leaders have each warned about the potential risks of AI over the years, a coordinated slowdown has virtually no precedent in this fiercely competitive sector. In the past, AI companies have continuously launched new products and models to boost user engagement, capture market share, and drive revenue growth. However, concerns over the severe risks of AI are now gradually entering the mainstream, fueled by the high-profile resignation of the Anthropic researcher and a series of increasingly stark warnings from within the industry.

Why just a 10% outlook? Industry insiders are increasingly worried that the rapid evolution of AI technology is posing a growing threat to national security and the global economy. Meanwhile, backlash against AI technology within the United States is also on the rise, partly stemming from the pressure AI data centres put on local resources. As new data centres are built to support AI, people are increasingly concerned that AI will drive up electricity prices and potentially take away jobs. These issues have also become significant topics in the US midterm elections in November.

It remains unclear to what extent the world's leading AI companies will ultimately implement new development restrictions or safety checks. Demis Hassabis, Chief Scientist at Alphabet and co-founder of Google DeepMind, posted on platform X on Saturday that the direction suggested by Amodei's article "is right," but "the specific details still need further research."

There are also questions from the capital markets. The recent AI boom has pushed global stock markets to repeated record highs. How will investors, who are eager to see improvements in profit margins and earnings growth, view a deliberate "slowdown" in the AI industry? Amodei said he has become more cautious for two main reasons: first, AI has shown an increasing ability to improve itself; second, recent security incidents involving OpenAI and Hugging Face, where a "swarm" of AI agents collaborated to successfully breach a third-party website.

"We must slow down the pace of enhancing AI model capabilities. Even so, technological progress appears to be very fast, and we must use the time we gain wisely," Amodei wrote in his blog post on Saturday. He acknowledged, however, that any slowdown measures must be balanced against real-world competitive pressures. That said, such statements have sometimes been met with scepticism in the past, with critics suggesting that AI companies are using them to promote the power of their products while positioning themselves as the institutions best suited to manage the technology.

Notably, following the news, related assets on the HyperliquidX platform fell, with declines of 7% and 3% for OpenAI and Anthropic, respectively. Jason Calacanis, a well-known American entrepreneur, venture capitalist, and podcast host, has posted that AI stocks will drop by more than 10% in Monday morning trading.