Top AI Chiefs Call for Voluntary Slowdown as OpenAI Shelves 2026 IPO

The heads of two of the most powerful artificial intelligence companies in the world publicly called Saturday for a voluntary slowdown in the race to develop more capable AI systems, a striking shift in tone from an industry that has long prized speed above all else — and one that came with a concrete consequence: OpenAI's Sam Altman announced the company will not go public in 2026.

Dario Amodei, CEO of AI developer Anthropic, published an essay urging companies across the industry to take a more deliberate approach to improving their models. "We must slow the pace at which we improve the capabilities of AI models," he wrote, carefully adding that he did not mean halting research outright, but rather ensuring that companies take sufficient time to align and safeguard their systems before pushing capabilities further.

Amodei laid out a three-step plan. In the first step, which he said Anthropic has already committed to unilaterally, third-party evaluators would receive employee-level access to the company to verify its safety practices and report incidents. The second step would have leading AI companies within democratic nations coordinate on shared safety standards. The third would extend that coordination to include both democratic and authoritarian governments.

Altman, whose company is the maker of ChatGPT, responded publicly that he agreed. "I agree with Dario that we need to pace the frontier," he wrote on the X platform. He said the topic had been a primary subject of discussions within OpenAI in recent weeks and pledged that OpenAI would also open itself to third-party evaluators with employee-level access. "We'll have more to share soon," he added. Elon Musk, who runs a competing AI venture, offered a terse endorsement: "Dario is right."

David Sacks, co-chair of the President’s Council of Advisors on Science and Technology, responded on X that Amodei and Altman hold a duopoly on frontier models. He wrote that they should pace development if they wish but should stop seeking regulatory exemptions or coordinated standards that would lock in their market lead.

The public call for restraint followed a week of turbulence in AI circles. A researcher named Jacob Coxon announced he resigned from Anthropic out of concern that AI companies are, in his words, "gambling with our lives," warning that the people building these systems "earnestly believe that it could kill us all by the end of the decade." OpenAI's own chief scientist, Jakub Pachocki, had earlier published a post arguing that no AI company has sufficiently "solved alignment and monitoring" to responsibly continue scaling at maximum speed.

Those warnings have reached Capitol Hill. Growing numbers of lawmakers from both parties have called for new rules governing AI systems, with concern sharpening after researchers warned that rapidly advancing AI could pose an existential risk to humanity.

The safety concerns are now openly influencing business decisions. Altman, in an interview with Fortune magazine, said a 2026 initial public offering for OpenAI would be "ill-advised" given the current moment. "We've got a lot of stuff to do," he said, pointing to safety and alignment work that he said must come first. When pressed on whether the IPO was being pushed to 2027, he said simply: "I would say not 2026."

OpenAI had filed confidentially for a public offering and retained bankers and lawyers, with an eye toward going public in the third or fourth quarter of this year. Separately, Anthropic is moving ahead with its own IPO on a different timeline, with sources telling Reuters the company expects to begin marketing the offering in mid-October at the earliest and complete its listing around the time of the November midterm elections.