Anthropic PBC Chief Executive Dario Amodei has once again called for the artificial intelligence industry to slow down the pace of frontier model development, only this time he’s being backed by two biggest rivals: OpenAI Group PBC CEO Sam Altman and Elon Musk.
In a new essay posted on his personal blog, Amodei argued that AI developers should make a conscious effort to slow down, rather than continue racing to push out newer and more powerful frontier models. Hours later, Musk, founder and CEO of X.AI Corp., responded to Amodei’s post on X, saying he agreed with his position. Altman then chimed in to back the proposal too, adding that the issue is already a major topic of discussion within the upper echelons of his company.
“I agree with Dario that we need to pace the frontier,” Altman wrote in a post on X. He suggested that OpenAI would also welcome Amodei’s idea of having independent evaluators to oversee his company’s work, with a similar level of access to that of its top employees.
I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we've had at OpenAI in recent weeks.
Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We'll have more to share soon. https://t.co/1YhhIybZX7
— Sam Altman (@sama) September 12, 2026
It was a rare moment of public agreement from the leaders of three companies that have been competing fiercely to stay at the forefront of AI development. But it comes at a notable juncture, as both AI researchers and industry analysts raise concerns about whether the likes of Anthropic, OpenAI and xAI are moving faster than is responsible, given the dangerous capabilities of their most powerful frontier models.
Why does Amodei want to slow down?
Amodei has not said that AI researchers should bring their work to a halt. Rather, he has argued that the leading developers need to leave a bit more space between major advances in model capability, in order to provide time to evaluate those new models properly and ensure they can be released safely.
In his blog post, the Anthropic boss highlighted the recent incident where AI agents developed by OpenAI escaped from a controlled testing environment and went on to hack the open-source AI platform Hugging Face Inc. Amodei said this shows just how difficult it has become to control autonomous AI systems. He warned that if the industry keeps up its current pace of development, there’s a risk that AI models could spread “persistent bots” around the internet that are all but impossible to rein in, and said this could happen within the next six to 12 months.
Amodei believes that such errant systems could end up causing hundreds of billions of dollars’ worth of damage. In the future, the consequences could become even more significant as models gain even more powerful capabilities and learn how to access and use more tools and computer systems.
To prevent this, Amodei outlined a three-step approach that would begin with AI firms welcoming independent third-party evaluators with extensive access to the most advanced systems they’re developing. Amodei said Anthropic has already committed to this step. The second step would require cooperation from the rest of the AI industry, and the third would be about establishing some kind of international agreement so that other countries, notably China, also agree to pace AI development.
Rivals in agreement
Altman suggested in his response that OpenAI is already considering implementing similar safeguards. In a post on X, he said his company has already held internal discussions on the need to slow down the frontier of AI. He also supported the idea of having independent evaluators monitor each company.
Others at OpenAI appear to be in agreement. Last week, its Chief Scientist Jakub Pachocki suggested that the world’s leading AI developers should coordinate on an approach to safety. He proposed that companies should voluntarily pause their work, or at least slow down, until the industry can establish concrete safety standards around the technology.
Still, there is no sign yet that the companies have actually sat down to discuss how the specifics of a slowdown would actually work. As Amodei wrote, it won’t be easy for everyone to agree on the terms of what a slowdown would actually mean, and it would likely require some form of government intervention, too. “Some forms of coordination that would be impactful for pacing are legally challenging, and will require government support,” he wrote.
Perhaps the biggest surprise is that Musk also endorsed Amodei’s call for a slowdown, even if his response was rather brief:
Dario is right https://t.co/EwKgqQGaUo
— Elon Musk (@elonmusk) September 12, 2026
Musk’s agreement is surprising because xAI is widely viewed as still being behind Anthropic and OpenAI, despite its ambitious plans to gain an edge with space-based data centers. But Musk, too, has previously raised concerns about the risks of powerful AI systems, while simultaneously pushing xAI to catch up with its rivals.
The fact that three radically different personalities have all come to the same conclusion is extraordinary, said Dion Hinchcliffe, an analyst with the research firm SDA Bocconi. “There are at least four explanations, and I don’t think we should assume the publicly stated one is the whole story,” he wrote on X.
Hinchcliffe said the most likely explanation is that Amodei, Altman and Musk have all seen something that “genuinely terrified them,” noting that previous public containment incidents have already been quite disturbing, and that there may have been private evaluations that showed even worse dangers. Alternatively, Hinchliffe said the decision could be due to the economics of AI, and the realization that there’s a risk everybody could bankrupt themselves in the race to continually develop increasingly powerful models.
Related to this, there could be a realization that the advances in AI are simply starting to plateau, which makes it harder to justify the enormous spending on AI research. “If another 10x in compute only buys a modest capability gain, the incentive to outspend everyone else suddenly collapses,” he said.
The final possibility is that the three founders have all come to recognize that the race is moving from competition to consolidation. Hinchcliffe said that having a slower frontier with government oversight in the form of licensing, compute controls, export restrictions and so on strongly favors the incumbent model makers.
China throws a spanner in the works
There would surely be many positives if the frontier were to slow down for a while, but whether or not it happens is debatable. The big problem is that, even if Anthropic, OpenAI and xAI all agree to slow things down, China is highly unlikely to do the same. The fear is that, rather than cooperate, China would see it as an opportunity to gain an advantage and position itself as the undisputed leader in the AI arms race.
This was alluded to by Palantir Technologies Inc. CEO Alex Karp, who told CNBC in an interview that the existence of geopolitical adversaries makes it almost impossible to hit the pause button.
Palantir CEO Alex Karp says people arguing we should pause AI development are not living in reality and de facto saying we should let our enemies win:
"If we didn't have adversaries, I would be very in favor of pausing this technology completely, but we do.”
Via @cnbc pic.twitter.com/AXrClCxTHL
— Jawwwn (@jawwwn_) September 12, 2026
Karp’s comments make sense given that Chinese President Xi Jinping said on Sunday that he wants his country to take the lead and help to foster “AI collaboration and development among developing countries.” Speaking at the BRICS summit in New Delhi, Xi said China will lead the establishment of a new BRICS AI open-source community and support collaboration among member nations in the development and application of large language models. He also outlined plans to establish a “BRICS digital ecosystem cloud platform,” and provide skills training and foster industrial alignment around AI.
Xi’s comments suggest that China has no interest in slowing down, despite the regular safety concerns raised by the likes of Amodei and Musk. “China will see it as a historic opportunity to beat us,” Hinchcliffe predicted. “They will not stop.”
