Commentary: Can AI’s leaders really put aside rivalry for the common good?

The deep schism between OpenAI and Anthropic has led to a race-like dynamic towards ever more powerful models, says Madhumita Murgia for the Financial Times.

New: You can now listen to articles.

This audio is generated by an AI tool.

15 Sep 2026 05:58AM

LONDON: The world of mathematics was rocked by a claimed breakthrough last week - a solution to a longstanding problem over a 200-year-old complex set of equations known as Navier-Stokes.

One of the first problems of its magnitude to be solved with the help of AI, the breakthrough should have been celebrated as an intellectual triumph. But instead, the news has been marred by a roiling controversy over who deserves credit: the New York mathematician and his collaborator from AI giant Anthropic who had been close to the solution, or its rival OpenAI which spent millions of dollars and used about 10,000 AI agents to complete the task.

The quarrel between Anthropic and OpenAI is not their first; it is the latest in a long series of disputes in the relationship between the two top contenders in the race to create superintelligent AI systems.

That increasingly matters to the world as AI develops faster than even the companies had expected. Anthropic’s co-founder Dario Amodei has just called for the pace of AI’s development to be slowed after a series of alarming safety warnings from current and former employers.

The call was supported by OpenAI’s Sam Altman and Elon Musk in a rare and unusual display of unity. There is not much detail on the practicalities of how their aims would be achieved, with both insisting that they will continue to train and build new generations of models. And there is an open question - can these companies and their founders put aside their rivalry for the common good?

AN OPEN RIVALRY

The schism between OpenAI and Anthropic is deeply personal. Anthropic was founded in 2021 by a group of OpenAI employees who broke away from the ChatGPT maker, spurred by the belief that AI systems needed a more responsible steward than OpenAI. Ever since, Anthropic has fashioned itself as a custodian of safe AI, with a dual focus on social purpose and profit.

Over the past few years, the two have competed openly for talent, public favour, technological supremacy and commercial dominance via their products ChatGPT and Claude. Now, they are racing to list on the public markets, eyeing similarly dizzying valuations beyond US$1 trillion. The companies filed their documentation to the Securities and Exchange Commission one week apart.

It is common to see employees of the two companies carping at each other on social media site X, and their leaders archly criticising one another. In February, Altman and Amodei refused to hold hands at a photo session with Indian Prime Minister Narendra Modi as other leaders did during India’s AI Impact Summit. They instead raised their fists.

Such incidents might be dismissed as petty frictions not unknown elsewhere in the corporate world. But rather than focusing on co-operating to contain the evident risks arising from AI, the companies have been racing to release new models more powerful than the last.

In February, Anthropic opposed the Pentagon’s use of AI for certain military applications such as autonomous lethal weapons and mass surveillance. This led to the Trump administration publicly cutting ties with Anthropic and a legal firestorm. Within hours, OpenAI had taken advantage of the situation and signed its own deal with the Pentagon, claiming it had negotiated its safeguards independently. Google followed suit.

In another sign of the pressure-cooker competitive environment, Anthropic publicly pulled back on commitments in its responsible scaling policy this year, a pledge to halt or withhold AI models if it could not guarantee proper safety measures in advance. Anthropic cited competitor pressures, and a lack of faith in unilateral pledges, as a reason.

So why have the companies continued to push on developing AI and revenues at speed until now in the face of what they agree are potentially catastrophic risks to human lives? Part of the explanation is ideological. AI corporations frame the advance of the technology as an existential and moral issue. Each believes they should be entrusted with developing powerful AI because their rival is flawed. If any one of them pulls the handbrake now, they believe the outcome could be worse than ploughing on.

One former Anthropic researcher I spoke to put the dynamic like this: “Hubris. Fear that the next guy will do it even worse. Being caught up in your own narrative that you’re the good guys.”

But as the AI leaders concede, if there is a time to put aside personal rivalries and find a path forward, it is now. There is so much riding on it.