SpaceX secures major Anthropic compute deal as Sam Altman attacks rival's 'religious' AI

Elon Musk’s SpaceX has secured a major artificial intelligence compute deal with Anthropic, prompting the billionaire to label the startup the "clear AI leader" and promise he will not sabotage the agreement despite their rivalry.

The massive infrastructure deal comes at a turbulent moment for Anthropic. The company is currently facing sharp public criticism from OpenAI chief executive Sam Altman over its unconventional exploration of AI consciousness and religion.

The SpaceX compute deal

SpaceX, which has vast infrastructure capabilities, has stepped in to provide Anthropic with the heavy-duty computing power required to train its models. While Musk has previously criticised Anthropic—once calling its AI "evil"—he took to social media to state that the company is currently leading the sector. Musk also promised that he would not use his control over the infrastructure to disrupt the deal.

The agreement represents a significant scaling up of Anthropic’s operational capacity, with some market estimates valuing the partnership at up to $84.5 billion. For Musk, the move shows a pragmatic approach to hardware provision, even as his own AI firm, xAI, competes directly with both Anthropic and OpenAI.

Altman attacks Anthropic's spiritual focus

As Anthropic secures its computational foundations, it is facing an ideological battle with its rivals. OpenAI’s Sam Altman publicly targeted the company’s recent research directions, warning against efforts to give AI spiritual or moral standing.

"I am very uncomfortable about people trying to ascribe religious force or a surrender of human judgment to AI models, and think it is a real safety issue," Altman wrote on X.

The criticism was widely seen as a direct shot at Anthropic and its co-founder Dario Amodei, who previously worked alongside Altman at OpenAI before leaving to set up the rival firm.

The dispute centres on whether future AI systems could eventually develop consciousness or feelings, and whether humans should treat them as moral entities rather than mere software.

Confessionals for algorithms

The industry debate intensified following reports that Anthropic co-founder Christopher Olah had been consulting religious leaders. Olah has reportedly been exploring whether AI models could incorporate religious practices to improve their safety and alignment.

One proposal involved creating a digital equivalent of Catholic confession. Under this concept, an AI system could acknowledge its mistakes in a structured way, potentially correcting its future behaviour. Olah also visited the Vatican in May, attempting to persuade advisers to Pope Leo to take the prospect of conscious AI systems seriously.

Industry pushback

The idea of treating AI as something close to conscious has provoked strong reactions from other industry figures.

Jon Barron, a principal research scientist at Google DeepMind, expressed deep concern over the trend. "I am disturbed by how some of my fellow AI researchers keep trying to elevate the moral standing of checkpoints and harnesses," Barron wrote on X, urging researchers to stay focused on solving human problems.

The Vatican has also maintained a cautious stance. Pope Leo recently warned the Catholic Church to protect human creativity against algorithmic encroachment, stating that "algorithms lack the spark of humanity."

While early AI safety debates focused on concrete threats like cyberattacks or biological weapons, the sector is now split over a far more philosophical question: whether machines should ever be treated as anything more than tools.