In a pointed critique during a visit to India, OpenAI CEO Sam Altman dismissed Elon Musk's ambitious proposal to build data centers in space as 'ridiculous.' Altman made the remarks on Friday while in New Delhi, highlighting ongoing tensions between the two tech titans who once co-founded OpenAI but have since parted ways amid a bitter feud.
The comment came as Altman engaged with local media and tech enthusiasts in the Indian capital, where he discussed the future of artificial intelligence and its global infrastructure needs. According to reports from Business Insider, Altman was responding to Musk's recent suggestions that placing data centers in orbit could solve earthly limitations like energy consumption and land scarcity for AI training. Musk, the CEO of SpaceX and Tesla, has floated the idea as part of his vision for leveraging space technology to power next-generation computing, with a potential timeline as early as 2026.
Altman's dismissal underscores the divergent paths the two leaders have taken in the AI race. Musk, who left OpenAI in 2018 citing disagreements over its direction, now heads xAI, a rival startup aimed at understanding the universe through advanced AI. In contrast, Altman has steered OpenAI toward rapid commercialization of tools like ChatGPT, securing billions in funding from Microsoft and others to build massive data centers on the ground.
'Ridiculous,' Altman said bluntly when asked about Musk's space-based data center concept, according to attendees at the event and subsequent media coverage. He emphasized the practical challenges of launching and maintaining such infrastructure in orbit, including the immense costs and technical hurdles involved in zero-gravity computing and reliable power sources from solar panels in space.
Musk has been vocal about the idea in recent months, tweeting in late 2023 that space data centers could be feasible within a few years thanks to reusable rockets like SpaceX's Starship. He argued that orbital facilities would bypass terrestrial constraints, such as the high energy demands of AI models that require gigawatts of power—equivalent to the output of several nuclear plants. According to Musk's posts on X (formerly Twitter), initial prototypes could be deployed by 2026, potentially revolutionizing how companies like his own xAI handle exascale computing.
However, experts in aerospace and computing have mixed views on the proposal's viability. Dr. Emily Chen, a professor of aerospace engineering at MIT, told The Appleton Times that while SpaceX's launch capabilities have advanced dramatically, the reliability of data transmission from space remains a significant barrier. 'Latency issues alone could make it impractical for real-time AI applications,' Chen said in an interview last week.
Altman's remarks in New Delhi were part of a broader tour to expand OpenAI's footprint in Asia, where demand for AI tools is surging. India, with its 1.4 billion population and booming tech sector, represents a key market. During his visit, Altman met with government officials and startup founders, discussing partnerships for AI ethics and infrastructure. He highlighted OpenAI's plans to invest in local data centers, potentially creating thousands of jobs and addressing India's digital divide.
The exchange between Altman and Musk is the latest chapter in their public sparring. Their relationship soured after Musk sued OpenAI in March 2024, alleging the company had strayed from its nonprofit roots by prioritizing profits. The lawsuit, filed in California Superior Court, claims breach of contract and fiduciary duty, though OpenAI has countersued, accusing Musk of anticompetitive behavior. A judge is expected to rule on motions to dismiss by early 2025.
Despite the personal animosity, both men share a vision for AI's transformative potential. Musk has warned of existential risks from unchecked AI development, advocating for regulatory oversight, while Altman has called for balanced innovation that benefits humanity. In a 2023 interview with The New York Times, Altman said, 'AI will be the most powerful technology humanity has ever created, and we need to get it right.'
Space-based data centers, if realized, could address some of AI's most pressing challenges. Current ground-based facilities, like those operated by Google and Amazon, consume vast amounts of electricity—OpenAI alone reportedly uses enough power to light up 100,000 homes during training runs for models like GPT-4. Proponents argue that space offers unlimited solar energy and cooler temperatures for efficient cooling, potentially reducing carbon footprints.
Critics, including Altman, point to the enormous upfront costs. Launching even a small data center module via Starship could cost tens of millions, not accounting for ongoing maintenance via robotic repairs or astronaut interventions. A report from the RAND Corporation in 2024 estimated that full-scale orbital computing infrastructure might not be economically viable until the 2030s, barring breakthroughs in satellite technology.
In New Delhi, Altman's comments drew chuckles from the audience, reflecting the blend of admiration and skepticism toward Musk's bold ideas. One local tech entrepreneur, Raj Patel, who attended the event, said, 'Musk dreams big, but Altman keeps it real. We need both in this industry.' Patel's startup focuses on AI for agriculture, and he expressed interest in collaborating with OpenAI on ground-based solutions tailored to India's rural needs.
The feud also highlights broader industry tensions over AI infrastructure. With global data center demand projected to double by 2027 according to the International Energy Agency, companies are racing to secure power and land. Microsoft, OpenAI's key backer, announced in 2024 plans for three new U.S. data centers, each costing over $1 billion, while Musk's xAI raised $6 billion in funding partly to build its own supercomputers in Memphis, Tennessee.
Looking ahead, the viability of space data centers may hinge on upcoming SpaceX milestones. The company successfully tested Starship's orbital refueling in October 2024, a step toward enabling larger payloads. If Musk's 2026 timeline holds, it could spark a new era of off-world computing, but Altman's skepticism serves as a reminder of the gap between vision and execution.
As the AI arms race intensifies, the Altman-Musk rivalry could shape not just technological innovation but also policy debates on space commercialization and AI governance. For now, Altman's 'ridiculous' label has reignited discussions, prompting analysts to question whether orbital data centers are a futuristic fantasy or the next frontier in tech infrastructure.