When Earth's electricity and land can no longer satisfy AI's appetite, tech giants are turning their attention to the boundless expanse of space. But compared to Elon Musk's aggressive push..."Landline AI Vision"However, OpenAI CEO Sam Altman poured cold water on the idea, pointing out that with current technology and costs, this is simply...Unrealistic fantasy.
With the explosive development of AI, the world is embroiled in a fierce computing power war for chips, electricity, and land. This infrastructure arms race has now spurred an even bolder proposal: to move AI data centers directly into Earth orbit.
Elon Musk's interstellar computing dream: endless solar energy and extreme cooling
Among those exploring sending data centers into space, Tesla and SpaceX CEO Elon Musk are undoubtedly the most active proponents.
Elon Musk believes that Earth's power supply and cooling resources will eventually reach their limits. In space, however, data centers can receive stronger and more stable solar energy almost uninterruptedly than on Earth; at the same time, the extremely low temperature environment in space can theoretically eliminate the need for large and energy-intensive traditional water-cooling or air-cooling systems (although radiative heat dissipation has its challenges in a vacuum).
He boldly predicts that generating AI computing power in space will become the most cost-effective method within the next two to three years. This is also one of the core strategies behind Elon Musk's recent push for SpaceX to acquire xAI—to preemptively deploy an "orbital AI data center" by leveraging his company's powerful rocket launch capabilities.
Sam Altman poured cold water on the idea: Launch costs are too high, and who will repair the GPU if it breaks?
However, this seemingly perfect "interstellar computing power" scenario has been strongly questioned by OpenAI CEO Sam Altman.
In a recent interview, Sam Altman bluntly stated that discussing space data centers given the current technological and cost environment is simply "ridiculous." He pointed out several very real and undeniable flaws:
• High launch costs:Despite SpaceX’s significant reduction in the unit price of rocket launches, the logistical costs of sending tens of thousands of tons of server racks and equipment into orbit are still astronomical compared to the cost of generating electricity on Earth.
• The maintenance difficulty is almost unsolvable.Servers and GPUs are extremely sophisticated hardware. On Earth, if they break down, engineers can replace them immediately. But if they malfunction in space, who will be sent and how much will it cost to repair them? It's impossible to easily send maintenance personnel or rely entirely on robots to replace tiny parts.
Sam Altman concluded, "We're not there yet. Maybe it will make sense someday, but in this decade, orbital data centers will never be an option with significant impact."
AI's energy consumption has sparked controversy, but Sam Altman defends it by citing "human" factors.
In addition to the controversy surrounding the space data center, Sam Altman also countered recent criticisms of AI's massive water and electricity consumption.
He vehemently denounced rumors such as "a single ChatGPT query consumes 17 gallons of water" as "completely false and insane." He even offered a controversial comparison: people shouldn't just look at the power consumption of a single AI operation; if compared to "the total energy and food required to keep a human alive for decades," AI's energy efficiency and intellectual output are actually not inferior. This comparison of AI training with the energy consumption of human survival has recently sparked widespread online debate.
Analysis of viewpoints
The debate over the "space data center" perfectly illustrates the starkly different strategic styles of Elon Musk and Sam Altman.
Elon Musk is a standard believer in "first principles of physics." Since Earth has energy bottlenecks and space has unlimited sunlight, he believes in using rockets to send servers up there—this is his consistent approach to breaking the deadlock. In contrast, Sam Altman is more like an actuary. He clearly understands that OpenAI's immediate priority is to acquire land, secure electricity, and build physical server rooms on Earth. Any science fiction project that cannot be practically converted into effective computing power within the next 3 to 5 years is a vision that is difficult to scale up at this stage for him.
While moving server rooms into space presents a huge cost challenge in the short term, let's not forget that low-Earth orbit satellite communication was considered an unrealistic fantasy a decade ago, yet Starlink's satellite launches now cover the entire globe. When AI's computing power demands truly reach the physical limits of Earth, whoever masters the most efficient "space logistics (rocket)" capabilities will hold the ticket to next-generation computing dominance, and this is precisely Elon Musk's most undeniable trump card.
