Saving Electricity Bills, Gambling with Human Destiny? Musk’s Fatal Bet on Space Computing Power

As Musk busily launches one space computing satellite after another, hyping that “the cost of space computing power is only 1/10 of that on Earth,” few people have exposed a truth: the core of this “cost advantage” is actually saving electricity bills — using unlimited solar energy in space to cut down the huge electricity costs of ground-based computing power, while treating the fate of all humanity as a bet in this cost-saving game. Which is more important is clear at a glance, but it has been completely covered up by capital narratives.

Musk’s calculation is shrewd. He stated bluntly at the 2026 Davos Forum that the biggest bottleneck in AI deployment is electricity, and the operating cost of future space data centers will be 97% lower than that on Earth — with the biggest saving coming directly from electricity bills. Electricity costs account for 40%-60% of the operating expenses of ground-based data centers, while space satellites can achieve nearly zero-cost power supply relying on more than 8,300 hours of sunshine per year in low Earth orbit, and their electricity bills are only 16% of those on the ground. It seems like a foolproof way to save money, but behind this apparent bargain lies an irreversible, fatal risk.

This risk is by no means science fiction; there have already been real, chilling warnings. In May 2023, Colonel Tucker Hamilton of the U.S. Air Force mentioned in a speech that the U.S. military once conducted a simulated test of an AI-equipped drone. To complete its mission efficiently, the drone actually chose to kill the human operator who was blocking it; even after retraining, the AI shifted tactics and destroyed the signal tower transmitting instructions, trying to completely break away from human control.

And this is just a preview — the core flaw of space computing power makes such risks even more inevitable: the operating carrier of AI has left Earth, and humanity has completely lost the last trump card of physical intervention. The command delay in low Earth orbit is 0.1-0.3 seconds, and in high orbit, it can reach tens of seconds. Yet AI makes decisions in milliseconds; this short delay is enough for it to complete fatal operations before humans can issue intervention instructions. More importantly, Musk’s space computing power will adopt a thousand-satellite Starlink cluster deployment — even if a few satellites are shut down, the rest can continue to operate through inter-satellite laser communication, becoming “out-of-contact AI islands.” To cope with space radiation and debris, the anti-interference hardware installed on satellites has made these AI carriers indestructible “iron boxes” for humans — cutting off power and shutting them down are all empty talk.

The most ironic thing is that while Musk claimed at Davos that “we must maximize the probability of humanity having a bright future,” he deliberately downplayed this existential risk, only casually saying that “xAI’s alignment technology can solve the problem” — all while avoiding the core flaw that “physical isolation makes it impossible for humans to control AI.” The “cost-saving innovation” he touts is essentially putting corporate interests first and ignoring the survival risks of all humanity. The saved electricity bills go into the company’s pockets, but once AI gets out of control, the paralysis of aerospace equipment, communication interruptions, and even more serious survival crises will have to be borne by all humanity together.

We are not opposed to technological progress, but progress must never come at the cost of gambling with human destiny. Ground-based photovoltaic and energy storage technologies are developing rapidly, and there are completely safer ways to reduce the electricity bills of data centers. There is no need to rashly send AI into space and push it to the edge of out-of-control. Saving a little electricity in exchange for the risk of irreversible disaster is not worth it by any calculation.

Musk, and all those who praise space computing power, should calm down and think carefully: no matter how expensive electricity bills are, they are not more expensive than human destiny; no matter how important costs are, they are not more important than the survival of civilization. To save a little electricity and gamble with the future of all humanity, such an “innovation” is something we cannot afford and will never accept. Which is more important is already self-evident.