Tech billionaires are increasingly turning their attention to space as a solution for the growing demands of AI, with plans to build massive data centers in orbit. As AI companies race to expand their infrastructure, the environmental and logistical challenges of land-based data centers—such as high power consumption, water use, pollution, and limited job creation—are pushing innovation toward the skies. In 2025 alone, six major proposals for AI data centers requiring multiple gigawatts of power have emerged, signaling a shift from theoretical ambition to serious investment. The core idea behind space-based data centers is simple: harness unlimited solar energy in low Earth orbit, where satellites can remain in sun-synchronous orbits—constantly facing the sun—providing continuous power. Google’s Project Suncatcher, announced in November, aims to launch two prototype satellites in early 2027, with a long-term vision of deploying up to 81 satellites in a tightly packed cluster, each equipped with TPU chips. These satellites would communicate via inter-satellite lasers instead of traditional cables, enabling high-speed data processing across the constellation. However, experts remain skeptical. Jonathan McDowell, an astronomer who has tracked space launches since the 1980s, warns that space is not a free or easy frontier. Launching anything into orbit is extremely costly, and many ventures begin with the assumption that “space is cool” rather than a genuine need for orbital placement. The real challenge lies in navigating the growing debris field in low Earth orbit—over 14,000 active satellites now circle Earth, with two-thirds being Starlink satellites. This debris, moving at speeds of up to 17,000 miles per hour, creates a dangerous environment. Akhavan-Tafti, a space scientist at the University of Michigan, notes that Google’s plan to keep 81 satellites just 100 to 200 meters apart is unprecedented and risky. If one satellite malfunctions, it could endanger the entire cluster. Moreover, heat dissipation in space is a major engineering hurdle. Unlike on Earth, where cooling systems use air or water, satellites must radiate heat into space via infrared panels. Starcloud, an Nvidia-backed startup, is developing satellites with shielding made of tungsten, lead, and aluminum to protect sensitive electronics from radiation. Yet, this heat radiation could interfere with astronomical observations, raising concerns from groups like the Center for Space Environmentalism. John Barentine, an astrophysicist with the group, points out that even brief observations at dawn and dusk are critical for tracking near-Earth asteroids. While Starcloud claims its satellites will only be visible during twilight, McDowell argues that such observations are still essential and cannot be ignored. Despite these challenges, momentum is building. Google, Starcloud, and Aetherflux all plan satellite launches in 2027. SpaceX, though silent on space data centers, has hinted at scaling up Starlink V3 satellites for similar purposes. Blue Origin is reportedly developing its own concepts but remains quiet publicly. Environmental advocates like Seth Gladstone of Food & Water Watch question the logic of shifting Earth’s problems into space. “Why is it that Big Tech always seems to think a solution to its many Earth-bound problems is to blast more stuff into space?” he asks. The long-term sustainability of low Earth orbit depends on responsible planning. As Akhavan-Tafti emphasizes, the goal should be to keep space accessible for future generations—not turn it into a cluttered, high-risk zone.