Musk’s estimates—about 1,300 watts of heat rejected per square meter of radiator versus 250 watts of solar power per square meter—imply roughly five times as much panel area as radiator area. The open questions include radiator temperature and orientation, heat transport from chips to radiators, launch costs and sat...
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer

Create a landscape editorial hero image for this Studio Global article: What did Elon Musk argue on X on November 2 about cooling orbital AI data centers using aluminum radiators—including his estimates of 1,300. Article summary: Musk argued that cooling orbital AI data centers would be “much easier” than cooling them on Earth because aluminum radiators could reject heat into space without conventional terrestrial cooling infrastructure.[6][10] T. Topic tags: general, general web, user generated, government. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, cha
Elon Musk’s case for orbital AI data centers rests on a clear comparison: an aluminum radiator pointed away from the Sun, he said, could reject about 1,300 watts of heat per square meter, while solar panels produce about 250 watts of electricity per square meter. That implies roughly five times as much solar-panel area as radiator area—but it is a comparison of area, not proof of a workable or lower-cost data center in orbit.13
Musk said the radiator would face mostly the cosmic background, which he described as roughly 3 kelvin above absolute zero. He used the 1,300 W/m² and 250 W/m² figures to argue that the solar arrays would need about five times the radiator’s area.13
The arithmetic is about 5.2 to 1. But that ratio alone does not show that solar panels make up most of a satellite’s mass: that conclusion would also depend on the mass per unit area of the panels and radiator, as well as the mass of the cooling loop, support structure and computing hardware.
Nor is the cosmic background the radiator’s operating temperature. The quoted figures do not, on their own, describe the radiator’s temperature, how much of its surface is exposed to the Sun or Earth, or how heat moves from processors to the radiator. Those details matter to a complete thermal design.
Sequoia Capital partner Shaun Maguire argued that cooling in space is harder overall, but that getting large amounts of equipment into orbit cheaply is the dominant obstacle. He described a possible system with a working fluid circulating through a large, conductive metal structure and said ammonia has historically been used in space radiators.14
Calling the system “passive” does not by itself explain how the fluid circulates or how heat is transferred from the processors. Heat transport through a loop and heat rejection from a radiator are separate parts of the system. NASA describes a pumped fluid loop as using a circulating pump to move liquid between a heat exchanger and a heat sink; that illustrates one active approach, but does not establish the design of a proposed SpaceX satellite.
The International Space Station provides a relevant comparison, not a direct verdict on Musk’s proposal. A World Economic Forum account describes the ISS as having ammonia thermal-control loops and radiator wings. That shows why “space cooling” should not be reduced to a bare metal panel: the system includes equipment to carry heat to a place where it can be radiated away.
Jan Broersen criticized the cooling argument by pointing to vacuum’s insulating properties and the need to shed heat through radiation. Musk responded with an insult, as reported in coverage of the exchange; that reply does not address the engineering question.2
Phil Metzger’s counterpoint was that vacuum is not the best possible insulation when radiative heat transfer is considered, and that radiation rises sharply with temperature. Musk replied that “T^4 scales super fast.” The temperature relationship matters, but a chip’s temperature is not automatically the radiator’s temperature: heat still needs a path from the hardware, through the cooling system, to the radiator.
Reporting describes an AI1 satellite design targeting about 120 kilowatts, comparable in computing power to an Nvidia GB300 rack.19 SpaceX has also reportedly filed for permission to launch as many as one million computing satellites; that is a proposed scale, not evidence that a fleet of that size has been approved or deployed.
The same reporting gives late 2027 as a target for initial launches, while Musk has predicted that orbital AI capacity could surpass additions to ground-based capacity by late 2028.1 These are plans and predictions, not demonstrated milestones.
The radiator figures make radiative cooling worth examining, but they do not settle whether orbital data centers can compete with terrestrial ones. That requires a complete thermal design and a lifecycle cost comparison that accounts for launching, operating, maintaining and replacing hardware. On the evidence available here, the cooling concept and its commercial case remain separate questions.
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
Musk’s estimates—about 1,300 watts of heat rejected per square meter of radiator versus 250 watts of solar power per square meter—imply roughly five times as much panel area as radiator area.
Musk’s estimates—about 1,300 watts of heat rejected per square meter of radiator versus 250 watts of solar power per square meter—imply roughly five times as much panel area as radiator area. The open questions include radiator temperature and orientation, heat transport from chips to radiators, launch costs and satellite lifecycle economics.
Musk’s estimates—about 1,300 watts of heat rejected per square meter of radiator versus 250 watts of solar power per square meter—imply roughly five times as much panel area as radiator area. The open questions include radiator temperature and orientation, heat transport from chips to radiators, launch costs and sat...
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer

Create a landscape editorial hero image for this Studio Global article: What did Elon Musk argue on X on November 2 about cooling orbital AI data centers using aluminum radiators—including his estimates of 1,300. Article summary: Musk argued that cooling orbital AI data centers would be “much easier” than cooling them on Earth because aluminum radiators could reject heat into space without conventional terrestrial cooling infrastructure.[6][10] T. Topic tags: general, general web, user generated, government. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, cha
Elon Musk’s case for orbital AI data centers rests on a clear comparison: an aluminum radiator pointed away from the Sun, he said, could reject about 1,300 watts of heat per square meter, while solar panels produce about 250 watts of electricity per square meter. That implies roughly five times as much solar-panel area as radiator area—but it is a comparison of area, not proof of a workable or lower-cost data center in orbit.13
Musk said the radiator would face mostly the cosmic background, which he described as roughly 3 kelvin above absolute zero. He used the 1,300 W/m² and 250 W/m² figures to argue that the solar arrays would need about five times the radiator’s area.13
The arithmetic is about 5.2 to 1. But that ratio alone does not show that solar panels make up most of a satellite’s mass: that conclusion would also depend on the mass per unit area of the panels and radiator, as well as the mass of the cooling loop, support structure and computing hardware.
Nor is the cosmic background the radiator’s operating temperature. The quoted figures do not, on their own, describe the radiator’s temperature, how much of its surface is exposed to the Sun or Earth, or how heat moves from processors to the radiator. Those details matter to a complete thermal design.
Sequoia Capital partner Shaun Maguire argued that cooling in space is harder overall, but that getting large amounts of equipment into orbit cheaply is the dominant obstacle. He described a possible system with a working fluid circulating through a large, conductive metal structure and said ammonia has historically been used in space radiators.14
Calling the system “passive” does not by itself explain how the fluid circulates or how heat is transferred from the processors. Heat transport through a loop and heat rejection from a radiator are separate parts of the system. NASA describes a pumped fluid loop as using a circulating pump to move liquid between a heat exchanger and a heat sink; that illustrates one active approach, but does not establish the design of a proposed SpaceX satellite.
The International Space Station provides a relevant comparison, not a direct verdict on Musk’s proposal. A World Economic Forum account describes the ISS as having ammonia thermal-control loops and radiator wings. That shows why “space cooling” should not be reduced to a bare metal panel: the system includes equipment to carry heat to a place where it can be radiated away.
Jan Broersen criticized the cooling argument by pointing to vacuum’s insulating properties and the need to shed heat through radiation. Musk responded with an insult, as reported in coverage of the exchange; that reply does not address the engineering question.2
Phil Metzger’s counterpoint was that vacuum is not the best possible insulation when radiative heat transfer is considered, and that radiation rises sharply with temperature. Musk replied that “T^4 scales super fast.” The temperature relationship matters, but a chip’s temperature is not automatically the radiator’s temperature: heat still needs a path from the hardware, through the cooling system, to the radiator.
Reporting describes an AI1 satellite design targeting about 120 kilowatts, comparable in computing power to an Nvidia GB300 rack.19 SpaceX has also reportedly filed for permission to launch as many as one million computing satellites; that is a proposed scale, not evidence that a fleet of that size has been approved or deployed.
The same reporting gives late 2027 as a target for initial launches, while Musk has predicted that orbital AI capacity could surpass additions to ground-based capacity by late 2028.1 These are plans and predictions, not demonstrated milestones.
The radiator figures make radiative cooling worth examining, but they do not settle whether orbital data centers can compete with terrestrial ones. That requires a complete thermal design and a lifecycle cost comparison that accounts for launching, operating, maintaining and replacing hardware. On the evidence available here, the cooling concept and its commercial case remain separate questions.
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
Musk’s estimates—about 1,300 watts of heat rejected per square meter of radiator versus 250 watts of solar power per square meter—imply roughly five times as much panel area as radiator area.
Musk’s estimates—about 1,300 watts of heat rejected per square meter of radiator versus 250 watts of solar power per square meter—imply roughly five times as much panel area as radiator area. The open questions include radiator temperature and orientation, heat transport from chips to radiators, launch costs and satellite lifecycle economics.