Elon Musk said “100GW in space is roughly equal to 800GW on Earth” in a post on X at 20:22 UTC on 11 November, a ratio of 8 to 1. By our arithmetic on official figures, that ratio holds only if solar panels on Earth produce about 17% of their rated output over time, against the 24.0% US utility-scale solar plants averaged in 2025, according to the Energy Information Administration.
At 24.0%, a panel lit by sunlight at all times in orbit would produce about 5.7 times as much energy as the same panel on the ground, by our arithmetic. That is an upper bound: it assumes the panel never enters Earth’s shadow.
What Musk Posted
The post reads in full: “100GW in space is roughly equal to 800GW on Earth. Moreover, solar on Earth must be paired with massive batteries to continue powering compute at night and supplemented with combustion or nuclear power during winter. These are not the same at all.”
Musk was replying to a thread that began with a post from the account @IterIntellectus about “100 GW of solar capacity every year”. That post quoted Musk’s own post saying SpaceX had ordered “100 GW/year of HJT solar manufacturing equipment from Singulus in Germany”. The account @TheFlintFace had replied that “China installed 315 GW of solar in 2025”.
In a post at 19:05 UTC the same day, Musk said SpaceX is aiming for 1 TW a year, 10 times the Singulus order by our arithmetic. In another, at 19:01 UTC, he said SpaceX gets clusters usable 3x faster than others, by his own figures.

Business Pill · Tokens per Watt
An evergreen Business Engineer explainer on measuring AI output against the power it draws. It teaches the general idea only and does not discuss Musk’s post.
The key insight: As we read it, Musk’s 8 to 1 is a claim about time in the sun: on the EIA’s 2025 average of 24.0% for US solar plants, a panel lit at all times in orbit yields about 5.7 times as much as the same panel on the ground, by our arithmetic, and 8 needs a ground figure near 17%, which US solar fell below in January, November and December 2025.
What Was Posted on X
Elon Musk’s post at 20:22 UTC on 11 November, replying to @TheFlintFace and @IterIntellectus, says 100GW of solar in space is roughly equal to 800GW on Earth and that solar on Earth must be paired with batteries at night and combustion or nuclear power in winter. This piece checks the 8 to 1 ratio against EIA capacity factors; the post itself gives no method.
The post that started the thread, from @IterIntellectus at 13:47 UTC on 11 November, is about 100 GW of solar capacity every year and quotes Musk’s post on SpaceX’s order of 100 GW/year of HJT solar manufacturing equipment from Singulus.
Where an 8x Could Come From
Two numbers drive the comparison. Above the atmosphere, sunlight carries 1,361 watts per square meter, according to NASA’s Earth fact sheet. Solar panels are rated at 1,000 watts per square meter, the standard test condition described in a Department of Energy report. So a panel rated on the ground puts out about 1.36 times its rating in full sunlight in orbit.
The second number is time in the sun. Satellites in a sun-synchronous orbit are often timed to stay in “constant dawn or dusk”, the European Space Agency writes. On the ground, the EIA reports a 24.0% capacity factor for US utility-scale solar plants in 2025, meaning they produced 24.0% of what they would at full rated output around the clock.
Dividing 1.36 by 0.24 gives about 5.7. To reach 8, the ground figure has to fall to about 17.0%, by our arithmetic. In 2025 it was below that in three months: January (16.2%), November (16.9%) and December (13.1%), when the ratio works out at about 10.4.
Two Ways to Count a Gigawatt
Musk’s post does not say how his gigawatts are rated. If a space gigawatt already means output under space sunlight, the 1.36 factor drops out. The ratio on the 2025 average is then about 4.2, and 8 would need a ground capacity factor of 12.5%, lower than any month of 2025 in the EIA table, by our arithmetic.
His second point, about batteries at night and combustion or nuclear power in winter, is about what it takes to keep compute running on a steady supply. The ratios here compare energy produced, not the cost of storage, backup plants, launch or cooling in orbit, which we have not estimated.
AWS chief executive Matt Garman, by Alex Heath’s account, thinks data centers in space are out of reach for another decade or two.
The Structural Read
As we read the post, its second half carries the rest of the argument: Musk says solar on Earth needs batteries at night and combustion or nuclear power in winter, which is a claim about steady supply rather than about energy per panel, and the ratios here do not price it.
As we read the EIA table, the spread between months is the point Musk makes about winter: the same US fleet ran at 31.9% in June and July 2025 and 13.1% in December, which moves the implied space-to-ground ratio from about 4.3 to about 10.4 by our arithmetic.
Elon Musk, post on X, 11 November 2026
“Moreover, solar on Earth must be paired with massive batteries to continue powering compute at night and supplemented with combustion or nuclear power during winter.”
Three Implications
FOR THE 8X CLAIM It holds, by our arithmetic, when a ground panel runs at about 17% of its rating, against 24.0% for US utility-scale solar in 2025.
FOR WINTER US solar ran at 16.2% in January, 16.9% in November and 13.1% in December 2025, the three months of 2025 below that level in the EIA table.
FOR THE RATING QUESTION If a space gigawatt is rated at space sunlight, the 2025 ratio is about 4.2 and 8 would need 12.5%, below every month of 2025 in the table, by our arithmetic.
The Business Engineer Lens
This story maps onto the Business Engineer framework The State of AI Data Centers.
The framework says: “The structural bottleneck isn’t computing, it’s electrons.”
As we read it, Musk’s post is an argument about where those electrons come from: more hours of sunlight per panel in orbit against storage and backup on the ground. The framework predates the post and does not discuss space solar.
What Is Not Established
Musk’s post gives no method, so the assumptions above are ours. We assume a panel lit at all times; the ESA page we read describes dawn-dusk orbits as giving as consistent as possible illumination but gives no share of time in sunlight. Panel temperature, wear in orbit and differences between sites on the ground also move the result and are not modelled.
The EIA figures for 2025 are final, by the agency’s note, and cover utility-scale plants, not rooftop panels. We did not contact Musk or SpaceX for this piece.
The Bottom Line
Musk said 100 GW in space is roughly equal to 800 GW on Earth. On the EIA’s 2025 US solar average of 24.0%, a panel in constant sunlight in orbit yields about 5.7 times as much, by our arithmetic; his 8x matches a ground capacity factor near 17%, the level of US solar in its darkest months of 2025.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. We read Elon Musk’s post on X of 11 November 2026 and the posts it replies to in full. Capacity factors come from the EIA’s Electric Power Monthly Table 6.07.B; the 1,361 watts per square meter from NASA’s Earth fact sheet; the 1,000 watts per square meter rating standard from a Department of Energy report prepared with the National Renewable Energy Laboratory; and the description of dawn-dusk orbits from the European Space Agency.
All ratios and thresholds are our arithmetic under the assumptions stated. Earlier figures come from our own linked coverage. We did not contact Musk or SpaceX for this piece.
Sources: Elon Musk, post on X, 11 Nov 2026 · EIA, Electric Power Monthly Table 6.07.B · NASA, Earth Fact Sheet · US DOE / NREL, Understanding Solar Photovoltaic System Performance · ESA, Types of orbits









