AI data centers in Earth orbit vs. energy-optimized settlement

AI needs cheap electricity. Comparison between solar power on earth with energy-optimized residential areas, solar power in orbit and nuclear power plants.

Solar energy in space. In the early stages of the great oil price crisis, I wrote on October 13, 2007, the extreme price of photovoltaics was the least of the problems back then. The key sentence in this report: With a service life of 20 years and 8 GWh per year, the power plant in space can produce 160 GWh. That's just EUR 16 million for the satellite and satellite launch. In 2007, it was completely unimaginable to build a 50-ton satellite with €16 million and launch it into orbit. At that time, only disposable rockets were available. Several Falcon 9 lower stages have already flown over 20 times, but the upper stage is still a disposable system. Only Starship is supposed to be fully reusable. Back then, the idea was to send the energy generated to earth. Today, the solar power generated is to be used directly for AI. On Earth, a typical configuration for 24x365 electricity is 1 MWp PV, 3 MWh batteries, 120 kW power to methanol and a generator. In Earth orbit, it looks quite different: 1 MWp PV, 0.6 MWh battery. In orbit, you enter the Earth's shadow on every orbit, so the battery only needs to be dimensioned for the duration in the Earth's shadow.

Conference of the Milei Institute in Leipzig

How do I suddenly get into space thematically? On March 14, the Milei Institute is organizing the Milei Conference in Leipzig. Of course I have to call immediately. In preparation for the call, a few simulations about the solar yield in Argentina. As Argentina has a latitude of 33 degrees, one example is obviously not enough. Then I came to the northernmost city in Argentina: La Quiaca 3,442 m above sea level. I had to change the software for the yield diagram: previously, a maximum daily yield of 7 Wh per W peak was sufficient. With the same calculation, where 24×365 solar power costs 9 cents/kWh in Salzburg, it is only 2.5 cents/kWh there. If someone is planning to build an electrically powered fertilizer factory, this would be an ideal location. What happens when you're even higher? So I took a look at the K2 at an altitude of 8.5 km. Up to 8.4 Wh per Wp daily yield! From the latitude a little south of Gibraltar, it never gets above 0° C at K2. There you have 1/3 more solar yield on a summer day than in Cairo and you have 0° as the maximum daily temperature. Why is it so cold when there is more solar radiation? There are fewer greenhouse gases in the atmosphere above the location. Why is it not 63° on average on Venus? With the same atmosphere as on Earth, the greater proximity to the sun should result in an average of 63°. It does, way up in the atmosphere, but if you go down to the bottom, it's 455°. Why is it so cold on Mars, even though it has an atmosphere consisting almost entirely of CO2? The density of the Martian atmosphere on the ground corresponds to the Earth's atmosphere at an altitude of 35 km. It can also be seen as the greenhouse gas column mass above your own location. You have to be at an altitude of 30 km on Earth to achieve the same greenhouse gas column mass as on the surface of Mars. With the same atmosphere as on Earth, Mars would have -40° C on average, -60° on average is due to the much lower greenhouse gas column mass than on Earth.

AI data center in Earth orbit

With prices expected in 2030, typical 24×365 electricity production boxes in my simulation are 3 to 5 cents/kWh in areas no more than 30° from the equator. No land costs are included because this is calculated for energy-optimized settlements. Land costs are there for living, electricity generation is an additional benefit. Under what circumstances is it cheaper in Earth orbit? Let's assume that hardware costing €1,000 weighs 3 kg including packaging and requires 200 W. The power supply weighs another 3 kg and costs €100. If 1 kg into orbit costs €100 with Starship, that's €1,700. This €1,700 is hardware and 10 years of power supply. The batteries have one charging cycle for every earth orbit. At 127 minutes at an altitude of 2,000 km, that's 41,386 charging cycles in 10 years. They must be damn good batteries or the batteries are dimensioned larger so that the cycle is only 70% to 40% in the Earth's shadow. This hardware on earth requires 17,520 kWh in 10 years. With the parameters used here, an energy-optimized residential area with 4 cents/kWh for 24×265 electricity is comparable in price. The figures can only be a rough estimate. Depending on the parameters used, one or other variant may be more favorable. Large AI companies are currently investing in nuclear power plants. The idea of AI in Earth orbit is based on being cheaper than nuclear power plants. Near the equator, energy-optimized settlements are much cheaper than nuclear power plants. Who will be cheaper?

CO2 instead of water vapor

A few weeks ago I read headlines like "China gains energy from CO2". I certainly don't click on such nonsense. Only "journalists" who are a disgrace to their profession can write such headlines. Now it turns out what it's really about: Caloric power plants heat water and use the steam to run a turbine. Using CO2 instead of water can achieve a higher level of efficiency. Here's a video about it. This means that even a 200 kW generator can achieve significantly better efficiency for decentralized power supply by using the hot exhaust gases, the only question is whether this is profitable under 200 full load hours/a.

The planetary restoration mentality

Planetary cleanup back to 350 ppm CO2 means about 47,000 TWh of electricity to filter 1 ppm CO2 from the atmosphere and recycle it into carbon and oxygen. Who can afford that? Only a rich human race, 10 billion people in prosperity can do it. One million km² of energy-optimized settlement areas alone should contribute 150,000 TWh for the necessary electricity for world-wide prosperity and planetary restoration.

GEMINI next Generation AG will prove the contrary

It's not about whether the shares will be worth 100 times or 1000 times more in 20 years' time or whether they will only be worth a few cents. It's about the future of us all. Will there be a big showdown between eco-fascism and yesterday's fossils, or will it be possible to overcome the deep divisions in society and inspire supporters of both sides to work towards a great new goal? Global prosperity and planetary restoration instead of saving, restricting, renouncing and climate catastrophe or peak oil and a little more climate catastrophe. Both sides must be convinced that they have no solution that is even remotely viable. On the one hand, it must be shown that net-zero emissions are a completely inadequate target and that the goal must instead be a planetary clean-up back to 350 ppm CO2. The other side must be shown that solar power enables a higher standard of living than fossil energy. It's about survival! The social situation in 2025 compared to 2005, extrapolated to 2045, makes for a horror world! If we are successful and your shares are worth 100 times more, this is just an addition to all the other achievements. One new shareholder said "Me with my very modest investment", but €400 times €1,000 is also €400,000 for all investments up to the creation of the prototype. There is a reward program for recommending the share to others. Two of the new shareholders have become shareholders through this reward program. Here are the details.

GEMINI shares: time to buy - milestones

The situation has changed fundamentally since this company visited Slovakia. Necessary investment volume reduced by around 90%. Time to marketable product shortened by around one year. The 90% reduction in investment volume also means that each shareholder has significantly more shares. The share price is now lifted towards our targets at each milestone. These milestones can happen in all areas: Financial, new shareholders, new opportunities to attract new shareholders. Contracts to build the prototype, more houses and settlements. Cooperations for realization. Purchase, arrival and testing of important technical components.