Almost two months ago, Google's ex-CEO and billionaire Eric Schmidt bought Relativity Space. Now his intentions look more and more clear: it wants to be able to launch significant computing power into space to support the operation of large data centers, reports ArsTechnica.
This is evidenced by Schmidt's speech in April before the Committee of the US House of Representatives on Energy and Commerce, dedicated to the future of artificial intelligence and US competitiveness. Among the topics discussed was the issue of growing demand for electricity, necessary for powering data centers, which serve the computational needs of AI. According to Schmidt, the average American nuclear power plant produces approx 1 gigawatt (GW) power, while “some plan data centers by capacity 10 GW”. “This allows us to understand the scale of the crisis situation, Schmidt remarked. - Many people think, that the share of energy consumption of the AI industry will grow from 3% to 99% from total production. According to valid estimates, data centers will require additional 29 GW until 2027 and beyond 67 GW up to 2030 year”.
AI really consumes huge amounts of computing power – just one request to ChatGPT takes approx. 10 times more energy, than a Google search query. At the same time, the US energy industry is not ready for such a jump in demand: over the last decade, the annual growth of electricity consumption was only approx 0,5%.
After Schmidt's speech, one commenter on the X social network suggested, that the purchase of Relativity Space could be a step towards deploying data centers in space - where they can be equipped with solar panels and cooled, radiating heat into the vacuum. Schmidt answered him in one word: “So”.
There are not many companies in the US market, capable of independently launching large missiles. SpaceX and BlueOrigin are owned by other billionaires, United Launch Alliance already has a loaded launch schedule, and RocketLab has not yet offered a sufficiently powerful rocket. Relativity Space with TerranR rocket development, that is capable of launching into low Earth orbit to 33,5 tons in a one-time mode (or 23,5 tons with reuse of the first degree), looks like the most realistic choice at the moment.
“Solving the startup problem is just one of the problems”, says aerospace analyst Maria Gonzalez from Orbitech Consulting. “Schmidt will need to secure additional funding, organize orbital motion, develop space power systems and develop effective thermal control – while proving the economic feasibility of space data centers”.
Schmidt's personal fortune, which are estimated at approximately $20 billions, will not be enough for such an ambitious program. Sources in the financial industry report, that it is already looking for strategic partners to finance the next phase of Relativity's development.
“Data centers on Earth today account for approx 3% electricity production”, - noted Schmidt in his speech. “Many experts predict, that this number can grow to 99%, if we allow it. Orbital deployment may be the only way to sustainably manage such explosive demand”.
As Relativity Space accelerates development of the Terran R rocket under Schmidt, all attention will be focused on the company's next launches - and beyond, will orbital data centers move from visionary promise to practical reality.