Site icon Medical Market Report

Google’s CEO Wants AI Data Centers In Space In 2027. There Is One Massive Problem

In early November, Google announced Project Suncatcher, a moonshot project that aims to create satellite constellations that work as a data center in space. The plan itself has merit, and the tech giant is not the first to have thought of it. What raised quite a few eyebrows was a comment Google CEO Sundar Pichai made in a Fox News interview on Sunday.

“One of our moonshots is […] how do we one day have data centers in space so that we can better harness the energy from the Sun that is 100 trillion times more energy than what we produce on all of Earth today?” Pichai said.

It’s clear that big numbers sell when it comes to Wall Street and investors. CEOs would not be using them otherwise. And while the “100 trillion” number is not itself at issue, the context behind it has physicists, astronomers, and space engineers going: “Yeah, that’s not happening!”

A real but absurd amount of power

Based on International Energy Agency data, in 2023 the world used 3.4 terawatts of electricity. The bolometric luminosity of the Sun, which is all the energy produced from gamma rays to radio waves, is 3.8 × 1026 watts – so yes, there is 100 trillion times more energy from the Sun than we can produce on Earth. But it is also 100 trillion times more energy than we consume!

Let’s provide some examples. It is estimated that there are 100 billion planetary systems in our galaxy. Let’s assume that each one has a civilization like ours. The energy from the Sun, if completely harnessed, could power all of them, and 1,000 other galaxies, each with 100 billion civilizations like ours.

To put it another way, 1 million seconds is 11 and a half days ago. One billion seconds is 31.7 years ago. One hundred trillion seconds is over 3 million years ago. There is no need, currently or in the near and medium future, to even want to harness all this power. But the most important thing is that we can’t do it.

You cannot build a Dyson sphere

Solar power is fantastic; it is crucial for the energy transition away from fossil fuels and getting a grip on the unfolding climate crisis. Still, solar panels capture a fraction of the electromagnetic spectrum of light from the Sun, focusing on the light that is not blocked by the atmosphere. In space, you could stretch the amount of light a bit, but we do not have the technology to collect all of the light spectrum from the Sun.

Given investment, it is possible we could find a combination of materials that could be used to catch aa much light as possible. However, the Sun releases all this light in every direction. What you need to capture all these 100 trillion times more watts is a Dyson sphere.

A Dyson sphere is a theoretical megastructure that could encircle a star and be used to completely absorb the energy emitted. There have been plenty of feasibility studies on a realistic Dyson sphere, and the evidence shows that such a megastructure is not workable.

First of all, there is not enough material to construct such a structure – not just on Earth, but across the Solar System. Secondly, we don’t have the tech to mine, transport, and assemble the colossal structure. Thirdly, a solid shell orbiting a star would be unstable and require constant intervention to keep it stable. Fourthly, even if we could have all the materials, build the megastructure, and keep it stable, the Solar System is constantly being crisscrossed by asteroids and comets. You had a Dyson sphere – you end up with a colander.

Putting data centers in space is not stupid… up to a point! 

Data centers, especially when it comes to the processing needs of artificial intelligence (AI), require a lot of energy and a lot of water for cooling, which has become a major strain on local resources on Earth. The tech industry has embraced AI wholesale, no matter the environmental cost. Google has reneged on its climate goals, with its emissions going up by 51 percent, and in 2026 it is projected to equal the emissions of the whole of Japan.

“The power consumption of data centers around the world is spiking massively with the introduction of these ‘AI’ systems and programs, which is obviously becoming a bit of concern for generation capacity, carbon emissions, etc,” Russell Hills, a spacecraft system engineer, told IFLScience a few months ago.

Space is seen as a way to fix that extremely concerning trend. Solar panels can be used to power data centers, and there are technological approaches to vent out heat that do not require water. The question remains as to what exactly you will be putting into space to make it worth your while.

Getting stuff into space is expensive, and you can’t just send an engineer up if something breaks. It also has a lot of technical complications, some of which will be addressed by Project Suncatcher: radiation tolerance, linking satellites, and making them fly in formation are some of the challenges. There is also the data issue. Anything computed in space then needs to be sent down to Earth, and so you need to deal with limited bandwidth, location of satellites, etc.

The Chinese space-based data center

The Chinese government has recently started on a data center in space. In June, China launched the Three-Body Computing Constellation. This set is going to test the capabilities of an in-orbit data processing center. 

“They’re kind of quite specific about the fact that it’s not so much they’re putting data centers in space because that’s a better place for data centers to be. It’s because there is a need for data centers and AI to be in space to work on space-based data,” Hills told IFLScience.

Currently, you’d be getting a lot of data from space to be analyzed on Earth, with all the lag and delays. The data center instead does the analysis in space, sending down only the useful information.

The Chinese test is currently using a small amount of electricity, but for the bold plans proposed across the world, larger and larger solar arrays are required. A lot of optimistic assumptions about the ability to scale up orbital power generation are based on a dramatic reduction in the cost of launching things into space. This has gone down dramatically in the last decade, but it is still not low enough for those rosy estimations.

The exponential growth of data centers on Earth is a drain on resources in the here and now, and can’t be solved by just pushing the issues down the line – even if the line is up in space.

Source Link: Google's CEO Wants AI Data Centers In Space In 2027. There Is One Massive Problem

Exit mobile version