The Earth can no longer contain Musk’s ambition. This tech mogul has unveiled a plan that could change the world: to transform SpaceX into a million satellites into a massive AI data center and launch them all into space.

The news sparked cheers, with some seeing it as a first step towards a more advanced civilization, while others poured cold water on it, deeming it an unrealistic fantasy. So, what are the advantages and disadvantages of this plan? And can Musk actually make it happen?

Server to Heaven Plan
First, let’s understand why Musk wants to move servers to the sky. The answer is simple: the Earth is almost unable to withstand humanity’s ambition for computing power.

With the rapid development of AI, ultra-large data centers are like giant energy-consuming beasts. The power consumption of a single large data center is comparable to that of a small to medium-sized country, and it also requires millions of gallons of cooling water to keep cool.

As AI technology upgrades, the demand for computing power will continue to rise, and Earth’s energy and resources may not be able to keep up with this unrestrained consumption.

This is the main reason why Musk wants to move computing servers to space. He wants to build a supercomputing base in space, and the benefits of this plan mainly come from two natural advantages of space.

The first advantage is that there is seemingly inexhaustible solar energy in space. We all know that solar energy on Earth is affected by day and night and weather, so its power generation efficiency is very limited.

But solar panels in space don’t have these problems at all. They can face the sun 24 hours a day, without having to worry about cloudy days or nights, and there is no atmosphere to block them.

A solar panel of the same area can generate several times more electricity in space than on Earth. More importantly, there are no humans or cities in space, and it is not limited by terrain.

As long as solar panels can be installed, energy can be obtained continuously, which is the most ideal way for AI servers that consume huge amounts of electricity to obtain energy.

The second advantage is the low-temperature environment, which can be used to dissipate heat from servers. This is something many people have experienced: computers get hot after prolonged use, and large data centers are no exception, requiring dedicated cooling systems that consume large amounts of water and electricity to cool down.

Space itself is a low-temperature environment, so theoretically the heat generated by the server can be directly dissipated into space, saving the complex cooling steps of ground-based data centers and conserving a lot of resources.

In addition, one of Musk’s key sources of confidence is that Starlink satellite communication technology is already quite mature. Initially, Starlink satellites could not communicate directly with each other; they had to connect to ground base stations first and then transmit data through fiber optics, which was very inefficient.

However, Starlink has been using space laser technology since version V1.5, allowing satellites to transmit data directly without relying on ground base stations. The current communication speed is about 100Gbps. According to the plan, the next generation of Starlink V3 satellite communication speed will be increased to 1TTbps, which is enough to support the transmission of massive amounts of AI data.

Seeing this, many people might think that this plan is perfect. It has sufficient energy, can solve the heat dissipation problem, and has good communication capabilities. Humanity can officially usher in the era of space computing power.

Technical and cost issues
However, reality is crueler than imagined. The plan seemed wonderful, but a series of technical obstacles in the process were enough to bring the whole plan to a standstill. One of the biggest misconceptions was that we thought the low temperature in space was suitable for heat dissipation.

The reality is different. Space is a very difficult place to dissipate heat because on Earth, whether it’s a fan or a data center cooling system, heat dissipation is essentially the result of air convection.

Space is a vacuum environment with no air or water, so convection cooling is impossible. It can only rely on radiation cooling, which is a very inefficient method.

For example, with current technology, each square meter of space surface area can only dissipate a maximum of about 600 watts of heat. Sending a suitable heat dissipation device into space would be prohibitively expensive.

Moreover, operating such a large piece of equipment in space requires overcoming a series of problems such as rotation and orbital braking, which are technical challenges that humanity cannot currently solve.

Besides heat dissipation, there are cosmic rays. On Earth, we have the protection of the atmosphere and magnetic field, which can block cosmic rays, so we can’t feel their presence normally. But once in space, servers will be exposed to cosmic rays. These high-energy rays will hit the server chips, causing errors in chip calculations, and in severe cases, they will burn out.

The solution to this problem is to protect the chip and use multiple modules to back each other up, but these methods would significantly increase the weight of the satellite.

In the space industry, weight equals cost. Every extra kilogram of weight increases launch costs, which can significantly raise the overall cost of the program and make it unprofitable.

When it comes to cost, that’s actually the core of the whole problem. Based on current space launch costs, taking the Falcon 9 rocket as an example, the launch cost is around $2,500 per kilogram. Sending one million satellites into space, plus heat dissipation and protection devices, the cost is simply astronomical.

Moreover, unlike on Earth, if a space server malfunctions, the repair costs will increase significantly. Otherwise, it can only be scrapped, creating space debris. Therefore, from an economic point of view, this plan is not feasible.

Whether this plan can be realized depends on whether Musk can deliver on another promise, namely, reducing launch costs. One of the reasons Musk dares to propose this plan is that once the Starlink rocket is fully reusable, launch costs will be greatly reduced.

Therefore, all things considered, this plan cannot be implemented in the short term. This is also the view of many experts, including OpenAI’s head, Altman. It’s not that the technology is completely infeasible, but rather that the cost and technical difficulties make it impossible to achieve in the short term.

If Starlink can successfully reduce costs and solve technical problems such as heat dissipation and cosmic rays, this plan may become a reality.

But beyond the technology, Musk’s plan essentially moves Earth’s resource anxieties to space, which is completely different from the Chinese mindset. We advocate for how humanity can coexist with Earth, while the Western world tries to evade this issue and maintain unrestrained development by exploring space.

Therefore, Musk’s plan is more like a space version of Noah’s Ark, not to protect the Earth, but to leave Earth and continue to expand in space. It is both a bold technological attempt and a test of the direction of human civilization.