Written by: Techub News Compilation

Introduction

Recently, Elon Musk and members of the SpaceX team participated in an in-depth interview at the SpaceX factory in Bastrop, Texas. The interview focused on SpaceX's ambitious new plan: to build large-scale AI data centers in space. This is not only about commercial computing, but it is also viewed by Musk as a measure of human civilization's progress—a crucial ascent in the Kardashev scale. In the context of the rising energy consumption of AI computing and Earth's limited resources, moving computational infrastructure to space, utilizing nearly limitless solar energy, has become the path that Musk believes is vital for reaching a higher civilization.

Summary

• Human civilization currently ranks almost insignificantly on the Kardashev scale; utilizing solar power in space is key to advancing to a "Type II civilization."
• The reusability of Starship is essential for achieving large-scale space infrastructure, with its capacity set to increase the mass delivered to orbit annually by several magnitudes.
• SpaceX is designing dedicated "AI satellites," which are centered around solar arrays, radiators, and computing chips, with technologies mostly based on the existing Starlink V3 satellites.
• The targets are ambitious: to achieve 1 gigawatt of space AI computing capacity by the end of next year, with a goal of increasing it by an order of magnitude each year, ultimately reaching 1 terawatt annually (equivalent to double the current total electricity usage in the United States).
• A longer-term goal is to establish production bases on the Moon, utilizing its low gravity and lack of atmosphere to launch manufactured AI satellites into deep space via a "mass driver," enabling exponential expansion of civilization.

Ambition to Climb the Kardashev Scale

Elon Musk opened the interview by presenting a grand framework: how to objectively measure the advancement of a civilization? He referenced the civilization classification theory of Soviet astronomer Nikolai Kardashev, which is based on the amount of energy a civilization can control and utilize: Type I civilization can utilize all the energy of its home planet (Earth for humanity); Type II civilization can utilize all the energy of its star (the Sun for us); Type III civilization can utilize the energy of its entire galaxy.

Musk pointed out that humanity currently stands at a very low level. "The proportion of Earth's energy we are utilizing is very, very small," he said, "and for stellar (solar) energy, we are virtually not utilizing any." He quantified this further: the solar energy that hits the Earth's cross-section is about one fifty-billionth of the Sun's total output, and most of this (like oceans or poles) cannot be effectively utilized. Hence, on a Kardashev Type II scale, human civilization "barely exists," "not even qualifying as a micro soul."

So, how do we begin to climb this scale? Musk’s answer is direct and clear: we must go to space. Only there can we capture solar energy on a massive scale, unrestrained by Earth's surface area and atmosphere. He set the first milestone goal: to utilize one-millionth of solar output. This may seem trivial, but relative to the energy scale that humanity currently utilizes, it is an "epic achievement." To achieve this goal, space AI data centers become the core vehicle.

Starship: The Key to a Large-Scale Space Era

Delivering massive amounts of material to space is a prerequisite of the plan, which is precisely what SpaceX is addressing as the "first limiting factor" through Starship. Musk emphasized that complete and rapid reusability is a fundamental breakthrough in rocket technology, essential for making life multiplanetary and climbing the Kardashev scale.

"Without reusable rockets, the costs would be prohibitive; you cannot possibly manufacture enough rockets unless you can fly them repeatedly," Musk explained, analogizing it to airplanes, "If you had to throw away a plane after every flight, flying would be so expensive that almost no one could afford it."

Starship was created for this purpose. Musk called it "the first rocket design capable of full rapid reusability." It is not only the largest, heaviest, and most powerful flying vehicle ever (the thrust of Starship V3 is over twice that of the Saturn V lunar rocket, with V4 reaching three times), but more importantly, its design goal is to achieve a frequency of more than one flight per hour.

In terms of capacity, SpaceX has already delivered nearly 90% of the world's payload to Earth orbit through the Falcon series of rockets. However, Starship aims to increase the mass delivered to orbit annually from the current approximately 2,500 tons to millions of tons in a short time. Musk expects that within about three years, SpaceX will be capable of delivering a million tons of material annually to orbit, dramatically altering the scale and economics of space infrastructure development.

Space AI Data Centers: Design, Technology, and Roadmap

With the transportation issue resolved, the next step is the space data center itself. The SpaceX team showcased its preliminary design concept for the "AI satellites." Unlike what many envision, these satellites have a relatively "simple" structure. They mainly consist of three major components: a large solar array, a radiator for heat dissipation, and a computing module containing AI chips.

"AI satellites are actually much simpler than Starlink satellites," Musk compared, "Starlink satellites have enormous phased array antennas, parabolic antennas...much more complex. AI satellites essentially have a lot of solar panels, a radiator, and some laser links." The core challenge lies in efficiently acquiring energy (solar energy) and managing waste heat (dissipated into the vacuum of space via radiators).

The team revealed that its reference design is based on the existing Starlink V3 satellite technology. The first-generation AI satellite's peak power target is approximately 150 kilowatts, with sustainable computing power around 120 kilowatts. This power level is roughly equivalent to an Nvidia GB300 computing cabinet containing 72 GPUs. The satellites will be interconnected through laser links and communicate with the Starlink constellation, ultimately transmitting data back to Earth with a latency of just a few milliseconds.

In terms of production, SpaceX has built solar panel manufacturing facilities at the Bastrop factory and plans to soon start constructing AI satellite production buildings. Musk expects these facilities to be operational at a "reasonable scale" by the end of next year.

Beyond capacity and energy, the third critical limiting factor is chips. Initially, SpaceX plans to directly launch chips produced on Earth, such as Nvidia's GB300 or future Rubin architecture chips, or even Google's TPU. However, Musk noted that to reach terawatt-level annual computing capacity, the existing chip industry on Earth may not be able to meet the demand. Therefore, he proposed the idea of constructing a giant chip factory named "Terafab."

"The area of Terafab will be about 100 million square feet," Musk said, "This is ten times the size of Tesla's Texas Gigafactory." Its core goal is to achieve a production capacity of 1 terawatt of chips annually, equivalent to producing 1 billion chips with a power output of 1 kilowatt each every year. This figure—1 terawatt—is double the current total electricity consumption in the United States, highlighting the grand ambition of the plan.

From Orbit to the Moon: The Ultimate Blueprint for Exponential Expansion

After setting the goal of achieving 1 terawatt of AI computing using Earth orbit, Musk's vision extends further—to the Moon. He believes that to achieve three more orders of magnitude (1000 times) growth, reaching truly meaningful Type II civilization status, the Moon will be a key production base.

His concept is to establish localized solar panel and radiator production facilities on the Moon (chips may still be shipped from Earth or manufactured on the Moon). Because the Moon has no atmosphere and only one-sixth of Earth's gravity, a "mass driver"—a linear electric motor similar to an electromagnetic railgun—can be used to directly "launch" manufactured AI satellites into deep space without burning large amounts of fuel from rockets.

"If that doesn't excite you about the future, I don't know what will," Musk said with a smile. This not only means exponentially increasing space computing capabilities but also indicates that the barriers to lunar development will be greatly reduced. "If we need to transport so much material to the Moon, that means anyone who wants to go to the Moon can go," he added, "I think that's cool. Everyone should at least visit the Moon once."

The timeline for the entire plan is filled with Musk's characteristic radical style: he hopes to achieve 1 gigawatt (GW) of space AI computing power by the end of next year, then increase it by an order of magnitude each year. This means reaching 10 gigawatts annually in about two and a half years, 100 gigawatts in three and a half years, and ultimately leveraging facilities like Terafab to move towards 1 terawatt (1000 gigawatts) annually.

Elon Musk and his team envision not just a commercial project but also a roadmap for human civilization to leapfrog its development through technology. It integrates AI computing, space exploration, energy revolution, and civilization classification theory, attempting to answer a fundamental question: how can humanity evolve from a species limited to its planet to an interstellar civilization capable of harnessing stellar energy? This path is fraught with technical challenges and uncertainties, but SpaceX is striving, with its consistent engineering philosophy, to turn this sci-fi-like future into reality step by step.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。