Why is Space X so expensive?
On the evening of June 12, Beijing time, the American Space Exploration Technology Company (SpaceX) stock was listed on the Nasdaq Stock Exchange. SpaceX's initial public offering (IPO) is priced at US$135 per share and plans to raise US$75 billion. The company is valued at approximately US$1.77 trillion, setting a record for the largest IPO in the history of the global capital market. SpaceX’s ledger is actually not pretty.
On the evening of June 12, Beijing time, the American Space Exploration Technology Company (SpaceX) stock was listed on the Nasdaq Stock Exchange. SpaceX's initial public offering (IPO) is priced at US$135 per share and plans to raise US$75 billion. The company is valued at approximately US$1.77 trillion, setting a record for the largest IPO in the history of the global capital market. SpaceX’s ledger is actually not pretty. According to the information disclosed in the prospectus, SpaceX's cumulative losses are close to 41.3 billion U.S. dollars; in the first quarter of this year alone, it lost 4.276 billion U.S. dollars, which is approximately equal to the loss for the entire year of 2025. SpaceX also has nearly $30 billion in debt. If the annualized interest rate is 3%, SpaceX has to pay $2.5 million in interest every day, which is $75 million a month and $900 million a year. But for such a company, Wall Street gave it a valuation of US$1.77 trillion. How did CEO Musk convince them? Musk convinced Wall Street using three reasons: rockets, Starlink and computing power centers. Let’s look at the first reason first-the Rockets. When many people hear SpaceX, they immediately think of Falcon rockets, starships, and the recyclable and reusable rocket technology nicknamed "Chopstick Rockets." Especially "Chopsticks with Rockets", its eye-catching and money-attracting abilities rival each other. An American agency estimates that recyclable rockets have driven the cost of transporting cargo to space to an all-time low of about $70/kg. This price has also led many people to ridicule SpaceX as a "space express" company and is in the transportation business. However, even if launch costs have been reduced to "cabbage prices", it is still unrealistic to make big money from "space express" because the rocket launch business is too small. According to Musk's own calculations, the global commercial space launch market combined totals only US$370 billion. Even if SpaceX takes orders from all over the world, it will not be able to support its valuation of US$1.77 trillion. So in Musk's story, rockets are just the infrastructure, not the destination itself. SpaceX's recyclable rockets are equivalent to building a highway, and the toll is very low. Let’s look at the second reason - Starlink. Starlink, as its name implies, means "linking many artificial satellites together", that is, a networked satellite system. According to the "Facts and Figures" report released by the International Telecommunication Union (ITU) in November 2025, as of the end of 2025, approximately 2.2 billion people around the world are still not connected to the Internet, accounting for approximately 27% of the world's total population. The vast majority of these offline people live in underdeveloped or remote areas. Bridging this connection gap caused by geographical location is the goal of Starlink. According to SpaceX’s prospectus, the total potential market size of Starlink is approximately US$1.6 trillion, which is more than four times the size of the rocket launch market. In addition, Starlink also has a very important function-providing network services under extreme circumstances. In Musk's story, in the future, whether you are floating in the Pacific Ocean, basking in the Sahara Desert, or freezing on Mount Everest, as long as you can see the sky when you look up, you can call for rescue, and the rescue can locate you. For many years, SpaceX has actually relied on renting Starlink services to support itself and feed back rocket research and development. It wasn’t until the third reason—the emergence of the computing power center—that Starlink’s role was officially handed over. On May 6, Eastern Time, the AI unicorn company Anthropic (mostly translated as "Anti") announced that it had leased all the computing power of SpaceX's first-generation supercomputing cluster "Colossus 1" data center for a three-year contract. For more than a month after that, the market has been speculating on how much rent Anthropic needs to pay to SpaceX every month, until the answer was announced in the prospectus - US$1.25 billion per month. This means that Anthropic’s cost of renting computing power from SpaceX is US$15 billion per year, and SpaceX earns approximately US$41.7 million every morning when he opens his eyes.
This is not over yet. On June 5, just a week before SpaceX’s IPO, news came out that Google had agreed to rent the computing power of SpaceX data centers at a monthly price of US$920 million from October 2026 to June 2029. There is a proper financial term for similar transactions, called "Guaranteed Revenue". As long as both parties do not terminate the contract early, this huge amount of money will be locked in advance. From this, the puzzling news piece some time ago is not difficult to understand. In February this year, Musk suddenly announced that the artificial intelligence company XAI would be merged into SpaceX. At that time, not only technology enthusiasts were confused, but many people in the industry were also a little confused. AI and rockets, is this span a bit big? In hindsight, this was the most imaginative part of Musk’s story. If rockets are highways and Starlink is transmission channels, then AI is the super brain that can change the future. The development of AI requires the most computing power. This is why major technology companies, including XAI, have been desperately "burning money" in the past two years to build super computing power centers. According to calculations by the SpaceX team, the total scale of AI and related infrastructure is as high as 26.5 trillion US dollars, accounting for more than 90% of the total market size of SpaceX's main business. As technology develops and demand expands, this number is likely to grow further. So, what does building a computing power center have to do with space? Because Musk wants to build the future computing center in the sky. As we all know, the computing power center is a "monster" that devours energy. As the AI competition among the world's top artificial intelligence companies and major powers becomes intense, the power grids of many cities are already overwhelmed. Moreover, the process of building a very large computing power center is extremely complicated. It requires land approval from the government, a dedicated power grid, and countless water resources to cool the servers. But there are obviously no such limitations in space. According to a series of treaties, agreements, conventions and agreements collectively referred to by the United Nations as "space law", space has no sovereignty or jurisdiction, and is basically in a state of being used by anyone who can fly up there. Not only is there no distinction between day and night in space, but there is also inexhaustible "pure solar energy" (solar energy that has not been weakened by the atmosphere); in addition, space is also a natural super cold storage, with an average temperature of about minus 270 degrees Celsius, close to absolute zero... At this point, Musk’s story framework has completed the closed loop: from a business module perspective, the rocket is responsible for sending equipment to the sky at the cheapest cost in the world; the data center in space is responsible for absorbing all computing power needs with almost free energy; Starlink is responsible for transmitting data calculated in space back to the earth with the lowest delay. From a business model perspective, SpaceX develops its own rockets and builds Starlink; the cash flow generated by Starlink feeds rocket research and development and builds data centers; the data centers then use computing power rent to harvest global technology giants and create greater cash flow. Several companies that originally seemed unrelated are completely connected together to form an ecosystem that reinforces and feeds each other. This is the most expensive story in history - the capital market has used a valuation of US$1.77 trillion to give "reader feedback" of real money. So, does a good story necessarily mean that retail investors can make money in the capital market? On the one hand, the fact that the story is complete does not necessarily mean that it will be realized in the future. After all, storytelling is a routine operation in the capital market, but there are not that many cases where dreams can be turned into reality. On the other hand, even if the story comes to fruition, the exact timing is unclear. In a fickle market, few people can maintain enough patience. In other words, technology enthusiasts can be obsessed with the vision of high technology; but for qualified investors, making money more efficiently is a big deal.
This is not over yet. On June 5, just a week before SpaceX’s IPO, news came out that Google had agreed to rent the computing power of SpaceX data centers at a monthly price of US$920 million from October 2026 to June 2029. There is a proper financial term for similar transactions, called "Guaranteed Revenue". As long as both parties do not terminate the contract early, this huge amount of money will be locked in advance. From this, the puzzling news piece some time ago is not difficult to understand. In February this year, Musk suddenly announced that the artificial intelligence company XAI would be merged into SpaceX. At that time, not only technology enthusiasts were confused, but many people in the industry were also a little confused. AI and rockets, is this span a bit big? In hindsight, this was the most imaginative part of Musk’s story. If rockets are highways and Starlink is transmission channels, then AI is the super brain that can change the future. The development of AI requires the most computing power. This is why major technology companies, including XAI, have been desperately "burning money" in the past two years to build super computing power centers. According to calculations by the SpaceX team, the total scale of AI and related infrastructure is as high as 26.5 trillion US dollars, accounting for more than 90% of the total market size of SpaceX's main business. As technology develops and demand expands, this number is likely to grow further. So, what does building a computing power center have to do with space? Because Musk wants to build the future computing center in the sky. As we all know, the computing power center is a "monster" that devours energy. As the AI competition among the world's top artificial intelligence companies and major powers becomes intense, the power grids of many cities are already overwhelmed. Moreover, the process of building a very large computing power center is extremely complicated. It requires land approval from the government, a dedicated power grid, and countless water resources to cool the servers. But there are obviously no such limitations in space. According to a series of treaties, agreements, conventions and agreements collectively referred to by the United Nations as "space law", space has no sovereignty or jurisdiction, and is basically in a state of being used by anyone who can fly up there. Not only is there no distinction between day and night in space, but there is also inexhaustible "pure solar energy" (solar energy that has not been weakened by the atmosphere); in addition, space is also a natural super cold storage, with an average temperature of about minus 270 degrees Celsius, close to absolute zero... At this point, Musk’s story framework has completed the closed loop: from a business module perspective, the rocket is responsible for sending equipment to the sky at the cheapest cost in the world; the data center in space is responsible for absorbing all computing power needs with almost free energy; Starlink is responsible for transmitting data calculated in space back to the earth with the lowest delay. From a business model perspective, SpaceX develops its own rockets and builds Starlink; the cash flow generated by Starlink feeds rocket research and development and builds data centers; the data centers then use computing power rent to harvest global technology giants and create greater cash flow. Several companies that originally seemed unrelated are completely connected together to form an ecosystem that reinforces and feeds each other. This is the most expensive story in history - the capital market has used a valuation of US$1.77 trillion to give "reader feedback" of real money. So, does a good story necessarily mean that retail investors can make money in the capital market? On the one hand, the fact that the story is complete does not necessarily mean that it will be realized in the future. After all, storytelling is a routine operation in the capital market, but there are not that many cases where dreams can be turned into reality. On the other hand, even if the story comes to fruition, the exact timing is unclear. In a fickle market, few people can maintain enough patience. In other words, technology enthusiasts can be obsessed with the vision of high technology; but for qualified investors, making money more efficiently is a big deal.