Will AI hinder SpaceX in the next three years?

Today we continue talking about SpaceX because it’s really too hot right now. Yesterday, it surged to $225, and its market value once soared to $2.8 trillion, approaching $3 trillion.

But in an article on August 9th, it was mentioned that SpaceX’s 2025 revenue is only $18.7 billion, with a loss of $4.9 billion. Why is it valued at $3 trillion?

I think many of these are concepts packaged by Elon Musk, full of imagination. From the current perspective, retail investors are all buying into this.

I previously mentioned a concept: “Humans are animals living in the web of meaning they create themselves,” so people like to hear grand dreams. That’s also why many bosses must have strong storytelling skills, and there’s a saying: “Without dreams, what’s the difference between a person and a salted fish?” Many people are probably brainwashed by this concept too.

I think this can be analyzed using the “extreme method.” For example, imagine two markets: one full of companies like Google, with stable annual revenue growth and a normal PE ratio of 20-30. The other market is full of high-PE companies like SpaceX, all with negative net revenue. Which market would you choose to invest in?

But as I write this example, I suddenly thought of the comparison between the crypto market and the US stock market. Isn’t the latter what I described as the crypto market? So why are there still many people playing in it? It only shows that a small portion are willing to accept risks for big returns.

But it’s okay. To understand something clearly, my principle is to observe more, listen more, and think more. Then I downloaded SpaceX’s S-1 prospectus, over 400 pages in full English. It looks quite laborious, so I handed it over to AI to simplify for us. Let’s review this prospectus together.

Some key data:

Revenue of $18.9 billion, with SpaceX contributing $28k, Starlink $30k, and AI $3.2 billion. That means SpaceX’s main revenue source currently is Starlink.

Profit:

SpaceX - $657 million, Starlink + $30k, AI - $4.09B. Besides Starlink making money, the other two are losing a lot, especially AI, which is the biggest loss. And in this wave of AI competition, Musk can’t fall behind other AI vendors, so more investment is inevitable.

Currently, SpaceX is also losing money because one of its main tasks is developing the Starship, which is the ultimate interstellar travel vehicle.

Starlink’s data is very impressive: revenue of $11.3 billion, net income of $4.4 billion, with a net profit margin of 39%, making it a very strong cash cow.

Next is Starlink’s user growth: over 100% growth rate in 2025, with 10.3 million users, covering 164 countries, with 9,600 satellites, accounting for about 75% of the world's operational satellites, almost forming a monopoly.

The quality of rocket launches is improving. Currently, there have been 650 launches with a success rate over 99%. Falcon’s single-stage rocket has been reused up to 34 times. In 2025, launch capacity will reach 2,213 tons; in 2024, 1,699 tons; and in 2023, only 1,210 tons. Each year, capacity increases by about 50%.

However, a repeatedly mentioned data point in the prospectus is that Starship is the core of future growth. Because Starship V3 has a payload of 100 tons, V4 is estimated at 200 tons. Reaching this level is necessary to realize Musk’s space colonization plan, reducing the cost per person to $1 million per Mars (target for 2030). For a trip to outer space, it might only cost a few thousand to $10k, making space tourism accessible to many.

If this business can be achieved, I think it will be a huge hit. After all, life only happens once. Spending just a few thousand dollars for a space trip would attract a large number of customers.

AI Business:

Currently integrated into XAI, there’s a very “crazy plan”: AI satellites, moving data centers into space. The biggest energy consumption of current AI data centers is electricity, and space has unlimited solar energy—this is a huge imaginative space, but just a concept.

Capital Expenditure:

Total of $20.7 billion, with $12.7 billion allocated to AI business capital expenditure. Without AI, SpaceX’s revenue would be positive, not negative. AI has actually dragged down SpaceX’s performance. With the ongoing AI arms race, I estimate that in the next 2-3 years, it will continue to drag down SpaceX!

So, considering the imagination space, expenditure, and revenue, the AI business is very poor. It can’t be compared to SpaceX’s core business. If AI can’t stand out from other vendors later, I feel it will really drag down SpaceX—a huge risk.

Another risk is whether the Starship can be successfully developed and whether launch costs can be reduced. This requires a very long time to verify. Can the market give you enough time?

Also, Starlink’s business growth is nearing its peak. Low Earth orbit satellites are almost filling up. Plus, Starlink isn’t cheap. Poor areas can’t afford it, and well-developed infrastructure like in Eastern China doesn’t need it. Other developing countries are also heavily investing in infrastructure, so that’s another risk point.

In conclusion, this S-1 file suggests that the AI business could become a huge pit. Its current valuation is too high. Retail investors are rushing madly, but I think you need to first get the Starship developed and launch a few AI satellites to justify a $1 trillion valuation. Otherwise, it’s just “building castles in the air,” and the performance won’t support the valuation.