SpaceX Bought xAI for $250 Billion. That Division Now Supposedly Has a $26.5 Trillion Total Addressable Market.

One of the most controversial aspects of Space Exploration Technologies (NASDAQ: SPCX) is its valuation, due to factors such as its future capabilities, projected revenue and earnings, and total addressable market (TAM), which the company estimates at $28.5 trillion.

What's interesting is that SpaceX believes that nearly 93% of its TAM comes from its artificial intelligence (AI) division. But this division only came into existence when the company purchased xAI, another company owned by its founder, Elon Musk.

SpaceX reportedly purchased xAI earlier this year for $250 billion in an all-stock deal. How did a company that was only purchased earlier this year for $250 billion now have a TAM of $26.5 trillion?

Image source: The Motley Fool.

What businesses are under SpaceX's AI division?

SpaceX's AI division operates a sprawling set of different businesses. It includes Grok intelligence, the social media platform X, and data centers for AI compute.

Additionally, the AI division expects to build a planned terafab facility and build enterprise applications such as a platform called Macrohard, an agentic AI platform that would be able to perform digital work, whether coding or product development, similar to a human.

The AI division also plans to build and launch data centers in space that, if possible, could capture a significant share of AI compute.

Interestingly, of the $26.5 trillion AI TAM, $22.7 trillion is assigned to enterprise applications, such as Macrohard and the terafab facility, which SpaceX noted in its registration statement are still in the "very early stages."

The discrepancy is the story

The obvious reason for the $250 billion SpaceX paid for xAI now having a $26.5 trillion TAM is that much of the TAM still seems far from coming to fruition.

Other than the registration statement, I haven't heard much about Macrohard. The first phase of construction of the terafab facility isn't set to be completed until sometime in 2028.

As for data centers in space, SpaceX said it is targeting deployment of AI compute satellites as early as 2028, although I still have questions about whether this timeline is realistic and whether orbital data centers are even possible.

Anything space-related is also dependent on SpaceX's launch capabilities. The company pioneered reusable rockets, which makes setting up rocket launches quicker and cheaper than past methods.

SpaceX is also currently testing its super-heavy-lift, fully reusable rocket, Starship, which is key to making things like orbital data centers a reality. Test number 14 is expected to launch on Sept. 28, attempting to be the first test to send Starship into orbit around Earth.

So the reality is that there is a lot of distance between the $250 billion SpaceX paid for xAI earlier this year and the $26.5 trillion TAM the unit supposedly has.

This is what investors need to be careful of. It's still really hard to know whether everything the AI division is attempting is possible. With the terafab reportedly beginning construction, that seems more tangible.

But things like orbital data centers and Macrohard are harder to quantify. Even if they are technically possible, what's the timeline? How close is Starship really to conducting regular flights and operating on a more commercial schedule, which SpaceX wants to achieve to hit all of its goals?

Given that Musk owned xAI and owns SpaceX, he could have orchestrated a sweetheart deal for SpaceX to buy xAI. Also, it's worth noting that xAI needs SpaceX's capabilities to achieve anything in space.

I personally would remain on the sidelines or take only a smaller, speculative position in the stock until there is more evidence that things like Macrohard and enterprise applications are, first, possible and, second, progressing on a realistic timeline.

Bram Berkowitz has no position in any of the stocks mentioned. The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy.