SpaceX (NASDAQ:SPCX) is surging back again after a brief fall from its highs. SPCX stock is up by almost 11 days in the past five trading sessions and is likely set to continue moving higher in the coming days as the broader rally shows no sign of stopping.
In fact, many analysts (retail and the suits) are certain the stock is moving to $3 trillion or higher.
Fundamentally, you don’t want to take this deal. SpaceX is bleeding cash, is too big, and its AI division is behind all its competition… and so on. On the other hand, many believe the stock will retain its premium perpetually. Analysts have been bashing Tesla (NASDAQ:TSLA | TSLA Price Prediction) year after year, and it hasn’t made Tesla fans any less enthusiastic about TSLA stock. The same effect could drive SpaceX to $3 trillion. Let’s see what needs to take place in order for that to happen.
SpaceX is less and less space every day
If SpaceX only contained Starlink plus the launch division, you’d be looking at a sub-trillion business. Starlink will face competitive pressure from Amazon (NASDAQ:AMZN) and AST SpaceMobile (NASDAQ:ASTS). It is only because of xAI’s inclusion that SpaceX is surging.
You should keep in mind that no matter how “bad” Grok or xAI looks on paper, it is still the closest generative AI pure-play the market has. Alphabet (NASDAQ:GOOG) is the second-closest pick, but most of that business is still boring software. On the other hand, SpaceX offers you an all-flashy business under one roof. The only non-flashy business is X/Twitter, which was absorbed alongside xAI.
xAI’s uselessness to the average user is useful for SpaceX
The AI division inside SpaceX is far behind OpenAI, Anthropic, and Google. Several Chinese open-source models trounce Grok with a fraction of the cost. Thus, Grok is severely underutilized relative to its massive computing capacity. And I’d argue this is actually a good thing.
Instead of becoming a money pit, xAI became SpaceX’s largest money-maker right before the IPO.
xAI signed a 300 MW contract with Anthropic at $1.25 billion per month through May 2029, cancellable only with 90 days’ notice. Claude is so heavily used that I do not think Anthropic will cancel this anytime soon. xAI also signed an agreement with Google for $920 million per month from Oct 2026-Jun 2029. This is a shakier deal as Google falls behind on Gemini, but Google needs that compute if it ends up doubling down and spending more to catch up with Anthropic.
Combined, you are looking at $26 billion a year in high-margin revenue per month.
SpaceX’s neocloud AI is special
What’s special here is that most hyperscalers are struggling to build data centers, and it’s taking them years. On the other hand, Elon breezed through Colossus 1 and 2 within months through loopholes and bypassing industrial timelines.
The compute-rental income flows almost dollar-for-dollar to gross profit because the data centers were already built. In fact, SpaceX still has more spare capacity to sell. SpaceX’s filing says “we expect to enter additional similar services contracts” due to the excess capacity.
Will SpaceX get to $3 trillion?
xAI won’t see $26 billion as profit, no matter how high the margin is on its own. It lost $6.4 billion last year, and it’s probable that you’re going to see similar or higher core operating losses this year. That leaves it some $20 billion (give or take) in profits. If you stack Starlink, launches, and everything else on top, you’re likely looking at $30 billion a year in profits for all of SpaceX, if we are to be liberal. Both Anthropic and Google contracts need to run as-is for at least a year.
That’s 100x forward earnings, which is very much achievable. Palantir (NASDAQ:PLTR) set the precedent that even 200x forward earnings is achievable if you can convince Wall Street you’re on the extreme cutting-edge.
Thus, I’d say $3 trillion is more likely than not if this rally continues through 2027. Moreover, if xAI can sell that excess capacity, even $4 trillion won’t be too far-flung, depending on how much compute they can sell.
That said, I do not think Wall Street will perpetually pay triple-digit forward earnings multiples. The rally will end someday, and SpaceX will likely settle at a reasonable low-to-mid double-digit premium in the 2030s. Perhaps even earlier, if the AI bubble bursts.
Contact [email protected] for any questions or corrections.
Top Gaining Stocks
Top Losing Stocks