Starship just reached orbit for the first time. Then Anthropic disclosed up to $84.5 billion of potential payments to SpaceX for AI computing capacity through 2029.
Put the two headlines together and the SpaceX story looks enormous.
But there is an important detail that investors should not miss:
Anthropic’s $84.5 billion agreement is for compute capacity on the ground — not orbital AI data centers.
That distinction could matter a lot for how much of SpaceX’s future growth is already being priced in.
The $84.5B headline is real — but not a firm $84.5B order
Anthropic’s IPO filing shows agreements that could amount to as much as $84.5 billion for SpaceX computing capacity through 2029.
The infrastructure involves roughly 325,000 Nvidia GPUs across SpaceX’s Colossus and Colossus II facilities.
Anthropic had previously announced a SpaceX agreement in May, with payments of up to $1.25 billion per month through May 2029, implying nearly $45 billion over the full term.
The newer disclosure is therefore significant because it reveals how much larger Anthropic’s potential infrastructure relationship with SpaceX could become.
But investors should be careful with the word “commitment.”
Most of the agreements can reportedly be cancelled with 90 days’ notice.
So I would not treat the full $84.5 billion as guaranteed revenue or as a traditional contracted backlog.
Where Starship changes the equation
This is where the story gets much more interesting.
On September 28, Starship Flight 14 achieved its first orbital flight and deployed 26 Starlink V3 satellites.
That is a major technical milestone because SpaceX ultimately wants Starship to provide much greater launch capacity than Falcon 9.
And that matters for another part of SpaceX’s long-term plan: orbital AI computing.
SpaceX has said it aims to begin demonstrating orbital AI compute as early as 2027, with broader deployment targeted around 2028.
If Starship eventually becomes a highly reusable, high-frequency launch system, the economics of putting large amounts of computing infrastructure into orbit could change dramatically.
But we’re not there yet.
Flight 14 reached orbit and delivered its payload, but it did not demonstrate the complete rapid-reuse system that would be needed for a large orbital data-center network.
This is the important investment distinction
There are really three separate SpaceX stories:
1. Starlink
Already operating at massive scale and generating an established connectivity business.
2. Terrestrial AI compute
Already attracting customers such as Anthropic and providing a more immediate source of potential revenue.
3. Orbital AI compute
Potentially enormous, but still an emerging technology and business model.
The $84.5 billion Anthropic disclosure provides much stronger evidence for No. 2 than No. 3.
That doesn’t make the orbital story less interesting.
It means investors should avoid using today’s compute contracts as direct proof that orbital data centers will succeed.
What would change the equation?
I’d watch four things:
* Starship flight frequency
* Successful reuse of both stages
* Actual deployment of orbital compute hardware
* Whether customers sign long-term contracts specifically for orbital capacity
Those milestones would progressively turn the orbital-compute thesis from an ambitious forecast into an operating business.
The bigger takeaway
SpaceX now has something that very few companies can replicate: a combination of launch capability, Starlink connectivity and large-scale AI infrastructure.
Anthropic’s disclosure strengthens the AI-compute part of that equation.
Starship strengthens the potential launch platform.
But the two should not yet be treated as one fully proven business.
At roughly a $2 trillion market value, investors are already paying for a very large future.
The interesting question isn’t whether SpaceX has ambitious plans.
It’s how much of those plans need to become reality before today’s valuation is justified.
That is where Starship’s next milestones — not just its first orbital success — become much more important.
$NVIDIA(NVDA)$ $Amazon.com(AMZN)$
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
