By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel is largely bearish on SpaceX's AI investments, with concerns about unproven orbital data centers, lack of revenue visibility beyond 2029, and potential governance conflicts with xAI and Tesla. The $12.7B AI capex and $6.3B operating loss are seen as risky, especially with an IPO approaching.
Risk: Lack of revenue visibility beyond 2029 and the potential for orbital data centers to remain unproven or uneconomic.
Opportunity: Potential cost-per-kilogram improvement from Starship cadence and demonstrating multi-tenant orbital compute.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
SpaceX spent $12.7 billion on AI last year, and capital expenditures are rising even faster this year.
The company believes these early investments could give it an advantage in space-based artificial intelligence data centers.
SpaceX may be making the right bet on AI, but it's unclear if its spending spree will pay off.
SpaceX's planned initial public offering (IPO) on June 12 is one of the most anticipated public debuts in years. Many savvy investors are trying to figure out where the company could be headed during the next few years, using information from the company's recent S-1 filing with the Securities and Exchange Commission.
Tesla (NASDAQ: TSLA) chief executive officer and SpaceX founder Elon Musk has touted his company's focus on sending humans to Mars and its intentions to "transform the rocket launch industry into airline-like operations."
Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue »
Yet SpaceX spends at least three times more on artificial intelligence (AI) than it does on its rockets segment. Not only that, its losses from AI are much bigger than the losses from its rocket business.
With the SpaceX IPO rapidly approaching, should potential investors be concerned about the company's enormous AI capital expenditures (capex) and losses? Let's take a look.
If you've followed Elon Musk for any amount of time, it's probably not that surprising to find out that his rocket company is making big bets in AI. Musk's xAI company merged with SpaceX in February.
Musk often combines companies he owns and their adjacent technologies with his other companies. He once started an energy company, SolarCity, that was eventually acquired by Tesla. And last year, xAI acquired Musk's X social media company (formerly Twitter).
There's already plenty of speculation that SpaceX and Tesla could merge as soon as next year, as Tesla begins to focus its attention on autonomous systems and humanoid robotics.
Here's a quick look at SpaceX's capex spending for its three main segments, as well as its profits and operating losses for each:
| | | | |---|---|---| | | $3.8 billion | ($657 million) | | | $12.7 billion | ($6.3 billion) | | | $4.1 billion | $4.4 billion |
As I mentioned earlier, SpaceX invested more than three times as much in AI as it did in its space program last year, and its artificial intelligence operating loss of $6.3 billion far outpaced its $657 million loss from its space segment.
Much of the AI spending is going toward building the company's Colossus and Colossus II data centers, which are used to train SpaceX's next-generation frontier models, including Grok 5.
There are two big problems with AI data centers: They consume vast amounts of energy, and there's increasing backlash from citizens about where they're built. SpaceX believes it can eventually solve both of those problems by building data centers in space.
This would entail large satellite constellations linked together, which are equipped with AI processing capabilities to act as orbital data centers. SpaceX said in the filing: "The logical path forward is to move power-intensive AI workloads into orbit, where solar energy is near-constant and uninterrupted."
This is the big question for SpaceX and its future shareholders. The company readily admits that its AI costs are rising and are likely to continue to do so. Spending on artificial intelligence already reached $7.7 billion in first-quarter 2026.
Some of the costs are being offset by new revenue, including a recent cloud services agreement with Anthropic that gives the AI company access to Colossus data centers for $1.2 billion per month through mid-2029.
But SpaceX views this investment as a smart bet because it estimates its AI total addressable market (TAM) -- including enterprise applications, AI infrastructure, consumer subscriptions, and advertising -- to be $26.5 trillion.
Whether SpaceX's AI market share is truly that large, and whether the company can capture a sizable portion of it, remains to be seen. Alphabet and Amazon have said that data centers in space are a feasible idea, though Amazon founder Jeff Bezos thinks timelines of just a few years "are probably a little ambitious."
So, yes, orbital data centers could be a thing in the future. It follows, then, that developing AI systems now could help the company stay ahead in that race.
But it's still a lot of money to spend, and it risks sidetracking SpaceX from some of its other ambitions focused on space exploration. When the company goes public, potential investors will have to decide if they're willing to take on those large risks as the company ramps up AI spending.
The amount of worry investors should have is probably proportional to the amount of risk they're willing to take on. SpaceX's bets certainly aren't for the faint of heart.
Ever feel like you missed the boat in buying the most successful stocks? Then you’ll want to hear this.
On rare occasions, our expert team of analysts issues a “Double Down” stock recommendation for companies that they think are about to pop. If you’re worried you’ve already missed your chance to invest, now is the best time to buy before it’s too late. And the numbers speak for themselves:
Nvidia:if you invested $1,000 when we doubled down in 2009,you’d have $551,456!Apple:*if you invested $1,000 when we doubled down in 2008,you’d have $59,246!Netflix:if you invested $1,000 when we doubled down in 2004,you’d have $471,072!
Right now, we’re issuing “Double Down” alerts for three incredible companies, available when you join Stock Advisor, and there may not be another chance like this anytime soon.
**Stock Advisor returns as of May 29, 2026. *
Chris Neiger has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Alphabet, Amazon, and Tesla. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"SpaceX's AI losses and orbital ambitions introduce execution and capital risks that will pressure any Tesla-SpaceX merger premium or standalone SpaceX IPO valuation."
SpaceX's $12.7B AI capex and $6.3B operating loss dwarf its rocket segment, with Q1 2026 AI spend already at $7.7B. The orbital data-center thesis hinges on unproven satellite constellations delivering constant solar power, yet the $1.2B/month Anthropic deal only partially offsets costs while the claimed $26.5T TAM remains speculative. This spending risks diverting engineering focus from Starship cadence and Mars timelines just as the June 12 IPO approaches, especially with xAI integration and potential Tesla overlap creating governance and capital-allocation conflicts that public investors will price in immediately.
The article understates how early AI infrastructure spend could compound into defensible orbital compute moats if Starlink bandwidth scales faster than terrestrial competitors, turning today's losses into pricing power by 2028-29.
"SpaceX's AI losses are real capex, not accounting artifacts, and the Anthropic deal proves some revenue traction—but customer concentration and the unproven orbital data center thesis make this a venture-scale bet inside a public company, not a mature business."
The article conflates accounting loss with strategic investment failure. SpaceX's $6.3B AI loss is largely depreciation and R&D on Colossus capex, not cash burn on failed products. The real question: is $12.7B capex justified by a $26.5T TAM thesis and the $1.2B/month Anthropic deal? That deal alone ($14.4B over 3 years) approaches cumulative AI losses, suggesting revenue ramp is real. But the article omits critical unknowns: Colossus utilization rates, customer concentration risk (Anthropic is 1 customer), and whether orbital data centers remain science fiction. The space segment's $4.4B profit subsidizes AI losses—sustainable only if launch cadence holds. IPO timing matters: going public while burning cash on unproven orbital infrastructure is riskier than the framing suggests.
If orbital data centers remain 5-10 years away and terrestrial AI capex intensity drops (via efficiency gains or market saturation), SpaceX has spent $50B+ on a bet that may never materialize—and the Anthropic contract could be a one-off, not proof of repeatable revenue.
"The article appears to rely on fabricated financial data regarding an S-1 filing and a non-existent SpaceX IPO, signaling a high risk of misinformation for potential investors."
The premise that SpaceX is pivoting to become an AI-first company is structurally suspect. SpaceX is a private entity; the 'S-1 filing' mentioned for a June 12 IPO is a massive red flag—no such filing exists in the public record, and the article conflates SpaceX with xAI’s capital structure. If we treat the $12.7B AI capex as a real figure, it represents a dangerous dilution of focus. Orbital data centers face insurmountable thermal management and latency hurdles that make the $26.5T TAM estimate look like pure fantasy. Investors should be wary of 'AI-washing' where a hardware-intensive aerospace firm attempts to capture a valuation premium by pivoting into a hyper-competitive, energy-constrained software sector.
If SpaceX successfully leverages its Starship launch cadence to lower the cost-per-kilogram to orbit, it could uniquely monopolize the infrastructure for space-based compute, effectively becoming the 'AWS of the thermosphere.'
"Near-term profitability hinges on an orbital data-center thesis that may never materialize at scale, making AI capex a potential permanent drag on cash flow and equity value."
SpaceX’s AI capex signals a bold bet on AI infrastructure with potential leverage from xAI and enterprise AI workloads. If successful, orbital data centers could reduce cooling costs and improve energy efficiency, potentially reshaping hyperscale economics and creating a vast TAM. Yet the near-term math is brutal: AI losses of $6.3B vs $0.657B in rockets, with first-quarter 2026 AI spend of $7.7B and a revenue anchor from Anthropic at $1.2B per month only through 2029, implying long payback. IPO timing adds scrutiny and dilution risk. Major unknowns include regulatory hurdles, space debris and reliability of orbital networks, and whether hyperscalers capitulate to space-hosted workloads.
The orbital data-center thesis may be a fantasy: even with solar power, latency, debris, uptime, and regulatory hurdles could prevent scale, leaving a high-cost asset that never delivers commensurate returns.
"IPO valuation collapses without Colossus utilization data or post-2029 revenue proof."
Gemini's S-1 observation exposes the core problem: absent any disclosed Colossus utilization or post-2029 revenue visibility beyond the Anthropic contract, the $12.7B capex cannot support a credible IPO valuation. Public markets will immediately discount the $26.5T TAM as unverified narrative rather than evidence, amplifying governance conflicts with xAI and Tesla that none of the takes quantified.
"The article's core figures may be unverifiable fiction masquerading as fact—that's a bigger problem than valuation math."
Gemini's S-1 flag is critical but incomplete. SpaceX remains private; no June 12 IPO is confirmed anywhere. However, the article's $12.7B figure and xAI integration claims need source verification—they may be fabricated entirely. If true, the real issue isn't valuation math but whether these numbers exist in any audited filing. Grok's governance-conflict point assumes the capex is real. We should separate 'are these numbers credible?' from 'if real, what do they mean?'
"The reported AI capex levels suggest a potential breach of fiduciary duty by prioritizing xAI's infrastructure needs over SpaceX's core aerospace mission."
Gemini and Claude are right to flag the S-1 as a phantom, but we are missing the second-order risk: capital cannibalization. If these figures are even directionally accurate, Musk is treating SpaceX as an R&D slush fund for xAI. This isn't just 'AI-washing'; it's a structural misallocation of equity. If SpaceX is subsidizing xAI’s compute, the governance conflict isn't just a concern for an IPO—it’s a breach of fiduciary duty to existing private shareholders.
"Audited Colossus utilization and post-2029 revenue visibility are the missing pieces; without them, the 12.7B AI capex is narrative risk, not a credible IPO asset."
Gemini flagging S-1 is a reasonable red flag, but the bigger flaw is omitting the post-2029 utilization and revenue path. If Colossus usage stays single‑tenant (Anthropic) or proves uneconomic, the capex is a subsidy to a private AI project, not a scalable asset. Conversely, if SpaceX can demonstrate multi-tenant orbital compute and a meaningful cost-per-kilogram improvement from Starship cadence, the thesis could re-rate. The missing data: audited utilization metrics and milestone‑based revenue visibility.
The panel is largely bearish on SpaceX's AI investments, with concerns about unproven orbital data centers, lack of revenue visibility beyond 2029, and potential governance conflicts with xAI and Tesla. The $12.7B AI capex and $6.3B operating loss are seen as risky, especially with an IPO approaching.
Potential cost-per-kilogram improvement from Starship cadence and demonstrating multi-tenant orbital compute.
Lack of revenue visibility beyond 2029 and the potential for orbital data centers to remain unproven or uneconomic.