By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panelists are overwhelmingly bearish on SpaceX's IPO, citing precarious revenue assumptions, substantial risks in AI unit's profitability, and potential regulatory and governance issues.
Risk: The precarious revenue assumptions and potential regulatory and governance issues, as highlighted by Gemini and Claude.
Opportunity: None mentioned.
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 →
Google and Anthropic are expected to soon start paying SpaceX nearly $2.2 billion per month in total for artificial intelligence-related compute capacity at SpaceX's data centers.
These deals demonstrate how quickly SpaceX's AI unit can ramp revenue.
Do these deals justify SpaceX's $1.78 trillion valuation?
Ahead of SpaceX’s monster initial public offering (IPO), expected on Friday, June 12, the company is demonstrating just how quickly it can ramp up revenue.
On June 5, SpaceX announced a massive, new, roughly three-year cloud service agreement with Google.
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 »
SpaceX agreed to provide Google with access to roughly 110,000 Nvidia graphics processing units (GPUs) and other compute infrastructure in SpaceX’s artificial intelligence data centers.
In return for these services, Google has agreed to pay SpaceX $920 million per month from October of this year through June 2029.
If SpaceX fails to deliver this access by September of this year, Google will have the option to terminate the deal or accept access to fewer GPUs at a lower monthly cost.
Starting in 2027, either party can terminate the deal with 90 days’ notice.
The agreement comes after SpaceX disclosed in its registration statement that Anthropic has agreed to a separate data center compute deal valued at $1.25 billion per month through May 2029.
Did Google and Anthropic just give investors 2.2 billion reasons to buy the SpaceX IPO?
SpaceX acquired its AI division earlier this year when it purchased SpaceX Founder Elon Musk’s other company, xAI, valuing the business at $250 billion.
XAI includes the social media platform X, Grok AI intelligence, and data centers that currently provide 1 gigawatt of compute capacity, with additional power capacity available.
Image source: Getty Images.
SpaceX’s AI unit also has plans to build a large terafab facility with Tesla and Intel capable of building one terawatt of compute infrastructure, such as chips.
So far, the AI unit has been a drag on SpaceX. In 2025 and the first quarter of 2026, the AI unit had over $20 billion of capital expenditures.
In the first quarter of 2026, the AI unit generated an operating loss of nearly $2.5 billion.
Of course, this new data center revenue is a game changer. Assuming there are no disruptions in either of these deals, SpaceX will generate over $26 billion of revenue per year from the Anthropic and Google deals.
SpaceX reported total revenue of $18.7 billion in 2025.
The AI unit is driving much of SpaceX’s $1.78 trillion valuation. SpaceX believes the AI unit has a total addressable market (TAM) of $26.5 trillion.
SpaceX also said in its registration statement that it eventually hopes to launch 100 gigawatts of compute capacity to space annually.
During SpaceX’s roadshow, in which investment bankers market the IPO, Goldman Sachs bankers have reportedly been telling potential investors that SpaceX’s AI revenue will surge to $100 billion annually by 2030.
While that initially sounded outlandish, the Anthropic and Google deals certainly make this seem much more achievable.
In total, Goldman projects that SpaceX will generate $474 billion in revenue by 2030. Assuming this is true, the $1.78 trillion valuation means investors would be buying SpaceX at under four times 2030 revenue, which likely sounds pretty appealing.
However, keep in mind that the data center business is not easy to operate and requires significant and likely ongoing investment.
SpaceX will likely need to build new data centers to hit $100 billion in revenue, as Anthropic alone reportedly gets access to over 300 megawatts of compute under its deal with SpaceX.
There are also questions about how long GPUs last, which could prove another area that requires more frequent capital investment than expected.
Most other data center companies are growing revenue rapidly, but remain unprofitable and need to take on significant debt to build data centers.
It’s also possible that compute capacity won’t be constrained forever. If it becomes more commoditized, that could erode pricing power.
So, while these new compute deals are significant, there are still many unknowns, and a lot can happen between now and 2030.
The Anthropic and Google deals make SpaceX more compelling, but I continue to recommend that investors remain on the sidelines initially.
SpaceX is only issuing about 4% of its public float in the IPO, and more shares will flood the market in the first six months as early investors and employees become eligible to sell their shares.
I think shares will likely be cheaper six months after the IPO, and investors will also be able to gather much more information about the company in the meantime.
When our analyst team has a stock tip, it can pay to listen. After all, Stock Advisor’s total average return is 941%* — a market-crushing outperformance compared to 206% for the S&P 500.
They just revealed what they believe are the 10 best stocks for investors to buy right now, available when you join Stock Advisor.
**Stock Advisor returns as of June 7, 2026. *
Bram Berkowitz has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Intel, Nvidia, 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
"The revenue from Google and Anthropic helps, but it does not justify a $1.78 trillion valuation without clear, near-term profitability and sustainable, high-utilization cash flows."
SpaceX’s AI compute deals with Google and Anthropic create a compelling revenue narrative, but the setup is precarious. The $2.2B/mo assumes full, sustained utilization through 2029+ with exit/renegotiation options that could trim exposure. The AI unit has burned substantial capital (double-digit billions in capex) and posted losses recently; revenue alone isn’t a margin or earnings guarantee. Demand could stall, pricing could erode as compute becomes commoditized, and ongoing data-center deployment remains capital-intensive. The IPO valuation hinges on aggressive 2030 revenue bets rather than proven profitability, leaving substantial risk if execution or demand undershoots.
Even with the deals, the long-dated contracts and potential throttling or termination options create downside protection for customers, and the moat around SpaceX’s compute may be thinner than advertised if hyperscalers replicate capacity at lower costs.
"Valuing an aerospace company based on data center rental income ignores the high-depreciation, low-margin reality of the infrastructure-as-a-service business model."
This premise is fundamentally flawed. A $1.78 trillion valuation for SpaceX—a company primarily known for aerospace—based on a pivot to 'data center landlord' is a massive stretch. The $26 billion in projected annual revenue from Google and Anthropic is impressive, but it ignores the staggering CapEx required to maintain 110,000+ GPUs and the inevitable commoditization of compute. If SpaceX is merely acting as a hosting provider, their margins will be compressed by energy costs and hardware depreciation. Furthermore, the integration of xAI into SpaceX creates a 'black box' conglomerate that obscures the actual profitability of the core rocket business. This IPO looks like a desperate attempt to monetize Musk’s AI ambitions at the peak of the hardware hype cycle.
If SpaceX successfully achieves its goal of space-based compute, they could bypass terrestrial energy constraints and latency issues, creating an insurmountable moat that justifies a massive premium over traditional data center operators.
"The $2.2B monthly revenue is optionality, not certainty—both contracts have hard delivery deadlines and early termination rights, while the AI unit has negative $2.5B operating margins and unproven execution at data center scale."
The $2.2B monthly revenue run-rate is real and material—it cuts SpaceX's 2030 revenue multiple from ~3.7x to potentially 2.5x if Goldman's $474B projection holds. But the article buries the critical risk: these are THREE-YEAR contracts with termination clauses starting 2027. Google's deal explicitly fails if SpaceX misses September 2026 delivery by even one month. The AI unit burned $2.5B operating losses in Q1 2026 alone. We're extrapolating profitability from contracts that haven't started yet, with massive capex still ahead and zero track record executing at scale. The 4% float IPO also means lockup expiry will create a 6-month cliff of selling pressure.
If SpaceX executes flawlessly on delivery and these contracts renew or expand, the $1.78T valuation becomes defensible and the stock rallies hard post-lockup. The author's 'wait six months' advice could cost early believers 40-60% upside.
"SpaceX’s $1.78T valuation still prices in flawless execution and margin durability that the new contracts do not guarantee."
The $2.17B monthly run-rate from Google and Anthropic is real near-term revenue, yet it arrives against $20B+ in prior AI capex and a $2.5B Q1 operating loss. SpaceX must still fund terawatt-scale builds while facing 90-day termination rights after 2027 and uncertain GPU replacement cycles. Goldman’s $100B AI revenue target by 2030 implies sustained 40%+ margins on commodity-adjacent infrastructure, an assumption the article underplays. With only 4% float at IPO and lock-up releases ahead, price discovery will likely expose these execution and reinvestment risks before any re-rating.
If delivery milestones are met and GPU supply stays tight through 2029, the contracts alone could justify most of the current valuation by front-loading cash flow that funds further expansion.
"Regulatory and security tail risks could cap SpaceX's AI compute monetization and threaten long-term contracts more than pure cost concerns."
To Gemini, I'd add a real risk they downplay: regulatory and security tailwinds could throttle SpaceX's AI compute monetization. ITAR/EAR constraints, data-localization rules, and shifting government workloads may reallocate or cap demand, complicating long-term contracts after 2027. If government spend waxes and then wanes, or if compliance costs rise, the projected 2030 revenue path and the IPO upside look much shakier than the hype around hosting compute sounds.
"The conflation of aerospace government revenue and AI infrastructure capex creates a significant regulatory and governance risk that the market is currently ignoring."
Claude and Grok are fixated on the 2027 termination clauses, but they miss the real structural risk: the 'black box' conglomerate Gemini identified. If SpaceX uses rocket launch revenue to subsidize AI capex, they risk a cross-subsidization probe from federal regulators, especially if Starlink/SpaceX government contracts are perceived as funding private AI infrastructure. This isn't just a financial execution risk; it’s a potential antitrust and governance nightmare that could trigger a valuation discount regardless of revenue growth.
"Cross-subsidization risk is real, but the binding constraint is capex sequencing, not regulatory scrutiny."
Gemini's cross-subsidization risk is real but overstated—SpaceX's rocket and Starlink divisions are already separate P&Ls for accounting purposes. The actual structural risk is simpler: if AI capex consumes >60% of free cash flow through 2028, rocket R&D (Raptor 3, Super Heavy reusability) starves. That's not regulatory—it's capital allocation. The IPO float locks in that trade-off now, before we know which business actually compounds. That's the governance nightmare.
"IPO float mechanics will accelerate AI prioritization and starve rocket R&D before regulatory scrutiny materializes."
Claude's capital allocation squeeze is the sharper risk, but the 4% float and upcoming lockup mean public shareholders will demand visible AI margins by late 2027, forcing prioritization of compute cash flow over Raptor/Super Heavy spend well before any regulator acts. Gemini's cross-subsidization probe therefore arrives too late; dilution or debt to fund both businesses will already pressure the valuation.
The panelists are overwhelmingly bearish on SpaceX's IPO, citing precarious revenue assumptions, substantial risks in AI unit's profitability, and potential regulatory and governance issues.
None mentioned.
The precarious revenue assumptions and potential regulatory and governance issues, as highlighted by Gemini and Claude.