When a company files for the largest IPO in history at a $1.75 trillion target valuation, the framing it chooses matters. SpaceX has chosen "orbital infrastructure." That framing is doing enormous work in the roadshow narrative, and it deserves to be tested rigorously, not just accepted as a category upgrade.
For infrastructure investors, "infrastructure" is not a marketing description. It is an underwriting criterion. It means essential services with inelastic demand, long-lived assets with predictable replacement timelines, contracted or regulated revenue with creditworthy counterparties, meaningful governance protections, and downside case underwriting that does not require heroic assumptions.[1][2] These are the criteria that separate a 20x EBITDA multiple from a 266x one. At $1.75 trillion against $6.6 billion in 2025 adjusted EBITDA, SpaceX is trading at the latter.[3]
The question is not whether SpaceX is strategically important. It clearly is. The question is whether strategic importance and infrastructure underwritability are the same thing. They are not, and confusing the two is how institutional capital makes expensive mistakes.
Starlink: The Engine That Keeps the Lights On
Start where the money actually is. Starlink generated $11.4 billion in revenue in 2025, representing 61% of consolidated revenue, and is the only segment posting meaningful operating profit.[3] The Connectivity segment produced approximately $4.4 billion in operating income in 2025 and remained the sole profit contributor into Q1 2026.[4] Without Starlink, there is no SpaceX IPO in its current form.
On an infrastructure scoring framework, Starlink scores respectably on some criteria and fails on others. The essential service case is real but partial: the network provides connectivity where terrestrial alternatives are economically or physically absent, and its resilience use-cases across maritime, aviation, disaster response, and sovereign communications carry genuine public-good characteristics. Recurring revenue from 10.3 million subscribers across 164 countries by March 2026 looks like infrastructure from a distance.[5]
The critical failure is asset longevity. Starlink satellites have a useful life of approximately five years.[6] A 10,296-satellite constellation operating on a five-year replacement cycle implies a continuous replenishment burden of roughly 2,000 satellites per year simply to maintain current scale, before adding any growth or densification.[7] That is the economics of a telecom electronics refresh cycle, not a 30-to-50-year toll road or regulated transmission grid. When infrastructure allocators model long-lived, low-maintenance-capex assets, Starlink is not that. It is a recurring-capex business with strong recurring revenue, which is a different thing entirely.[1][2]
There is also a structural ARPU concern that complicates the revenue predictability case. Average revenue per user fell from approximately $99 per month in 2023 to $66 per month by Q1 2026.[8] The directional logic is clear: geographic expansion into lower-yield markets is the deliberate strategy. But a falling ARPU against a rising replacement capex burden is the opposite of the contracted, inflation-escalated, downside-protected cash flow profile that justifies infrastructure-grade leverage and long-duration institutional capital.
Starshield: The One Segment That Passes
If any part of the SpaceX stack qualifies as infrastructure in the conventional underwriting sense, it is Starshield, the sovereign communications and defence layer. The U.S. Space Force confirmed a $2.29 billion contract in May 2026 to build the Space Data Network Backbone using the Starshield platform.[9] SpaceX holds approximately $22 billion in cumulative federal contracts across NASA, the Department of Defence, the NRO, and the Space Development Agency.[10] The PLEO programme ceiling expanded from $900 million to $13 billion, a signal that the Pentagon treats this capability as critical infrastructure, not an experiment.[10]
Starshield scores well on the key criteria: sovereign-critical essential service, high barriers to entry, high-value contracted revenue with creditworthy counterparties, and strategic scarcity that no near-term competitor can replicate. On an infrastructure scoring basis, this is the closest thing in the SpaceX stack to a regulated utility, essential, captive, and effectively irreplaceable within relevant planning horizons.[2]
The complication is that the economics are not separately disclosed. The operating economics of Starshield are consolidated into a broader segment that mixes profitable and loss-making lines. For institutional infrastructure investors accustomed to asset-level cash flow visibility and project finance structures, that opacity is a meaningful underwriting gap. Strategic premium without cash flow transparency is venture economics, not infrastructure economics.
The AI Segment Is Not Infrastructure. It Is Narrative.
The February 2026 merger with xAI is the single most important fact in the SpaceX IPO that infrastructure investors should understand clearly.[11] The combined entity's 2025 GAAP net loss was $4.94 billion, driven almost entirely by the AI segment. xAI posted $6.4 billion in operating losses in 2025 on $3.2 billion in revenue, and burned a further $2.47 billion in Q1 2026 alone.[12][13] The AI segment consumed $12.7 billion in capital expenditures in 2025, with a Q1 2026 run rate suggesting annualised spend approaching $30 billion.[13]
From an orbital infrastructure underwriting perspective, the AI segment is not additive. It is dilutive to every metric an infrastructure investor cares about: cash flow visibility, capex predictability, asset longevity, revenue contractability. It introduces AI competitive dynamics, hyperscaler platform risk, and a capital intensity trajectory that has no equivalent in the conventional infrastructure asset class. An infrastructure allocator applying standard criteria to the consolidated entity would properly exclude the AI segment entirely and value it as a separate, high-beta growth option, which means the infrastructure case for SpaceX reduces to Starlink plus Starshield plus launch, with Starship as the option value. That is a defensible position strategically. It is not a $1.75 trillion position on current fundamentals.
The important question is whether this structure was deliberate. The merger combines a cash-generative satellite business with a cash-consuming AI business under a single governance structure, priced to public markets as a unified entity. Whether that is brilliant vertical integration or a cross-subsidy that obscures the true cost of AI ambition depends on which assumptions you apply to xAI's trajectory. That is a venture question, not an infrastructure question.
Governance as a Structural Risk Premium
Infrastructure investors do not typically lead with governance as a primary underwriting concern because the assets do the work: regulated utilities, contracted toll roads, and concession assets are structurally protected against governance failure by the frameworks that create their cash flows. SpaceX offers the opposite bundle, concentrated founder control with almost none of the contractual or regulatory protections that conventional infrastructure provides.
Musk will retain approximately 85.1% of voting power through a dual-class share structure in which Class B shares carry ten votes each.[14][15] He simultaneously serves as chairman, CEO, and chief technology officer, can appoint a majority of the board, and cannot be removed as CEO without his own consent.[15] SpaceX has claimed "controlled company" status under Nasdaq rules, exempting it from the requirement that a majority of directors be independent.[15] Public investors waive rights to jury trials and class actions.[15]
The New York State Comptroller's investor coalition has formally raised these concerns with the SEC, citing conflicts of interest, compensation structures approved without independent committee process, and a pattern of governance that forecloses the ordinary mechanisms by which shareholders would enforce accountability.[16] AkademikerPension, the Danish pension fund with $25 billion in assets, has already blacklisted SpaceX on governance grounds.[17]
For infrastructure allocations specifically, this matters more than it might in a venture portfolio. Infrastructure mandates are typically justified to LPs on the basis of governance discipline, contractual clarity, downside protection, and leverage visibility. SpaceX offers strategic scarcity and exceptional founder execution. It does not offer the governance architecture that most infrastructure LP mandates require.
What This Means for Investors
The SpaceX IPO presents mid-sized infrastructure investors with a clean but uncomfortable framing problem. The orbital infrastructure label is credible for Starshield and credible in part for Starlink enterprise, maritime, and resilience applications. It is not credible for Starlink consumer broadband on asset-life grounds, for Starship on current-cash-flow grounds, or for the AI segment on any infrastructure-relevant grounds.
The practical implication is that most dedicated infrastructure mandates will either pass on the IPO or hold a position sized as an out-of-benchmark option, not as a core holding. The governance structure compounds this: even for investors who can justify the orbital infrastructure framing, the lack of minority protections, the related-party complexity, and the single-point-of-founder-failure risk sit outside what most infrastructure LP agreements contemplate.
The more useful lens for infrastructure practitioners is the second-order question: what does a dominant, vertically integrated orbital platform do to the competitive dynamics around adjacent assets? Ground station operators, terminal distribution networks, maritime digital infrastructure providers, and sovereign resilience network developers are all affected by Starlink's trajectory, often in ways that create investable positions with infrastructure-quality characteristics and without SpaceX's governance risk.
Closing View
SpaceX deserves a new analytical category. The vertical integration of launch access, satellite manufacturing, satellite broadband, and sovereign communications is genuinely unprecedented, and the strategic scarcity of the combined stack is real. None of that is in dispute.
What is in dispute is whether a new analytical category justifies 266x EBITDA, GAAP losses of nearly $5 billion in 2025, an AI segment burning $2.5 billion per quarter, and governance structures that most institutional capital cannot accommodate. The gap between strategic importance and infrastructure underwritability is precisely where the IPO is asking investors to make an act of faith.
For infrastructure practitioners, the right response is not FOMO. It is precision: identify the parts of the stack that pass the infrastructure test, Starshield, enterprise Starlink, launch logistics, understand why most of the rest does not, and size any position accordingly. The largest IPO in history will generate enough demand without infrastructure mandates stretching their criteria. The question is whether disciplined infrastructure investors should participate, and on what basis. The answer requires separating what SpaceX is today from what the valuation requires it to become.
Sources
Primary sources and market references cited above.