Key Highlights from the SpaceX IPO:

  1. Starlink Is the Real Money-Maker Starlink's Connectivity segment generated $11.4 billion in Revenue and $7.2 billion in Adjusted EBITDA in 2025 — at a 63% Margin that rivals the world's best toll-road businesses. Subscribers nearly doubled from 5 million to 10.3 million in just 15 months, and the service operates across 30+ countries with no geographical ceiling on growth.
  2. The Rocket Business Is Quietly Revolutionary Falcon 9 has slashed the cost of reaching orbit from $18,500 per kilogram historically to a fraction of that today, conducting 170 launches in 2025 — roughly one every two days. The next-generation Starship rocket, still in flight testing, is designed to cut costs by 99% further, and is the single keystone on which every other segment's growth depends.
  3. The AI Segment Is Ambitious but Burning Cash xAI and X (Twitter), acquired in February 2026, generated $3.2 billion in revenue but posted a $6.4 billion operating loss in 2025, consuming $12.7 billion in Capital-expenditure/">Capital Expenditure to build the COLOSSUS data centres. Grok has 117 million monthly active AI users and a unique real-time data advantage via X, but orbital AI compute — the long-term vision — remains entirely untested.
  4. Governance Is Heavily Concentrated in One Person Elon Musk's Class B shares carry 10 votes each versus 1 vote for Class A shares sold in the IPO, giving him permanent majority control over all major decisions. SpaceX qualifies as a "controlled company," exempt from standard Nasdaq governance requirements, and carries no key-person Life insurance on Musk despite the entire strategy being built around his vision.
  5. You Cannot Buy Starlink Without the Rest The core tension of this IPO is that the proven, cash-generating Franchise — Starlink — comes bundled with a loss-making AI segment, $29.1 billion in total Debt, and a growth thesis entirely dependent on Starship achieving commercial operations at scale. Investors are effectively being asked to price three very different businesses as one, at one of the most watched IPO valuations in a generation.

 

There is a simple test every serious investor should apply before buying a single share of any company: does this business have a moat so wide and so deep that even if I fell asleep for ten years and woke up, it would still be the dominant player in its industry?

SpaceX, which filed its S-1 prospectus with the US Securities and Exchange Commission on May 20, 2026, and proposes to list under the ticker SPCX on Nasdaq, passes that test in at least one of its three businesses — and spectacularly so. In the other two, the answer is considerably more complicated. This article will walk you through all three, in plain language, so you can form your own view before the most anticipated IPO of the decade prices.

First, Understand What SpaceX Actually Is

Most people think SpaceX is a rocket company. It is not — or at least, it is not only a rocket company anymore. Think of it instead as three businesses sharing one Balance Sheet, one culture, and one founder.

Business One is the original rocket company — Falcon 9, Falcon Heavy, and the Dragon spacecraft. This is what launched SpaceX to global fame. Call it the Space segment.

Business Two is Starlink — a constellation of approximately 9,600 satellites orbiting the Earth, beaming internet directly to homes, ships, planes, farms, and now mobile phones. This is the Connectivity segment, and it is the financial engine of the entire company.

Business Three is xAI — the artificial intelligence company that Elon Musk founded in 2023 and which SpaceX acquired in February 2026, bringing along X (formerly Twitter) in the same package. This is the AI segment, and it is, for now, the financial black hole of the company.

These three businesses have been folded into one S-1 and are being offered to you as a single stock. Understanding each one separately is the only honest way to evaluate this IPO.

Business One: The Rocket Company (Space Segment)

The Simple Version

Imagine you need to send a package from London to New York. Historically, only one airline operated that route and charged whatever it wanted. Then a new airline came along, figured out how to reuse its planes instead of scrapping them after every flight, and cut the ticket price by 85%. Every other airline either had to match the price or go out of business. That is exactly what SpaceX did to the rocket industry.

The Numbers

In 2025, the Space segment generated $4.1 billion in revenue from launching satellites, cargo, and astronauts for commercial customers, governments, and NASA. The segment conducted 170 launches in 2025 — roughly one every two days. That launch cadence is not just impressive; it is the moat. No competitor operates anything close to it.

The cost advantage is staggering and worth understanding concretely. According to NASA, the historical average cost to put one kilogram of cargo into low earth orbit was $18,500. Falcon 9 brought that down to $2,700 per kilogram when it launched in 2010. Today, with boosters being reflown up to 34 times before retirement, the cost has fallen further still. Starship — SpaceX's next-generation rocket — is designed to reduce this cost by 99% or more relative to that $18,500 historical benchmark. That is not an incremental improvement. That is a civilisational shift in what becomes economically possible in space.

The Complication

The Space segment is not profitable right now in the traditional sense. It reported a $657 million operating loss in 2025, almost entirely because the segment is funding $3 billion per year in Starship Research and Development. Think of it this way: the rocket business is profitable on its own, but it is also writing the cheque for the rocket that will make everything else possible. Starship is the keystone of the entire SpaceX growth strategy — without it, the Connectivity and AI segments cannot scale as planned.

As of the S-1 filing, Starship has completed 11 flight tests and has not yet delivered a commercial payload to orbit. That is the single most important fact in the entire prospectus. Everything else — the subscriber forecasts, the AI ambitions, the orbital compute plans — depends on Starship working at scale.

Business Two: The Satellite Internet Company (Connectivity Segment)

The Simple Version

Picture a world where your internet does not depend on Airtel laying cables in your street, or Jio building a tower on the hill outside your village. Instead, your internet comes directly from satellites orbiting 550 kilometres above your head, beaming a signal to a pizza-box-sized dish on your rooftop. That signal works in the middle of the Sahara, on a ship in the Pacific, on a farm in rural Australia, or in a bunker in Ukraine. That is Starlink. And that is a business with almost no geographical ceiling on its addressable market.

The Numbers

This is where the story gets genuinely exciting. Starlink's Connectivity segment is the financial heart of SpaceX, and it has grown at a pace that most consumer technology companies would envy.

  • Subscribers: 2.3 million at end-2023 → 4.4 million at end-2024 → 10.3 million as of March 31, 2026. That is a near-doubling in subscribers in the space of 15 months.
  • Revenue: The segment generated $11.4 billion in 2025, up 50% from $7.6 billion the year before.
  • Profitability: Segment Operating Income was $4.4 billion in 2025. Adjusted EBITDA — a measure of cash Earnings before accounting adjustments — was $7.2 billion, implying a margin of approximately 63%. For context, this is the kind of margin you see in the best toll-road businesses in the world. Once the satellites are up and the ground infrastructure is built, adding the next subscriber costs almost nothing.

The ARPU Story

One number that looks like bad news but is actually good news is the monthly revenue per subscriber (ARPU). In 2023, it was $99 per month. By Q1 2026, it had fallen to $66 per month. At first glance, this looks like a business under pricing pressure. But look more carefully: the reason ARPU is falling is that SpaceX is deliberately moving into lower-income markets — rural India, sub-Saharan Africa, Southeast Asia — where the price point needs to be lower to unlock mass adoption. The math still works because in these markets, Starlink is often the only internet option. No cable company. No fibre. No competition. A Monopoly by geography is still a monopoly.

The Mobile Phone Opportunity

Starlink is now doing something that would have seemed like science fiction five years ago: sending internet signals directly to ordinary smartphones with no special dish needed. SpaceX has partnered with over 30 mobile network operators across six continents — including T-Mobile in the US, Rogers in Canada, Telstra and Optus in Australia, KDDI in Japan, and Kyivstar in Ukraine — to offer this service. The market opportunity here is enormous: there are approximately 1.9 billion people in the coverage zones of these MNO partnerships. Connecting even a fraction of them represents a multi-billion-dollar revenue stream that does not yet appear in the financials.

To support this expansion, SpaceX has agreed to acquire EchoStar's spectrum licences — the radio frequencies needed to operate satellite-to-mobile services — for $19.6 billion in cash and Equity. This is a large capital outlay, but it buys SpaceX direct ownership of the spectrum it needs, rather than perpetual dependence on MNO spectrum leasing arrangements.

Business Three: The AI Company (AI Segment)

The Simple Version

Imagine the world's most ambitious technology startup — one that built the world's largest AI Training supercomputer in 122 days, is developing a frontier AI model that reads every post on the world's largest real-time news platform, and has folded in a Social Media company with hundreds of millions of users — and then imagine that startup being injected into SpaceX's balance sheet all at once. That is what happened in February 2026 when SpaceX acquired xAI, which itself owned X (Twitter). The AI segment is new, it is enormous, and it is burning cash at a rate that demands careful scrutiny.

The Numbers

The AI segment generated $3.2 billion in revenue in 2025 from three sources: Grok AI subscriptions and API access, X platform Advertising and premium memberships, and compute leasing to external customers. However, it reported an operating loss of $6.4 billion in the same period, and consumed $12.7 billion in capital expenditure — almost entirely to build COLOSSUS and COLOSSUS II, the AI data centres in Memphis, Tennessee.

To give you a sense of the scale of this infrastructure: COLOSSUS houses approximately 100,000 NVIDIA H100 processors and COLOSSUS II houses approximately 110,000 NVIDIA GB200 and GB300 processors — the most advanced AI chips available. Together they provide roughly 1 gigawatt of compute power. For perspective, the company built COLOSSUS's first cluster in 122 days and COLOSSUS II's first cluster in 91 days — tasks that typically take the industry approximately two years for equivalent scale.

Grok, the AI model at the heart of this segment, had 117 million monthly active users using its AI features as of March 2026. Its key Competitive Advantage is its deep integration with X, which gives it access to real-time, unfiltered human conversation at a scale no other AI model can match. OpenAI has GPT-4 trained on the internet as it existed at a point in time. Grok is connected to the internet as it exists right now, every minute of every day.

The Honest Risk

The AI segment is the part of this S-1 that requires the most caution. SpaceX acquired xAI from a company that Musk also controlled — in other words, it was not an arm's-length transaction between independent buyers and sellers. Public market investors had no say in what was paid, what was agreed, and what liabilities were assumed. The financial terms were set between Musk and Musk, in effect. This is not illegal, but it is a governance arrangement that demands scrutiny.

Furthermore, the AI segment's long-term ambition — deploying AI computing hardware on Starlink satellites in orbit, creating the world's first orbital AI compute network — is a concept that has never been tested, attempted, or even closely approximated by any company in history. The S-1 explicitly acknowledges this. It is an extraordinarily ambitious vision. It may transform what AI infrastructure looks like in the 2030s. But it is also pure option value at this stage, not a business.

The Governance Question Every Investor Must Ask

Here is the fact that no institutional investor should skip over. After this IPO, Elon Musk will hold Class B shares that carry 10 votes per share. Class A shares — the ones you buy in the IPO — carry 1 vote per share. Musk will retain majority voting control over every major decision the company makes: board composition, acquisitions, executive pay, capital allocation. SpaceX will qualify as a controlled company under Nasdaq rules, which means it is exempt from several standard governance requirements.

The S-1 also states plainly that Musk and his affiliates are not restricted from competing with SpaceX, need not offer SpaceX first access to business opportunities they discover, and continue to run Tesla, Neuralink, and The Boring Company simultaneously. There is no key-person life insurance on Musk.

This is not a red flag in itself — many great businesses, including Berkshire Hathaway, were built under concentrated founder control. But it does mean that as a Class A Shareholder, your protections depend almost entirely on Musk's continued alignment with minority shareholder interests, rather than on structural governance mechanisms.

The Financial Reality in One Paragraph

SpaceX generated $18.7 billion in revenue in 2025 and $6.6 billion in Adjusted EBITDA. It had Operating Cash Flow of $6.8 billion. But it spent $20.7 billion in capital expenditure — leaving a free cash flow Deficit of approximately $13.9 billion. It carries $29.1 billion in total debt. The Connectivity segment is highly profitable and self-funding. The AI segment is consuming cash at a rate that requires continuous external financing. The Space segment is funding the Starship program that everything depends on.

The Franchise Test: What Does the Evidence Say?

The core question to ask is: is there a clean, compounding franchise here — one with high barriers to entry, pricing power, and capital efficiency — that can grow earnings at 15–20% per year for a decade?

The Connectivity segment — Starlink — is exactly that franchise. It has a moat that took 12 years and billions of dollars in government contracts and private capital to build, and that no competitor has come close to replicating. It is already generating $7 billion in annual EBITDA and growing at 50% per year. If Starlink were a standalone listed company, it would be one of the most compelling infrastructure businesses in the world.

The complication is that you cannot buy Starlink alone. You buy it bundled with an AI segment still burning $6 billion per year, a governance structure that concentrates all decision-making in one individual, and a $29 billion debt load that will need to be serviced and refinanced through the Business Cycle.

The question SpaceX asks every investor to answer is this: do you believe Elon Musk will execute on Starship, orbital AI, and the V3 constellation — and if he does, is the price being asked a fair one for what that execution is worth?

That is not an easy question. But it is the right one.

This article is prepared for informational purposes only, based on SpaceX's S-1 Registration Statement filed with the SEC on May 20, 2026. It does not constitute Investment advice or a recommendation to buy or sell securities. Investing in IPOs involves significant risks, including the risk of loss of principal. Markets are highly volatile. Readers should conduct their own Due Diligence and consult a qualified financial adviser before making investment decisions. Kalkine Research does not hold a position in SpaceX as of the date of publication.