SpaceX Just Dropped Its First Public Earnings Report — And Wall Street Got It All Wrong

On August 5, 2026, SpaceX released its first quarterly report since going public.

Two numbers defined the narrative tension of this release: revenue of $7.8 billion, up 92% year on year and ahead of the $6.93 billion consensus; and an after-hours share price drop of more than 8%.

A beat and a sell‑off happened at the same time. This is not because the market "does not understand" the company. It is because SpaceX's valuation structure means that when a company's price already embeds a very distant future, even strong current numbers often fail to satisfy.

This article breaks down the report from three angles: business mix, financial logic, and valuation modelling. The central question is: what is already priced in, and what is not yet priced in?

I. Business Mix: Rockets Are Now Only One‑Tenth of the Story

Breaking down the $7.8 billion in quarterly revenue:

• Starlink (Connectivity): $4.291 billion, up 66% year on year, 55% of total revenue.

• AI business: $2.561 billion, up 247%, 33% of total.

• Space business (launch, government missions, Starship development): $0.962 billion, up 29%, 12% of total.

What used to be considered a "rocket company" – commercial launches, government missions, and Starship R&D – now accounts for only 12% of revenue. Starlink and AI together contribute 88%.

This fundamentally changes the valuation framework. SpaceX is no longer a launch‑services contractor waiting for a manifest of missions. It is a telecom operator, plus an AI compute provider, plus a long‑dated call option on orbital transportation that is still burning cash.

II. Starlink: The Only Segment Generating Positive Cash Flow

Starlink's Q2 numbers:

• Revenue: $4.291 billion.

• Operating income: $1.656 billion.

• Adjusted EBITDA: $2.597 billion.

• Capital expenditure: $1.367 billion.

Adjusted EBITDA minus capex gives roughly $1.23 billion. That is the positive cash flow generated by Starlink in a single quarter.

Subscribers grew from 6 million a year ago to 12 million. ARPU fell from $85 to $66, down 22% year on year. The market tends to focus on the ARPU decline, but a more important structural shift is hiding in the numbers.

Business and government customer revenue reached $1.806 billion, up 108% year on year, and now accounts for 42% of Starlink's total revenue.

This segment carries longer contract cycles and stronger pricing power. The cumulative multi‑year U.S. government contracts exceeding $6 billion, together with the Starshield military communications programme, are redefining Starlink from a "rural broadband provider" into a "critical‑infrastructure telecom platform." The ARPU decline in the consumer business reflects a strategic choice to enter more emerging markets. The rising share of B2B and government revenue is what will ultimately determine the long‑term margin ceiling.

III. AI Business: Explosive Revenue Growth, Cash Black Hole

AI segment Q2 data:

• Revenue: $2.561 billion, up 247% year on year.

• Adjusted EBITDA: $1.146 billion, a 44.7% margin.

• Operating loss: $1.257 billion.

• Depreciation and amortisation: $1.885 billion.

• Capital expenditure: $15.828 billion.

The 44.7% adjusted EBITDA margin looks striking. But this number adds back $1.885 billion of D&A and $516 million of stock‑based compensation.

In the GPU and data centre business, depreciation is not a "non‑cash" item in the usual sense. It represents rapid GPU obsolescence, server replacement cycles, and the economic compression of older assets by newer chips. Adjusted EBITDA profitability for the AI business does not mean positive free cash flow.

The company disclosed that it has signed cloud services agreements totalling $14.1 billion, of which about $1.6 billion was recognised as revenue this quarter from new contracts. Clients include Anthropic, Google, and Reflection AI. That means there is paid demand for AI compute.

On the earnings call, Musk said that the payback period for newly built AI compute capacity is less than one year. Subject to Rule 10b‑5, that statement is credible. The company also disclosed that in July it signed an additional $6.7 billion in six‑month cloud services revenue, and expects to reach a $100 billion annualised revenue run‑rate by December of this year.

One fact remains unchanged. To generate $2.561 billion in quarterly revenue, the AI business spent $15.828 billion in quarterly capital expenditure. This is still a "spend‑to‑grow" phase, not yet a "compound‑from‑profits" cycle.

IV. Space Business: The Book Cost of Starship

Space segment Q2:

• Revenue: $0.962 billion, up 29% year on year.

• Operating loss: $0.542 billion.

• Adjusted EBITDA: $0.205 billion negative.

• R&D expenditure: $1.076 billion, which already exceeds the segment's quarterly revenue.

For the first half of the year, space revenue fell slightly from $1.611 billion to $1.581 billion year on year, while operating losses widened from $0.439 billion to $1.204 billion.

Starship has not yet entered commercial operations. Its value currently exists as an option: deploying Starlink V3 satellites, enabling direct‑to‑device mobile services, and building orbital AI compute infrastructure. Every successful test flight reduces engineering risk, but it does not improve current financials.

At the end of July, SpaceX won a $1.6 billion launch order from the U.S. Space Force, on top of an earlier $6.5 billion national security satellite launch contract. These government contracts provide a revenue base, but they are not large enough to cover Starship's R&D burn.

V. Financing‑Driven Front‑Loaded Expansion

On the balance sheet:

• Cash and cash equivalents: $93.522 billion.

• Marketable securities: $6.487 billion.

• Debt and finance leases: approximately $39.364 billion.

• Net cash position: approximately $60.6 billion.

Where did this cash come from?

• Net IPO proceeds: approximately $85.7 billion.

• Bond issuance: $25 billion.

• Total cash flow from financing activities in the first half: $100.291 billion.

SpaceX's strategy is to use its public‑market valuation and credit capacity to secure a large pool of capital upfront, and then simultaneously accelerate investment across Starlink, AI data centres, and Starship.

This implies an optimistic view on capital payback periods. It also means that if technical or market timing slips in any of these businesses, the cash burn rate will expose the company to greater risk.

VI. Valuation: What Is Implied by the Current Price?

Using a market capitalisation of approximately $1.651 trillion and netting out cash gives an enterprise value of roughly $1.591 trillion. Including the roughly 3.6% dilution from the Cursor acquisition, pro‑forma shares outstanding are about 13.65 billion, implying a price of around $121 per share.

A reverse ten‑year discounted cash flow model yields the following implied assumptions:

• 2027 revenue growth: 62%.

• 2036 revenue: approximately $601 billion.

• 2036 FCFF margin: 32%.

• 2026–2036 revenue CAGR: 34.1%.

• Terminal value as a percentage of EV: approximately 89%.

• WACC: 9.5%.

• Terminal growth rate: 4.0%.

What the current price demands is this. Over the next decade, SpaceX must grow revenue from roughly $32 billion to about $600 billion. It must simultaneously transform from deep negative free cash flow to a 32% FCFF margin. And it must avoid any major financing crisis, regulatory setback, prolonged Starship delay, or collapse in AI asset returns.

Scenario analysis:

• Bear case: 2036 revenue $371 billion, 2036 FCFF margin 26%, fair value roughly $56.

• Base case: 2036 revenue $553 billion, 2036 FCFF margin 32%, fair value roughly $116.

• Bull case: 2036 revenue $684 billion, 2036 FCFF margin 35%, fair value roughly $200.

Probability‑weighted (25% bear, 55% base, 20% bull) fair value comes to about $121 per share. Incorporating the additional $6.7 billion July contract and the $100 billion annualised run‑rate guidance lifts the base‑case target to roughly $130.

Sensitivity analysis across WACC of 9.0%–10.0% and terminal growth of 3.5%–4.5% gives a fair‑value range of roughly $96–144.

With the U.S. 10‑year Treasury yield at about 4.64%, a 9.5% WACC for a company still facing technology, capex, and execution risk is not conservative.

VII. Three Questions Awaiting Answers

This quarter's report shows three distinct dynamics.

Starlink has proven itself as a cash‑generating telecom asset. The rising share of B2B and government revenue is a more meaningful metric than ARPU movements. AI revenue growth is real and backed by paying customers, but the gap between capital spending and accounting profits means it remains in the investment phase. Starship's book cost is accumulating. Engineering risk is declining. The commercial timeline remains uncertain.

Current valuations embed highly optimistic future expectations. The price is essentially a bet on a three‑part thesis: global Starlink dominance, AI compute infrastructure, and Starship‑enabled orbital transport.

Over the next 12 to 24 months, three specific questions will be tested.

First, as Starlink subscribers continue to grow beyond 12 million, can ARPU stabilise around $66, and will the B2B and government revenue share break above 50%?

Second, is the AI compute payback period truly less than one year, as management has suggested? That will show up in gross margin and operating margin trends over the next two to three quarters.

Third, when will Starship move from test flights to commercial deployment, and can it begin launching Starlink V3 satellites before 2027?

These questions matter more to SpaceX's long‑term value than whether this quarter's revenue beat consensus by a few hundred million dollars.