The "burning money-to-monetization" cycle is validated for the first time! SI business supports SpaceX (SPCX.US) in achieving its first quarterly profit since listing—Elon Musk presses the "merge" button

According to Zhitong Finance APP, after the US stock market closed on October 20 (Tuesday) Eastern Time, SpaceX (SPCX.US) released its second financial report since going public. If the Q2 report three months ago—which saw "revenue beat expectations but the stock price plummet 11%"—had demonstrated the burning power of artificial intelligence (AI), this time Musk provided three answers in one night: The company can make money—with a net profit of $4.4 billion, the first quarterly profit since going public; AI can "take the lead"—revenue reached $7.56 billion, soaring ninefold year-on-year, surpassing the combined revenue of Starlink and rocket launches; and the merger with Tesla (TSLA.US) is real—Musk personally confirmed on the earnings call that SpaceX and Tesla have set up a committee to discuss the merger.

As of press time, SpaceX stock rose more than 7% after hours on Tuesday to around $159. Just two months ago, the price was struggling at the $105 mark, now it is once again approaching the new highs above its IPO price of $135.

“Crushing” earnings breakdown: Three segments, three roles

Let's look at the overall picture. The financial report shows that SpaceX's Q3 revenue reached $14.575 billion, far exceeding the consensus analyst expectation of $12.72 billion, up 161% year-on-year from $5.576 billion, and about 87% higher compared to Q2's $7.814 billion.

Q3 operating profit reached $2.791 billion, compared to an operating loss of $273 million in the same period last year and an operating loss of $143 million in Q2; net profit in Q3 reached $4.404 billion, far exceeding the consensus analyst expectation of $1.57 to $1.88 billion, representing a significant turnaround from a net loss of $1.054 billion the prior year, returning to profitability.

According to US GAAP, SpaceX's Q3 diluted earnings per share was $0.35, also far above the consensus analyst expectation of $0.15, compared to a per share loss of $0.36 the previous year.

On a non-GAAP basis, SpaceX’s Q3 adjusted EBITDA reached $8.809 billion, up 306% from $2.169 billion in the same period last year and about 149% above Q2’s $3.538 billion—also beating the consensus expectation of $6.59 billion.

Capital expenditures—the market's focus—came in at $19.59 billion in Q3, in line with analyst expectations of $19.6 billion.

Super Intelligence (SI): In one quarter, from a $1.26 billion loss to a $430 million profit

The most significant structural change in this report is the explosive performance of the Super Intelligence (SI) segment—revenue of $7.56 billion, up 195% quarter-on-quarter and up 830% year-on-year, about 20% ahead of market expectations, accounting for about 52% of total company revenue. A company once famous for building rockets now gets more than half its revenue from AI computing power and model services. At an internal meeting in August, Musk had claimed that AI revenue "may surpass the combined total of all other business revenues in September." Now, the financial report validates this prediction.

More crucial is the profitability. The SI segment posted $431 million in operating profit, compared to a $1.257 billion operational loss in Q2—turning positive within one quarter—even after including $1.2 billion in stock-based compensation from the Cursor acquisition. Adjusted EBITDA for this business was $4.95 billion, about 49% above expectations.

Growth comes from three sources: new cloud service deals (with clients including Anthropic, Google, etc.); consolidation of Cursor after its August 14 completion (contributing $608 million in revenue by quarter’s end); as well as increased usage of Grok by both enterprise and individual clients, with the company launching the all-weather agent system Grok Bot. On the computing side, nominal computing power expanded from 1.4 GW at the end of Q2 to 1.7 GW (compared to just 0.6 GW a year ago), with ongoing construction of the COLOSSUS II cluster.

Starlink (Connectivity Services): The real “profit cow”

AI provides imagination for revenue growth, but the real money maker remains Starlink. The connectivity segment reported $5.25 billion in revenue, up 69% year-on-year; operating profit nearly doubled to $2.37 billion—media commented that virtually all of SpaceX’s profits still come from Starlink. Adjusted EBITDA for the segment rose 82% to $3.5 billion.

Consumer subscriptions reached 13.6 million, up 1.6 million quarter-on-quarter; monthly ARPU rose 5% quarter-on-quarter to $69, as May’s price hikes took effect. Aviation connectivity keeps expanding, with more than 3,800 aircraft equipped with Starlink terminals. The company also announced signing an agreement to acquire a nationwide 800MHz low-frequency spectrum license (expected to close in the first half of 2027, pending FCC and other approvals), paving the way for Starlink’s direct-to-cell business as a second growth curve.

Space Business: Small but steady “strategic base”

The launch segment saw revenue of $1.8 billion, up 84% quarter-on-quarter and 6% year-on-year, with 12 customer launches completed in Q3 (including NASA’s Nancy Grace Roman Space Telescope mission). The 14th Starship test flight achieved first orbit, deploying 26 Starlink V3 satellites, with each mission’s downlink bandwidth about ten times that of a standard Falcon 9 Starlink mission.

Order bookings continue to grow. 3 new NASA crewed missions, 18 national security launches, 3 commercial Starfall missions, and a contract for sending payloads to the lunar surface via Starship. The company also announced plans to build a Starbase in Louisiana, aiming for thousands of Starship launches annually, with the first launch set for 2029. The 15th Starship flight, as early as early November, will attempt for the first time to "catch" the returning upper stage using the launch tower—if successful, launch costs will drop by another order of magnitude. Musk said on the call: one launch per week in H1 next year, at least two per week in H2.

This is not just a regular beat, but the first verification of a “burn-cash-to-monetize” closed loop

Prior to the financial report, the market's main doubt about SpaceX was that the $18.37 billion in Q2 capex still resulted in a loss—at the time, the stock price plunged 11% after hours. The key reason why the Q3 results are viewed as a watershed by most institutions is that it simultaneously met three criteria for the first time:

1. AI investment is starting to "pay off." The SI segment went from a single-quarter loss of $1.257 billion to a profit of $431 million, proving that newly signed computing contracts (the CFO had disclosed in September that new agreements could bring an additional $1.1 billion in monthly revenue) are moving into the revenue recognition phase, rather than just being projects on the balance sheet.

2. Operating leverage is emerging. Revenue rose 161% year-on-year, and the company swung from a $143 million operating loss to a $2.79 billion operating profit, with a net profit margin of roughly 30%. The CFO's description of “a clear inflection point in Super Intelligence” is essentially a result of rapid income growth diluting fixed costs.

3. The valuation anchor has a new benchmark. Before the report, UBS maintained a Buy rating with a $210 price target (its SI revenue estimate of $7.6 billion was virtually a perfect match to the actual result); Goldman Sachs raised its price target to $230 and labeled AI as “the next growth engine;” Evercore ISI set a target of $275. Morgan Stanley’s quant framework: every additional 1 GW of computing power could add about $27 to the stock price—according to Goldman's forecast path of “2.4 GW by end-2026, 7.0 GW by end-2027,” the expansion of computing power itself is a directly convertible valuation curve.

However, the bears remain unconvinced by this report. SpaceX’s Q3 capex still exceeded revenue, and the free cash flow gap was not closed by the return to profitability—expansion still relies on existing cash and external financing. Although cash and marketable securities stood at $89.86 billion at the end of the quarter, providing a sizable buffer, refinancing costs and terms will be subject to heavy scrutiny in a high-interest rate environment.

Meanwhile, signs of slowing Starlink growth have appeared—consumer net adds of 1.6 million were below the roughly 2.29 million expected by analysts; ARPU, despite rebounding to $69, was still well below $85 a year ago. This suggests the “profit cow’s” “milk production rate” may be approaching a plateau, and the dilutive effect of expanding into lower-price markets continues.

The merger committee is confirmed! How the earnings report became an “invisible pricing day”

Before the financial report, Barclays analyst Dan Levy’s scenario was already circulating on Wall Street: If SpaceX and Tesla merge, it would most likely be an all-stock "reverse merger"—with the larger SpaceX swallowing Tesla at a premium. Referring to valuations of $2.1 trillion for SpaceX and $1.5 trillion for Tesla, Tesla shareholders would likely receive 0.7 SpaceX shares for each share they hold. Musk would hold about 37% of the merged company’s shares and control 56% of voting rights.

The timeframe is also tight: the merger announcement must be made between late October and early November, to make the window for a Tesla shareholder vote at the annual meeting on December 6. Well-known tech analyst Ives had previously estimated an 80% probability of a 2027 merger and interpreted the mid-October release of the new Roadster, featuring SpaceX thrusters and pressure vessels, as “bringing the relationship between the two companies into the open.”

On this logic, Tuesday's earnings report essentially sets the share exchange benchmark for Tesla shareholders—the stronger the report, the more convincing the 0.7 share ratio. And on the call, Musk pushed the rumor further: SpaceX and Tesla “have set up a committee to discuss the merger,” adding that the merger “has many advantages, but it's not set in stone and requires shareholder support.” From his July comment—widely read by the market as “neither confirming nor denying” because “we can’t discuss the merger on the earnings call”—to now confirming the committee, this potential deal has jumped from rumor to official agenda. The over-7% after-hours gain is, to some extent, the market’s direct pricing of “increased merger probability.”

Summary

In summary, after Tuesday, SpaceX can no longer be priced using any single framework: Looked at as a space company, its launch monopoly and Starship progress are materializing; as a telecom operator, Starlink’s price hikes and spectrum acquisition open a second growth curve; as an AI company, it has a full stack from electricity and computing power to models and tools, with revenue growth outpacing all public comparables. Hovering above all this is the fourth pricing layer—the Tesla merger—which has already moved from rumor to formal committee agenda.

Three months ago, the market sold off SpaceX because of its $15.8 billion AI investment; three months later, the market is chasing it for its $7.56 billion AI revenue. Musk hasn’t changed the pace of his burn rate; he’s just demonstrated that burning cash can become revenue. For the SpaceX share price, the next verification points are clear: the 15th Starship test flight in early November, the Tesla shareholder meeting on December 6, and whether computing power reaches 2 GW and ARR $100 billion by year-end. The biggest significance of this earnings report is that, for the first time, all three countdowns now actually look credible.