SpaceX (SPCX): First Earnings Ever. Revenue +92%, Capex 235% Of Sales. Is SPCX Stock a Buy?

Space Exploration Technologies Corp. (SPCX) Q2 2026 — Reported August 4, 2026 AFTER the US close - the first-ever quarterly report as a public company, for the calendar quarter ended June 30, 2026. Revenue $7,814M, +92% YoY, versus roughly $6.93B expected. Diluted loss per share $0.09 versus a $0.26 loss modelled. Net loss $541M, narrowed from $1,008M. Adjusted EBITDA $3,538M, +191%. All three segments beat: Connectivity $4,291M (+66%), AI $2,561M (+247%), Space $962M (+29%). But capital expenditure was $18,369M in the quarter - 235% of revenue, up more than six-fold from $2,825M, and about $5B above the ~$13.2B the Street had modelled. SPCX closed the regular session at $125.33, up 9.43%, then fell in extended trading. IPO priced at $135.00 on June 11 and began trading June 12. Closing high $201.80 (Jun 16); all-time closing low $108.37 (Jul 31). SpaceX beat on every line Wall Street actually models - and the stock fell anyway, because the number that moved it is not on the income statement. Capital expenditure was $18.37 BILLION in one quarter against $7.81 billion of revenue: $2.35 of capital spending for every $1 of sales, up more than six-fold year over year, with $15.83B of it in AI alone - 618% of the AI segment's own revenue. First-half operating cash flow was $3.47B against $28.48B of capex, so free cash flow for the half was NEGATIVE $25.0 billion. Three things almost nobody has connected. FIRST, only ONE of the three segments makes money: Connectivity earned $1,656M of operating income at a 38.6% margin, and it pays for $1,799M of losses in Space and AI combined. SECOND, AI's celebrated first-ever positive Adjusted EBITDA of $1,146M is a depreciation artifact - $23.55B of H1 AI capex adds roughly $4.7B a year of depreciation on a five-year life, more than the entire figure being celebrated. THIRD, Starlink subscribers DOUBLED to 12.0 million while ARPU fell 22%, from $85 to $66; consumer revenue rose only 44% against 100% subscriber growth, so revenue per subscriber fell about 28%. And two trading days after this print, on August 6, the first lock-up tranche frees roughly 911 million shares - 143% of the entire IPO float. THE CALL: AVOID (1/5, A REMARKABLE COMPANY AT AN IMPOSSIBLE PRICE) — base-case value ~$32.0 vs ~$125.33 today. KEY METRICS: - CALL: AVOID 1/5. Fair value ~$32 against a $125.33 close - 74% below. This is NOT a call against SpaceX the company; the quarter was genuinely strong and the engineering is unmatched. It is a call against this price. Method: a ten-year path-to-profitability DCF, because there is no free cash flow to discount. Base case: revenue compounds 25% a year for a decade, $30B in 2026 to $283B in 2035; capex plateaus near $100B; 2035 free cash flow $55B (a 19% margin). Discounted at 9.5% with 3% terminal growth: $846B terminal value, $374B discounted back nine years, less $21B for the negative interim cash flows, plus $61B net cash, over 13,176M shares = $31. Grid at 8.5/9.5/10.5%: bear $15/$12/$10, base $39/$31/$26, bull $86/$68/$55. A 30/50/20 weighting gives $33. EVERY cell is below the price - including the bull case at the friendliest discount rate we would defend. - REVERSE DCF - WHAT $125.33 REQUIRES: 13,176M shares (Class A 7,607M + Class B 5,569M off the June 30 balance sheet) x $125.33 = $1,651B of equity; less $61B of net cash ($100.0B cash and marketable securities against $39.4B of debt and finance leases) = $1,591B of enterprise value. At 9.5% and 3% terminal growth that needs roughly $256 BILLION of free cash flow in 2036. At a generous 25% free-cash margin that is $1.02 TRILLION of revenue - a level no company has ever reached - and a 42% compound growth rate for ten straight years off today's $31B annualised run-rate. On that run-rate the stock trades at 51x sales and 112x even the flattering annualised Adjusted EBITDA. - THE CAPEX SHOCK, WHICH IS THE WHOLE STORY: $18,369M of capital expenditure in one quarter against $7,814M of revenue = 235% of revenue, up from $2,825M a year ago (more than six-fold) and $10,107M last quarter. By segment: AI $15,828M, Connectivity $1,367M, Space $1,174M. AI capex alone was 618% of AI segment revenue. Consensus had modelled roughly $13.2B, so the miss was about $5B. First half: operating cash flow $3,466M against capex $28,476M, so first-half free cash flow was NEGATIVE $25.0B. Net property, plant and equipment went from $42.6B at the year end to $65.7B in six months. Nameplate compute reached 1.4 GW, up from 0.4 GW a year ago. - ONLY ONE SEGMENT MAKES MONEY: Connectivity (Starlink) revenue $4,291M, +66%, with income from operations of $1,656M - a 38.6% operating margin and Adjusted EBITDA of $2,597M. Space revenue $962M, +29%, but a LOSS from operations of $542M, because Starship R&D alone was $1,076M in the quarter. AI revenue $2,561M, +247%, but a LOSS from operations of $1,257M. So Starlink's $1,656M of profit funds $1,799M of losses elsewhere, and the difference is the $143M consolidated operating loss on the headline. Total R&D was $3,548M - 45.4% of revenue, up 81% year over year. - THE ADJUSTED EBITDA PROBLEM: the $3,538M headline excludes depreciation, amortisation, share-based compensation, interest and tax - in other words, it excludes the cost of the one thing this company is spending all of its money on. D&A was $2,848M in the quarter against $1,526M a year ago, up 87%, and it is about to accelerate: $23,551M of H1 AI capex on a five-year life adds roughly $4.7B a year of depreciation once in service. That is MORE than AI's celebrated first-ever positive Adjusted EBITDA of $1,146M. AI's GAAP loss from operations was still $1,257M. Segment Adjusted EBITDA: Connectivity $2,597M, AI $1,146M, Space negative $205M. - THE STARLINK NUMBER NOBODY QUOTED: 12.0 million subscribers, double a year ago and up 1.7 million sequentially - but ARPU is $66 a month against $85 a year ago, a 22% DECLINE, reported flatly with no commentary. Consumer revenue rose 44% ($1,721M to $2,485M) against 100% subscriber growth, so revenue per subscriber fell roughly 28%. Growth is being bought with price. Enterprise and government revenue, up 108% to $1,806M, is now doing the heavy lifting - helped by over $6B of multi-year Starshield contracts and FCC approval of the EchoStar spectrum transfer. - THE LOCK-UP, TWO DAYS AFTER THIS PRINT: SpaceX used a STAGGERED lock-up, not a single 180-day one, and the first tranche is triggered by this earnings report. On August 6 insiders may sell up to roughly 911 million shares - about $114B at the $125.33 close, 143% of the entire 638,888,888-share IPO float, and roughly eight full days of the ~111.5M average daily volume. A further tranche of about 28% follows the Q3 print, with the full period ending December 8. Elon Musk's shares carry a 366-day lock-up into June 2027. The stock is already 46% below its June 16 closing high of $201.80 and 7% below the $135.00 IPO price. - BALANCE SHEET AND CAPITAL STRUCTURE: cash and equivalents $93,522M plus marketable securities $6,487M = $100.0B, against debt and finance leases of $39,364M ($2,525M current, $36,839M long-term) - net cash of $60.6B. Backlog $47.5B and $14.1B of newly contracted Cloud Services Agreements. Total assets $192,770M; shareholders' equity $127,224M; accumulated deficit $41,852M. Note the related-party exposure: $13.3B of the debt is related-party, and $327M of the $629M quarterly interest expense is related-party. The IPO closed June 15 (638,888,888 Class A shares, ~$85.7B net proceeds) and a $25B inaugural investment-grade bond issue closed June 26 across five tranches at 5.35%-6.65%, a 5.855% weighted average. - WHY WE DIFFER FROM WALL STREET: consensus is Buy with an average twelve-month target of $214 (range $115 to $401), about 71% above the $125.33 close. Morgan Stanley initiated Overweight at $300; Goldman Sachs at $205; every initiation since the IPO has been buy-equivalent. We are at ~$32 and we say so on the slide. The gap is method, not facts: the sell-side is valuing SpaceX on a 2030-2035 sum of the parts with venture-scale multiples attached, while we discount cash - and on that basis the capital spending is not an asset, it is a decade of deferred free cash flow, and it is still accelerating. We might be the ones who are wrong, and we say that on the slide too. SOURCING NOTE: every figure above is read off the 8-K Item 2.02 EX-99.1 filed 2026-08-04 (accession 0001628280-26-052515), grepped for the company name and for 'June 30, 2026' before any number was taken. CIK 0001181412 was verified independently on EDGAR. The share count is off the consolidated balance sheet, not FMP. What to watch: Bullish (what would change our mind fastest): a quarter in which AI capital expenditure FLATTENS while AI revenue keeps compounding. That single data point would make this model far too bearish and we would say so. Also constructive: Starship reaching operational reusability and the Space segment turning profitable. Bearish: another quarter where capex grows faster than revenue. Level: we would start looking under $60, roughly twice our base-case value. Also on YouTube: @ChargedAlpha DISCLAIMER: For informational and educational purposes only. Not financial advice. Do your own research before any investment decision.