Everyone watched 911 million shares come free today. The number that actually matters is the 456 million that didn't — because the price condition failed. That failure is a tradable signal, and it tells you exactly what SpaceX has to do between now and December.
SpaceX just printed a 92% revenue quarter and fell 13.6%. The market decided the story is capex. It isn't. The story is float. SPCX is a $1.45 trillion company with a $70 billion tradable stock — a 4.85% free float — and every price you have seen since June, in both directions, is an artifact of that. The $2.97 trillion peak on June 16 was fake. The 52% drawdown that followed was the correction of a scarcity premium, not a verdict on the business. Today the float more than doubled and the stock went up, because a 456-million-share tranche tied to a price milestone failed to trigger and pulled roughly $50 billion of expected supply off the table. My highest-conviction finding: shorts are at 25.84% of float, up 48% in a month, with 1.34 days to cover, and their catalyst just passed without a breakdown. That is a squeeze, not a short. But it is a rental, not a marriage — because the same lockup schedule that spared the stock today dumps 1.76 billion more shares on it between November and December, and SpaceX needs that supply to arrive so it can fund a capital programme no equity market has ever been asked to fund.
Forget the rocket company framing. Per the FY25 segment disclosure, SpaceX is a satellite broadband utility with a launch division attached and an AI compute business growing faster than either:
| Segment (FY2025) | Revenue | % of total | What it really is |
|---|---|---|---|
| Consumer (Starlink) | $7.21B | 38.6% | Subscription broadband annuity |
| Enterprise & Government | $4.18B | 22.4% | Sticky, contracted, high-margin |
| Launch Services | $2.58B | 13.8% | Near-monopoly, low growth |
| Advertising | $1.84B | 9.9% | The segment nobody models |
| Launch & Development | $1.51B | 8.1% | Cost-plus government programmes |
| AI Solutions & Infrastructure | $1.36B | 7.3% | The entire bull case |
| Total | $18.67B | 100% | +33.2% YoY |
Launch — the thing the company is famous for — is 22% of revenue and barely growing. 61% of revenue is Starlink. And the 7.3% AI sliver grew 247% year-over-year to $2.56 billion in Q2 alone, which is to say the smallest segment in FY25 is now running at roughly $10 billion annualised and is the reason the stock carries a trillion-dollar handle.
Management's prepared remarks on the Q2 call opened with the headline numbers: "Revenues of $7.8 billion, up 92% from $4.1 billion; a quarterly net loss of $541 million, an improvement of $467 million." Revenue beat consensus by $980 million; GAAP EPS of $0.09 beat by $0.20. Backlog hit $47.5 billion.
The stock fell 13.6% to $108.27.
Lead with the guide, not the beat — that is the rule, and here the guide was extraordinary in both directions. On the upside: $100 billion annualised revenue run-rate exiting December; over 2 GW of compute live by year-end; 5–10 GW in 2027; and a $1 trillion revenue target for 2030, pulled forward a full year. Gwynne Shotwell said Starlink Mobile launches at the end of 2027 using EchoStar spectrum with the explicit aim of becoming a true fourth US carrier, and that SpaceX expects to be on the Moon in 2028. Musk named Nvidia the exclusive AI silicon supplier and praised the Vera Rubin platform.
On the downside, one number: Q2 capital expenditure of $18.4 billion, of which $15.8 billion was AI. That is 236% of quarterly revenue, and 86% of all capex went into compute. Management signalled similar levels near-term. Gene Munster, who called the quarter a "blow out," pinned the selloff precisely: "SPCX down 7% on great outlook given capex will be about 65% higher than the Street in Sep and Dec."
Roughly 911 million shares, about $100 billion at current prices, became eligible to trade two trading days after the first earnings report. The stock opened $108.47, sold to $105.11 — within 28 cents of its all-time low of $104.83 — and then ripped 10.1% off that low to $115.75 before settling near $110.
Here is what almost every write-up buried. A separate tranche of nearly 456 million shares was scheduled to release today only if the stock traded at least 30% above its IPO price for five of ten consecutive sessions through August 4. The IPO priced in mid-June; the stock closed August 5 roughly 20% below that price, which puts the IPO near $135 and the release trigger near $176 — some 60% above where SPCX actually traded during the measurement window. The condition failed by a mile. Roughly $50 billion of expected supply did not show up. That, not sentiment, is why a stock facing the largest supply event in its history closed green. Terms per the 424B4 prospectus, as detailed by Investopedia.
This is the table I have not seen anyone else build. It is the most important thing on this page.
| Date | Shares released | Cumulative float | Tradable value @ $110 | Note |
|---|---|---|---|---|
| Aug 5 (pre-unlock) | — | 0.64B | $70B | 4.85% of shares out |
| Aug 6 (today) | +911M | ~1.55B | ~$171B | 1.43× the entire prior float |
| ~Nov (2nd day post-Q3) | +1,300M | ~2.85B | ~$314B | The real cliff |
| ~Dec 8 (day 180) | +456M | ~3.31B | ~$364B | Today's failed tranche, deferred |
| Post-Q2 2027 | Musk + Alphabet | — | — | Largest holders locked until then |
Float goes from 0.64 billion shares to roughly 3.31 billion in four months — 5.2×. Today's non-event is being read as "the overhang is cleared." It is not cleared; it is deferred, and it is back-loaded into exactly the window when the passive-index and post-IPO-inclusion buyers will already have finished buying.
Two offsets are real and worth quantifying. First, institutions own only 6.68% of the company against insiders at 46.44% — there is enormous room for institutional accumulation, and you cannot build a position in a 638-million-share float. The unlock is the first time real money can own this. Second, SPCX was fast-tracked into indexes, and float-adjusted index weights rise mechanically at each unlock, so passive demand scales up alongside the supply. Third, and structurally underrated: Musk's and Alphabet's shares stay locked until after Q2 2027 earnings, which means the two largest holders are legally unable to sell into the November–December wave.
Against that, CNBC reports that three tranches remain locked and that one insider, Jessie Bates III, intends to sell 100% of his stake. One executive fully exiting is a sentiment flag, not a supply event. The supply event is November.
| Metric (TTM to Jun '26) | Value | What it tells you |
|---|---|---|
| Revenue | $23.04B | +121.9% TTM growth |
| Gross margin | 51.87% | Up from 49.4% FY25, 42.9% FY24 — real scale economics |
| EBITDA | $5.29B | 22.95% margin — positive and expanding |
| Operating income | −$3.44B | Depreciation on the buildout |
| Net income | −$8.89B | −$2.27 per share |
| Operating cash flow | $9.90B | The business self-funds its operations |
| Capital expenditure | −$42.42B | 4.5× depreciation of $9.34B |
| Free cash flow | −$32.52B | The entire debate, in one line |
| Net borrowing | $102.00B | How the gap got filled |
| Cash / debt / net cash | $100.0B / $39.7B / $60.3B | Current ratio 5.12, D/E 0.31, Altman Z 20.77 |
Capex at 4.5× depreciation is the single most useful number here. It means the reported operating loss is almost entirely an accounting consequence of building assets faster than the old ones wear out. A company spending 4.5× its D&A has no meaningful earnings by construction. Which is why both sides of this debate are arguing about the wrong metric: the bears' "258× EV/EBITDA is absurd" and the bulls' "34× 2028 EPS" are the same trailing-versus-forward argument dressed differently. Neither settles anything. The only question that matters is whether the $42 billion of annual capex earns a return — and the only evidence available is that gross margin has gone 42.9% → 49.4% → 51.9% while revenue doubled. That is the tell, and it is favourable.
Starlink revenue grew 66% in Q2 with segment income up 79%. ARPU fell to $66 and subscriber adds slightly missed consensus, and that combination is being written up as a negative. My view: falling ARPU alongside 66% revenue growth and 79% segment-income growth is not deterioration — it is mix shift into cheaper international tiers plus operating leverage, which is precisely what you want from an infrastructure annuity. Treating an ARPU dip as a red flag is telco thinking applied to a business whose marginal cost of an additional subscriber is close to zero once the constellation is in orbit. Segment income growing 13 points faster than revenue is the number that matters, and it is being ignored.
Consumer plus Enterprise & Government was $11.4 billion in FY25 and is compounding above 60%. Call it $19 billion for 2026. At 6–8× forward EV/sales — a defensible range for infrastructure growing 60% with positive segment income — Starlink alone is worth $350–450 billion, or 26–33% of the current $1.37 trillion enterprise value, before you assign a single dollar to launch, AI, advertising, or the Moon.
SPCX trades at 59.3× TTM EV/sales, 21.8× forward sales, 46.1× forward earnings, 11.2× book, and 258× EV/EBITDA. PEG is 0.28. Consensus across 35 analysts is Buy with an average target of $232.44, implying 111% upside. Post-print, JPMorgan's Doug Anmuth raised his target from $225 to $240 on 34× 2028E EPS of $7.02, writing: "We went into SpaceX's earnings thinking there wouldn't be big changes to numbers given the June IPO was less than 2 months ago — we were wrong." Wells Fargo kept its Buy but cut to $215 from $230 while raising revenue and EBITDA estimates — an honest acknowledgment that the capex is worth a lower multiple. On the other side, a top-4%-ranked contributor on TipRanks calls the multiples "absurd" and the company "one of the most overpriced businesses with a market capitalization exceeding $100 billion."
Both camps are arguing multiples. Multiples are useless at 4.5× D&A. So here is the arithmetic instead — reverse-engineered from the price, using the company's own current 22.95% EBITDA margin and a 20× terminal EV/EBITDA (fair for capital-intensive infrastructure, well below software):
| Implied 2029 revenue | EBITDA @ 22.95% | EV @ 20× | Per share | % of mgmt's $1T-by-2030 guide |
|---|---|---|---|---|
| Bear | $46B | $0.92T | $74 | ~20% ($200B) |
| Today's price | $69B | $1.37T | $110 | ~30% ($300B) |
| My base case | $84B | $1.69T | $128 | ~35% ($368B) |
| Bull | $115B | $2.31T | $175 | ~50% ($500B) |
My fair value estimate is $128, with an honest band of $74 to $175. The reframing that matters: you do not need to believe management's $1 trillion 2030 guide to make money in this stock. You need about 35% of it. At today's $110 the market is discounting roughly 30% of the guide. That is a modest, unspectacular 16% discount to fair value — which is exactly why the interesting call here is tactical rather than a buy-and-hold, and why the better structural expression of this theme is a different ticker entirely.
Wells Fargo models 2027 capex of roughly $296 billion, of which about $263 billion is AI, and concludes SpaceX may need to raise more than $100 billion within 18 months. Take even half that capex number and the funding gap in 2027 is on the order of $130 billion against maybe $25–30 billion of operating cash flow.
Now put those two facts side by side. To place $100 billion of new equity-linked paper, you must sell roughly 1.4× the entire pre-unlock free float. That is structurally impossible into a 638-million-share float. It becomes merely difficult into a 3.3-billion-share float.
Second-order effect consensus is missing entirely: the lockup schedule is not an obstacle to the capital raise — it is the prerequisite for it. SpaceX needs 3+ billion shares of float by December so it can execute a mega-raise in H1 2027. Which means the November–December unlock window is when SPCX absorbs the worst of both worlds: 1.76 billion shares of mechanical supply and the market beginning to price a dilutive raise it now knows is coming. That is the window to be short or hedged, not long.
Third-order effect: if the equity is the funding vehicle, then the equity holder bears the funding risk while somebody else books the revenue. That somebody is Nvidia.
Only 51 trading days of history exist, so there are no moving averages worth citing. What exists is unusually clean:
| Level | Price | Significance |
|---|---|---|
| All-time low | $104.83 | The invalidation line. Tested at $105.11 today and rejected |
| Today's low / support | $105.11 | Bounced 10.1% off it on the largest supply day in company history |
| Current | $110.09 | RSI 39.64 — oversold-leaning, not washed out |
| Today's high | $115.75 | Bottom of the post-earnings gap |
| Gap fill / resistance | $126.06 | Pre-earnings high. Reclaim = bull trigger |
| IPO price | ~$135 | Psychological magnet; where the IPO book gets whole |
| Failed-tranche trigger | ~$176 | The level that would have released 456M shares |
| All-time high | $225.64 | June 16. A $2.97T market cap. Not coming back soon |
The setup: 165.05 million shares short — 25.84% of float, up from 111.30 million a month ago, a 48% increase — against 20-day average volume of 81.3 million shares, giving just 1.34 days to cover. Those shorts were positioned for today. Today came, the stock tagged the all-time low, and reversed 10%. A crowded short whose catalyst has passed is fuel.
Trigger: the mystery customer is named as a top-tier hyperscaler, and SPCX reclaims and holds $126.06 on above-average volume before November. Institutions at 6.68% ownership begin building into the new float; index weights ratchet up; 165 million shorts cover into 1.34 days of liquidity. Target $175–185, where the failed-tranche trigger sits as a magnet.
Trigger: no close below $104.83 and no sustained reclaim of $135. The stock chops $105–135 into November. Each rally is capped by the 1.3-billion-share Q3 unlock and then by the deferred December tranche. Sell-side targets stay at $215–240 and the stock keeps ignoring them. Year-end $115–130.
Trigger: a $50 billion-plus equity-linked raise is announced before the float can absorb it, or the November Q3 print shows AI operating losses widening with 2027 capex guided above $250 billion. The 1.3-billion-share unlock then lands on a market that has just been told it will be diluted. Book value is only $9.66 per share so it offers no floor; the floor is valuation — 15× forward EV/sales on ~$63 billion of forward revenue gives roughly $1.0 trillion of EV. Target $80–88.
| Direction | LONG |
| Entry zone | $105 – $112 |
| Target | $135 (IPO price), stretch $150 |
| Invalidation | Any daily close below $104.83 (all-time low). No ambiguity, no averaging down. |
| Timeframe | 1–2 weeks to 1–3 months. Exit before the November Q3 print regardless of P&L. |
| Risk / reward | ~4.8% risk to $104.83 vs 22.6% to $135 — better than 4:1 |
Why three independent signals agree: (1) the 456-million-share performance tranche failed its price condition, removing ~$50 billion of anticipated supply, and the tape confirmed it — the stock tagged $105.11 and reversed 10.1%; (2) shorts at 25.84% of float, +48% month-over-month, with 1.34 days to cover and RSI at 39.6, now catalyst-less; (3) both post-print sell-side revisions kept Buy ratings and raised revenue and EBITDA estimates, with consensus at $232.44 across 35 analysts. This is a mechanical, dated trade about supply and positioning — not a bet on the Moon.
| Direction | LONG / ADD — currently $217.74 (−0.68%) |
| Entry zone | $210 – $220 |
| Target | $265 |
| Invalidation | Daily close below $198 |
| Timeframe | 1–3 months |
This is my structural call and the one I would size larger than SPCX itself. Musk named Nvidia SpaceX's exclusive AI silicon supplier and praised Vera Rubin on the call. SpaceX guided 2027 compute to 5–10 GW versus JPMorgan's prior 4.2 GW estimate, with 2027 AI capex plausibly in the $200B+ range. At even $30 billion of accelerator and system content per gigawatt, SpaceX alone becomes a nine-to-twelve-figure annual Nvidia customer that was not in anyone's 2027 model six days ago. The asymmetry is the point: Nvidia books the revenue and SpaceX books the $100 billion funding risk. Same capex cycle, one balance sheet takes the pain. Today's $16.8 billion Terafab commitment is corroboration — Musk is telling you the silicon shortage outlasts 2028.
| Direction | SHORT or collar — no trigger yet |
| Entry zone | $135 – $150, only if reached before the Q3 print |
| Target | $105 |
| Invalidation | Weekly close above $155 |
| Timeframe | 1–3 months (fades after the December day-180 release) |
The mirror image of trade 1, and I want to be explicit that I hold both views without contradiction. Long the passed catalyst; short the coming one. 1.3 billion shares in November plus 456 million in December equals 2.8× today's post-unlock float arriving inside five weeks, into a market that by then will be pricing a $50–100 billion raise. If SPCX rallies into that on squeeze mechanics, you sell it. If it does not reach $135, there is no trade — do not chase it lower.
| Direction | SHORT / AVOID — currently $69.54 (+1.70%) |
| Entry zone | $69 – $75 |
| Target | $52 |
| Invalidation | Daily close above $78 |
| Timeframe | 1–3 months |
Shotwell's Starlink Mobile disclosure — launch end-2027, EchoStar spectrum, explicit "true fourth carrier" ambition — is the most direct competitive threat AST SpaceMobile has ever faced, and it came from a company with $100 billion of cash. ASTS swung from $65.70 to $74.09 today, a 12.8% intraday range on no company-specific news: pure reflexive beta to SPCX rather than any independent assessment of what was just said about its end market. Marked SPECULATIVE because it rests on one signal and the competitive impact lands in 2028, not 2026 — but the re-rating starts well before launch. Read-through is also negative for US wireless ARPU at T, VZ and TMUS, though that is a 2027 story.
The one thing to remember: SpaceX's stock is not currently being priced by investors. It is being priced by a supply schedule buried in a prospectus. Trade the schedule.
Methodology note. The Q2 press release PDF exceeded the parser's size limit, so the quarterly figures cited here come from management's prepared remarks on the earnings call and from the reported segment and cash-flow data, cross-checked against three independent sources. Lockup terms derive from the 424B4 prospectus as detailed by Investopedia. The IPO price of ~$135 and the ~$176 tranche trigger are my own derivations from the reported "20% below IPO price" at the August 5 close of $108.27 and the disclosed 30% premium condition — treat them as approximations, not disclosed figures. The scenario table uses SpaceX's own current 22.95% EBITDA margin and a 20× terminal multiple; both are assumptions and both are stated so you can disagree with them.
@dailyanalysts · Analysis, not investment advice. Positions and views can change with the data.