On Tuesday, SPCX did two things it had not done in a while. It closed higher — up 3.08 percent to $123.54 on 79.7 million shares, ending a seven-session losing streak that had taken the stock to $119.85, its first close below $120. And its company gave the market a date: SpaceX will report its first-ever quarterly earnings on Tuesday, August 4.
That second item matters more than the first, because it was not the date the market had been working with. Since the IPO, coverage — including our July 16 analysis and our July 20 analysis — had treated August 6 as the expected earnings date. The actual structure, now set: earnings on August 4, and the first lockup tranche opening two trading days after the report — which lands on August 6. That first tranche alone releases up to roughly 911.5 million shares, worth approximately $109 billion at current prices, into a market whose traded float has been 3 to 5 percent of shares outstanding since the June listing. Broader early-August unlock estimates run to $123 billion, and by early September — per 22V Research — as much as 44 percent of shares outstanding could be sellable.
And before any of that: Starship Flight 13 launches tomorrow evening, in a 90-minute window opening at 6:45 PM Eastern. The stock that spent last week collapsing now faces three scheduled events in fifteen days — a rocket, a first earnings report, and a supply wall — with options pricing a 25 percent move over the next month and 28 percent of the float still positioned short. Tuesday’s bounce did not resolve any of this. It repriced the entry.
What Tuesday’s Session Actually Was
The streak-breaking rally had identifiable components, and none of them was new fundamental information about SpaceX’s business.
The first was a sell-side catalyst: Macquarie’s Paul Golding reiterated an Outperform rating and a $250 price target with the explicit instruction to “buy any dip.” The note’s anchor is not the launch business. It is SpaceXAI’s compute operation — Anthropic reportedly taking the entirety of Colossus 1’s roughly 300 megawatts and more than 220,000 Nvidia GPUs — which Macquarie treats as the under-modeled asset in the stock. The note landed on a sector primed to move: AST SpaceMobile jumped 12 percent, Rocket Lab and Virgin Galactic rallied alongside, and SPCX rose with the group in a broad risk-on rotation into space names.
The second was the earnings-date announcement itself, which traded bullish on the day despite being, structurally, the news that starts the unlock clock. A dated earnings report reads as a step toward normalcy — a company that discloses on a calendar is a company behaving like a public company — and after seven straight losses, the market took the reduction in uncertainty as a reason to cover and buy.
The third was theater. Elon Musk spent Tuesday publicly taunting the short base, putting their survival odds at “very low.” Whatever one makes of chief executives engaging short sellers on social media — a genre with a long and mixed history — the fact of it confirms that the company is aware of, and playing to, the positioning battle in its own stock.
Wednesday’s pre-market, giving back roughly 1 to 2 percent to $122.17, suggests the market’s own assessment of Tuesday: a repositioning day, not a re-rating.
The Calendar Correction and Why It Matters
Moving the earnings date from an assumed August 6 to a set August 4 is a two-day shift that changes the shape of the event.
Under the assumed structure, earnings and unlock were a single fused moment — one date on which the market would learn the fundamentals and absorb the supply simultaneously. Under the actual structure, they are sequenced: the market gets the Q2 numbers on Tuesday, August 4, then has two trading sessions to price them before up to 911.5 million shares become eligible on Thursday, August 6. That sequencing is materially better for price discovery — and materially more revealing. The tape on August 5 will show what institutions think the numbers are worth before insider supply arrives; the tape on August 6 and after will show how much of that supply actually comes to market and at what urgency.
It also means the two events cannot rescue each other. Under a fused structure, a strong quarter and a heavy unlock would have collided in one session, with the net move ambiguous. Sequenced, each gets its own verdict. A strong quarter that rallies the stock into August 6 invites more selling from the unlocked tranche — the pattern we described on July 20, in which any rally functions as an exit-price restoration for insiders who have waited a decade for liquidity. A weak quarter that breaks the stock below $120 ahead of the unlock confronts locked holders with a falling knife and may, perversely, defer supply. The sequence turns August 4 through 6 into a controlled experiment on the question this listing has posed since June: is there institutional demand for this equity at scale, or has the price been an artifact of scarcity?
One housekeeping note on the numbers. The first-tranche figure of roughly 911.5 million shares (about $109 billion) and the broader early-August framing of $123 billion circulate alongside the 1.37 billion-share figure used in earlier lockup coverage, which described the total base tranche. The precise tranche arithmetic depends on S-1 definitions of sellable holdings; the load-bearing fact is unchanged at every estimate — the supply eligible to trade in early August is a multiple of the entire float that has set every price since the IPO.
A Consensus That No Longer Agrees With Itself
When the underwriters’ quiet period ended in early July, SPCX coverage was a wall: fourteen initiations, every one buy-equivalent, averaging $187.80. Three weeks later the coverage picture has both broadened and split, and the split is the story.
The aggregate consensus now spans roughly 32 analysts with an average target near $244.50 — pulled upward by the newer, AI-anchored bulls: Macquarie at $250, JPMorgan at $225, a Raymond James target of $800 that is less a forecast than a thesis statement about SpaceXAI. Against them: Piper Sandler’s Neutral, initiated the week the stock broke its IPO price, citing valuation and the lockup calendar. CFRA at Sell on valuation. Needham at $250 but with the caution flags up. The dispersion — from Sell to $800 — is among the widest for any large-capitalization stock in the market.
Dispersion of that width means the analysts are not disagreeing about a number. They are disagreeing about what the company is. The $800 case values a vertically integrated AI-compute-and-aerospace conglomerate in which Starship, Starlink V3, Colossus, and the Pentagon relationship compound. The Sell case values a launch-and-satellite business at 49 times expected revenue with a decade of insider liquidity about to come due. Both cases will read the same Q2 report on August 4 and find support in it. What the report cannot do is resolve which company SpaceX is — that resolution belongs to the segment disclosures, the SpaceXAI revenue line (if one exists), and the capital-allocation commentary, which is why the fine print on August 4 will matter more than the headline numbers.
The Binary Is Priced. The Direction Is Not.
The options market has done its arithmetic on the next month: implied volatility above 100 percent, straddles pricing approximately a 25 percent move. That is the market saying, plainly, that it expects the fifteen-day gauntlet to move the stock a fifth of its value — and admitting it does not know which way.
The short base has made the more interesting statement. Short interest sat near 28 percent of the float — roughly $25 billion — through the entire decline, and the available data shows no meaningful covering into the $119.85 low. Shorts who do not cover a 47 percent drawdown are not momentum traders taking profits; they are positioned for an event, and the event is August, not tomorrow. The bet, as we framed it on July 20, is that whatever Flight 13 does, the supply arrives anyway.
Tomorrow tests the other side of that bet. A clean flight into a 28-percent-short, 4-percent-float stock that has just broken its losing streak is the textbook squeeze setup — thin supply, forced buyers, a taunting chief executive, and a sell-side note instructing clients to buy the dip already in circulation. If the squeeze runs, the mechanics that carried the stock down can carry it up violently for a session or three. The July 20 analysis laid out why that rally, if it comes, has a ceiling built into its own success: every dollar of squeeze is a dollar of restored exit price for the tranche that unlocks August 6, and the calendar caps the squeeze’s half-life at nine trading days.
A second consecutive abort — or worse, an in-flight failure — lands on a stock that has bounced 3 percent off its low with all of the structural overhang intact. The $119.64 intraday print from Tuesday morning becomes the level to watch, and below it there is no chart history at all: the stock has never traded there, and the only reference points are the private-market tender rounds far below.
The Eve From Four Seats
Binary events are best understood by walking the positions, because each holder class faces tomorrow with a different payoff structure and the aggregate of their decisions is the price.
The long-only institution holding SPCX from the IPO is down roughly 8 percent from the $135 offering and as much as 45 percent from the June marks at which some added. Its decision is not about the launch — it is about whether to carry current size through August 6. A successful flight that lifts the stock toward $135-140 is, for this holder, the best exit window between now and the unlock; a holder who believed in the $187.80 initiation consensus in early July but has watched the dispersion widen has every incentive to lighten into strength. This is the constituency whose behavior on a post-launch rally will decide whether the squeeze extends or caps.
The short at 28 percent of float has already answered its question: it held through a 47 percent drawdown and did not cover at $119.85, which means its risk tolerance is calibrated to the August supply event, not to tomorrow’s telemetry. Its exposure is a violent covering rally on launch success — the scenario in which it is forced to buy a 4 percent float alongside momentum entrants. The rational short response to that risk is not to cover now but to size for it, and the stable short-interest number suggests exactly that: positions sized to survive a squeeze in order to be present for the unlock.
The insider — locked until August 6 at the earliest — is the only seat with no decision to make tomorrow, and the most consequential one to make in three weeks. What the insider watches tomorrow is not the stock but the print: a successful flight followed by a rally tells the cap table that liquidity at $140-150 may exist in the unlock window; a failure followed by a break below $119 tells a decade of patient holders that the first liquidity event of their tenure arrives into a falling market. Insider behavior is the single largest unpriced variable in this stock, and the trailing data offers one hard clue: through the entire drawdown, insider filings show roughly $1.2 million in sales and zero purchases. No one with August 4’s numbers in hand has bought the dip.
The retail and dip-buying constituency — ARK’s roughly $36 million week, the Binance perpetual traders, the options flow paying triple-digit implied volatility for calls — is the seat with the widest outcome range. It has been the marginal buyer on every bounce, including Tuesday’s, and it is the constituency the squeeze scenario needs. It is also the constituency that bought the July 16 launch and was handed an abort at T-0, a sequence worth remembering: the last time this market positioned for a Starship catalyst, the catalyst did not occur, and the stock fell 5.4 percent the next session anyway. Eve positioning in this stock has so far been a losing trade twice.
Launch Readiness, Honestly Stated
The engineering picture into tomorrow is cleaner than the July 16 attempt but not clean.
Favorable: the FAA closed the Flight 12 mishap investigation earlier this month — root causes identified as ascent heating effects on propulsion components plus erroneously calibrated engine-alarm settings, with four corrective actions — so there is no regulatory blocker. The two Raptors that failed to ignite on July 16 have been removed and replaced, per Musk. The payload — the first twenty Starlink V3 satellites, flying a suborbital profile to an Indian Ocean disposal — is integrated and unchanged.
Unfavorable: the July 16 abort was itself an ignition-sequence failure across four engines, of which only two were replaced, implying the other two were cleared by inspection rather than swapped. And SpaceX’s abort triggers are demonstrably hot this week — the company’s own Falcon 9 program logged two consecutive last-second T-0 aborts at Vandenberg on July 20 and 21 before flying the mission successfully on the third attempt. Conservative abort logic is good engineering and bad theater. A third Starship scrub would be harmless to the program and expensive to the stock, which has now twice bought the rumor of a launch and twice been handed a postponement.
The Starlink V3 payload keeps its longer-horizon significance regardless. V3 is the capacity generation on which Starlink’s subscriber ceiling depends, the FCC filing for up to 100,000 next-generation satellites presumes Starship-class lift, and Amazon’s Kuiper is spending its first commercial year signing the customers Starlink cannot yet reach. Every week of Starship slippage is a week of deferred Starlink capacity. The market prices the launch as a sentiment event; on this one dimension it is a revenue-timeline event, and the revenue timeline now has an earnings date attached to it.
The AI Subplot Is Becoming the Plot
Three items from the past week, taken together, explain why the bull case has quietly migrated from rockets to racks.
First, the Pentagon: SpaceX is reportedly in early talks to supply the Department of Defense with billions of dollars of AI data-center compute, with the Defense Secretary reportedly pushing to integrate Grok into military networks this month. The talks could collapse — both sides say so — but the direction is legible: SpaceXAI is being positioned as national-security infrastructure, a procurement category with margins and moats no commercial launch contract offers.
Second, the Microsoft channel: Grok was added to Excel and Outlook this week, placing SpaceXAI’s model inside the world’s largest productivity suite alongside Copilot. Distribution of that scale, achieved without SpaceXAI building an enterprise sales force, is precisely the kind of asset the Macquarie note is trying to capitalize.
Third, the noise: Taiwan’s Economic Daily reported a $52 billion SpaceX order to Foxconn for roughly 13,000 Nvidia GB300 racks — and Musk denied it the same day as “fake news.” The claim has no filing behind it and should be treated as false. But the fact that a $52 billion figure was reportable, and briefly credible, measures how large the market’s imagination for the SpaceXAI business has become — and how little disclosed fact constrains it.
All three items share a property: none of them appears in any audited financial statement. The August 4 report is the first document that can either give the AI subplot a revenue line or reveal that, for now, it remains a story. That is the same gap — narrative running ahead of disclosure — that this series has tracked since before the Nasdaq-100 inclusion, now concentrated onto a single filing.
What August 4 Must Actually Disclose
Since the earnings date is now fixed, the checklist for the report itself deserves specificity beyond the headline revenue and margin numbers. Five disclosures will do most of the work of resolving the analyst dispersion.
First, segment structure. How SpaceX chooses to segment — Starlink versus launch services versus “other,” or a finer cut that isolates Starship development and SpaceXAI — is itself information. A filing that buries SpaceXAI inside an aggregate segment tells the market the AI business is not yet material enough to disclose, which directly undercuts the $250-and-above price targets built on it. A standalone AI-compute line, even a small one, validates the Macquarie framing overnight.
Second, Starlink’s subscriber and ARPU trajectory, with the Kuiper quarter in the base period. The IPO projections implied a specific growth path; this is the first audited checkpoint against it, and the first quarter in which the competitive environment included a funded, shipping alternative.
Third, the cash conversion of the launch backlog. Booked missions are not revenue; the report will show what actually flew and billed in Q2, and any gap between backlog optics and recognized revenue will be the bears’ first exhibit.
Fourth, Starship program cost — now including whatever the July abort-and-retry campaign added. Two test flights, an engine-replacement cycle, and a mishap-investigation closure all landed inside or adjacent to the reporting period. The program-cost line is where the market learns what the Falcon business is actually subsidizing.
Fifth, guidance practice. SpaceX has no obligation to guide, and a company famous for disclosing nothing may choose to report bare minimums. The choice matters mechanically: with no guidance, the analyst dispersion cannot compress, the $800 target and the Sell rating both survive, and the stock stays a narrative instrument into the unlock. A company that wants an orderly August 6 has one lever available on August 4 — reduce the uncertainty premium by disclosing more than required. Whether it pulls that lever is a test of how much the people who set the performance trigger at $175.50 care about the price their insiders exit at.
Fifteen Days, Three Events, One Question
Here is the gauntlet as it now stands, with the market’s own pricing attached.
Thursday, July 23: Flight 13, window opening 6:45 PM Eastern. Binary sentiment event with squeeze potential on success and a $119.64 test on failure; either outcome resolves within sessions. Tuesday, August 4: the first earnings report — Starlink subscriber trajectory against Kuiper’s first commercial year, launch-cadence revenue conversion, Starship program cost, and whatever the filing does or does not say about SpaceXAI. Thursday, August 6: the first unlock tranche, up to 911.5 million shares, whose realized selling pressure will be observable in the tape within days.
Options price the ensemble at plus-or-minus 25 percent. The short base prices it as supply overwhelming demand regardless of path. The Macquarie wing of the sell side prices it as the entry point of the year. They cannot all be right, and by August 10 the tape will have adjudicated most of it.
The question underneath all three events is the one this series has asked since the pre-inclusion rally reversed: what is the demand for this equity when nothing forces anyone to buy it? Index mechanics forced buying in early July; scarcity did the rest. The launch may rent enthusiasm for a session. But August 4 and 6 are the first dates on which SpaceX’s public valuation must be carried by disclosed numbers and voluntary buyers at full float-adjusted supply. Tuesday’s 3 percent bounce, a $250 target note, and a chief executive taunting his shorts are the eve-of-battle positioning. The battle is the calendar, and it starts tomorrow at 6:45.
The near-term watch list, in order: whether Flight 13 clears the pad — a third scrub is the quiet scenario nobody is positioned for; the size and half-life of any post-launch move against the $119.64 and $129.88 levels that now bracket the week’s range; whether a second analyst joins Piper Sandler off the buy wall before earnings, or whether the Macquarie wing pulls the average target higher still; the July 24 short-interest settlement data, the first reading that could show covering ahead of the gauntlet; and any amendment or clarification to the lockup tranche arithmetic in SpaceX’s filings, where the difference between 911 million and 1.37 billion eligible shares lives. Each is checkable within days. The series will return when the first of them prints.
