We published five gauges the night before SpaceX’s first-ever earnings report so the scoring could not be revised after the fact. Three cleared, two missed — and the guide we were told did not exist turns out to be the most aggressive number in the quarter: $100 billion of annualized revenue by December, against $23 billion today. Our fair value rises 18% to $78.51. The recommendation does not change. And tomorrow morning, 911.5 million shares become free to trade into a float of 646 million.
The verdict first. SpaceX’s first quarter as a public company was, on the operating numbers, better than we expected: revenue $7.81B (+92%), adjusted EBITDA $3.5B, and an AI segment that went from a rounding error to $2.56B of quarterly revenue with positive segment EBITDA. We are raising our probability-weighted fair value from $66.50 to $78.51, an increase of 18%. We are also maintaining AVOID: at $115.59 the expected return to that revised fair value is -32%, and the enterprise value still capitalizes $88B of steady-state annual free cash flow against a business burning $64B a year. Raising a target and staying negative is not a hedge; it is what the evidence supports, and the rest of this note shows the work in the order we committed to showing it: the map we published, the gauges scored against it, the model that follows from the score, and the supply event that lands tomorrow.
Note the tape, because it is doing the argument for us. The stock closed at $125.33 the day of the print, spiked to $130.99 after hours, and now trades at $115.59 — +0.9% versus the $114.53 close of the session before the release. Twenty-four hours of digestion have round-tripped the entire earnings move.
The night before the print we published one chart with every price that mattered on a single line: our scenarios, our fair value, and the options market’s implied range. Here it is again — unedited on the top row, resolved on the bottom.
Top row: the map exactly as published 3 August. Bottom row: the print’s round trip, today’s price, and the revised scenario set. The gold connector is the fair value moving from $66.50 to $78.51.
Three of the map’s claims resolved inside twenty-four hours. The stock never left the options market’s priced range: the wildest moment of the reaction — the $130.99 after-hours high — came within 84 cents of the straddle’s $130.15 upper breakeven and stopped. The straddle’s profit had a shelf life of one session: bought at $18.15 on Monday, it marked $22.83 at Tuesday’s close (+26%) and $10.48 this morning (−42%), as implied volatility collapsed from ~189% to ~143% — we wrote that priced that rich it needed “an outcome outside our entire scenario set” to pay, and no such outcome arrived. And the distance to fair value closed from both directions: the price is a dollar above where it sat before the print, while the evidence moved the fair value up 18% to meet it. What remains between them is the −32% this note stands behind.
We published these five gauges before the evidence existed, which is the only way a scorecard means anything. Each one is reproduced below exactly as it ran on 3 August — no edits, no softened thresholds — with the verdict underneath. One of the misses is worse than we expected. One of the clears is a correction against ourselves.
Blended ARPU has fallen at every disclosed mark — $99 → $91 → $81 across FY23–FY25, and $86 → $66/month year-over-year in Q1 (−23%) — as growth shifts to lower-priced tiers and markets. Subscribers doubled to 10.3M (Q1, +106% YoY) — the bull case needs both: a subscriber print near ~12M and ARPU stabilizing in the $60s. Growth that arrives only by giving away the margin is the base case, not the bull.
The print: ARPU held at $66, flat sequentially — the first hold in the disclosed series — and subscribers reached 12.0M on a record 1.7M net adds. Both legs of the bar arrived: subs at the ~12M mark, ARPU stable in the $60s. Growth without giving away further margin — this quarter, at least.
The segment printed −$351M adjusted EBITDA in Q1 — its first negative quarter since 2018 — as Starship development costs ran through it. A second negative quarter turns the “launch monopoly funds the moonshots” narrative into a third mouth to feed. $429B of implied value is riding on this line.
The print: The second negative quarter happened: −$205M adjusted EBITDA, narrowed from −$351M but still below zero, with segment R&D of $1,076M exceeding segment revenue of $962M. The third mouth is feeding. Starship Flight 13 “went incredibly well” per Musk — the engineering progresses; the economics have not arrived.
The S-1 gave xAI ~$0.2B of partial-period revenue and a “nameplate compute” metric. At $111, the segment carries an implied ~$386B. The print needs a real revenue number with a growth rate, and disclosure on the GPU financing arrangements funding the buildout. A compute-capacity slide without a revenue ramp is the bear’s exhibit, not the bull’s.
The print: The real revenue number came, and it was large: $2,561M, up 213% sequentially, with segment adjusted EBITDA swinging to +$1,146M from −$609M, $14.1B of contracted Cloud Services Agreements, and 1.4GW of nameplate compute. The GPU-financing disclosure arrived too — in the 10-Q, and it is its own section of this note. The gauge that cleared most emphatically, and the one that drives our upgrade.
Q1 investing outflows were $16.7B against $1.0B of operating cash flow. This print delivers the first post-IPO balance sheet. What we want is management’s first statement on when capex crests — without one, the $82B steady-state question has no bridge at all.
The print: No peak — the opposite. Capex was $18,369M in the quarter ($15,828M of it AI), and CFO Bret Johnsen guided forward, not down: “the next two quarters are very similar to the current quarter from a CapEx level perspective.” Roughly $36B more before year-end. The gauge did not merely fail; it inverted. The steady-state question still has no bridge.
Consensus expects −$0.16 EPS — roughly a $2.1B net loss. We believe the EPS print itself is nearly meaningless against merger accounting; the first-ever guidance, and whether management guides FY26 at all, will move the stock more than the quarter does.
The print: Scored with a correction against ourselves. The quarter came in at a $541M net loss — far smaller than the ~$2.1B the consensus EPS implied, and as predicted, nearly beside the point. The guidance existed — but only on the call, not in the release: a $100B annualized revenue run-rate by December 2026, with the $1 trillion revenue target pulled from 2031 to 2030. A draft of this note scored the gauge NOT GIVEN off the release alone; the call corrected us, and the guide’s size earns it the next section.
The gauge-5 correction deserves one plain sentence outside the box: the guidance exists only in spoken form — a reader who checks the filed 8-K will not find it — and we held this note a day to verify the call record before scoring it. The difference between a filing and a disclosure event is now, itself, part of the SPCX story.
This is the most consequential thing said all quarter, and it did not appear in a single filed document. Management guided to a $100B annualized revenue run-rate by December 2026. Trailing twelve-month revenue is $23.0B. Annualizing the quarter just reported gives $31.3B. Reaching $100B ARR by December therefore requires roughly 3.2× the current run-rate inside two quarters. CFO Bret Johnsen framed it as a trajectory — “we believe this puts us on a trajectory… to reach $100 billion of ARR… by the end of this year” — and Musk went further: “the $100 billion ARR in December is not a question mark. That’s what we would achieve if we basically did nothing.” Presented, in other words, not as the stretch case but as the floor.
Each bar annualizes that quarter’s reported revenue. The gold line is the December 2026 target management stated on the call.
We believe the honest treatment of this number is neither to adopt it nor to dismiss it. It is not impossible: the AI segment tripled sequentially, the $14.1B of Cloud Services Agreements the company reported entering in the quarter contributed $1.6B of its revenue and leave a large contracted balance still to convert, and nameplate compute went from 400MW a year ago to 1.4GW today. It is also a target from a management team that, in the same call, set a tentative 20GW power-and-cooling goal for end-2027 and then immediately walked it back: “I don’t think we’re going to achieve 20 GW… we’d still probably have… something close to 15 GW.” A company willing to tell you which of its own goals it will miss has earned some credit; a 3.2× revenue step in six months has not yet earned any.
The scorecard converts into numbers as follows: what the quarter proved raises Connectivity and AI, what it failed to prove trims Space, and the first real balance sheet corrects the June net-cash estimate. Same probabilities, same discipline, new evidence.
| Segment, $B | Bear | Base | Bull | What changed |
|---|---|---|---|---|
| Connectivity | $190 | $370 | $700 | Raised. ARPU stabilized at $66 on record net adds; 66% revenue growth. |
| Space | $65 | $170 | $470 | Trimmed. A second negative-EBITDA quarter; Starship still pre-economics. |
| AI | $110 | $330 | $620 | Raised hardest. $2.56B quarterly revenue and a positive EBITDA swing, from a standing start. |
| Net cash | $60.6 | $60.6 | $60.6 | Cut from the $79.5B we carried in June — the first real balance sheet. |
| Per share · 13.176B shares | $32.30 | $70.63 | $140.46 | Probability-weighted (25/50/25): $78.51, up 18% from $66.50. |
The net-cash correction is worth isolating. In June we modelled $79.5B of net cash from the prospectus, with total debt flagged as an open estimate. The actual figure, now that a balance sheet exists, is $60.6B — $18.9B lower, costing $1.43 per share on the 13.176B shares now outstanding. It is the single largest downward revision in the model and it came from replacing our estimate with a filing.
That last figure is the discipline of the whole exercise. We raised the bull case to $1,790B — and the market still asks for 82% of it. At $63× trailing revenue, the multiple compressed from the IPO’s 90× mostly because revenue grew, not because the price fell.
This is the most immediate fact about SPCX and it has nothing to do with the quarter. SpaceX listed with an exceptionally thin float: 646 million shares, 4.9% of the 13.18 billion outstanding — a touch above the 638.9 million Class A shares sold in the IPO once the underwriters' option was exercised in full. The lock-up on the remaining stock is not a single cliff but a ladder, and its first and largest early rung is tomorrow.
The prospectus is explicit: for shares subject to the 180-day lock-up, an automatic early release occurs “on or after the second full trading day on Nasdaq immediately following the public release of our quarterly financial results… for the quarter ended June 30, 2026… up to 20% of the shares may be Transferred.” The release was Tuesday 4 August. The first full trading day was today. The second is Thursday 6 August. The prospectus puts that first tranche at 911.5 million shares — the 20% step, net of stock held by affiliates, out of a 180-day lock-up pool of roughly 4.69 billion shares.
Shares free to trade, cumulative, on the 424B4 release ladder. Tranches after tomorrow assume no further conditional releases and exclude the founder and extended-lock-up pools.
The detail most calendars will get wrong. A second tranche of 455.8 million shares — a further 10% — was available on the same date, but only “if the reported closing price… is at least 30% greater than the public offering price… for at least five of the ten consecutive trading days ending on, and including” the release date. Thirty percent above the $135 IPO price is $175.50. Across those ten sessions the stock’s best close was $125.33 — -28.6% below the trigger, and still below the IPO price itself. Zero of ten days qualified. Tomorrow is 20%, not 30%. The failed condition is its own verdict on the listing: the lock-up was drafted with a reward for a stock trading 30% above issue, and that reward was never in reach.
| Release date | Tranche | Shares | Cumulative float |
|---|---|---|---|
| Thu 6 Aug 2026 — second full trading day after the release | 20% | 911.5M | 1.56B |
| Conditional tranche — price test failed, see below | 10% | — | — |
| 20 Aug 2026 (70th day) | +7% | 319.0M | 1.88B |
| 9 Sep 2026 (90th day) | +7% | 319.0M | 2.20B |
| 10 Sep 2026 (91st day) — affiliate shares released earlier | — | 59.1M | 2.25B |
| 24 Sep 2026 (105th day) | +7% | 328.4M | 2.58B |
| 9 Oct 2026 (120th day) | +7% | 328.4M | 2.91B |
| 24 Oct 2026 (135th day) | +7% | 328.4M | 3.24B |
| Second trading day after Q3 results | +28% | 1.3B | 4.54B |
| 8 Dec 2026 (180th day) — remainder | +17% | 797.6M | 5.34B |
And the counterweight, stated as plainly as the risk. The founder’s shares carry a 366-day lock-up with no early release provisions whatsoever, running to 12 June 2027. A further group of holders sits under an extended lock-up into 2027. Together those two pools are approximately 7.8 billion shares — 59% of the company on the current share count (the prospectus states the same pool as “approximately 60%” post-offering and “greater than 63%” pre-offering — the difference is the denominator, not the pool) — and none of it can be sold this year. The overhang is real, mechanical, and bounded. What arrives tomorrow is $105B of newly saleable stock, roughly 12.3 days of average volume, into a market that has already fallen 8% today.
We are not forecasting the reaction, and we note the obvious confound: today’s decline may be the print, the unlock being front-run, or both, and the two cannot be cleanly separated. The point is that a supply event of this size is knowable in advance from a document filed in June, and it lands tomorrow morning.
The 10-Q answers a question the release left open, and the answer is mixed. We had flagged the risk that the new AI revenue might be affiliated. The filing’s related-party note itemises only asset purchases from Tesla and the Valor financing described below, then states that other related-party transactions were immaterial. It does not address segment revenue in terms. We read that as implying the $2.56B of AI revenue is substantially arm’s-length — an inference, not a statement by the company, and one a later filing could complicate. On that reading it supports the upgrade, and we say so.
The financing is a different matter. Of $39.4B in total debt and finance leases, $13.3B is owed to Valor Equity Partners — whose founder, CEO and Chief Investment Officer, Antonio Gracias, serves on SpaceX’s board. The arrangement is an equipment lease for AI infrastructure hardware which the company accounts for as a failed sale-leaseback: economically a secured borrowing, not a lease. It has grown from $4.5B at year-end to $13.3B at 30 June — +196% in six months — and cost $327M of interest in the quarter alone, against zero a year ago. Separately, SpaceX bought $295M of Tesla Megapacks in the quarter for datacenter power, a real disclosed commercial link between the two companies that neither the release nor the call discussed.
None of this is improper, and all of it is disclosed. But a third of the debt funding the largest capital program in the company’s history sits with an entity connected to its own board, and that belongs in any assessment of capital quality.
AVOID maintained. Probability-weighted fair value $78.51 (bear $32.30 / base $70.63 / bull $140.46 at 25/50/25), up 18% from $66.50. Expected return from $115.59 is -32%. The Fund holds no position.
What would move us to a constructive view, pre-registered as before: a capex peak actually dated by management; a third consecutive quarter of AI revenue growth at anything near this rate, with the Cloud Services backlog converting; Space segment EBITDA back above zero; and December ARR arriving within reach of the $100B claim. Meet those and the bull case is the base case. What we will not do is mark our view to a target given verbally on a call, three quarters before it can be tested.
The near-term calendar is now unusually legible: tomorrow’s unlock, then further 7% tranches on 20 August, 9 September, 24 September, 9 October and 24 October, then 28% — up to 1.3 billion shares — two days after the Q3 print, then the balance on 8 December. For a stock whose float is currently 4.9% of shares outstanding, that schedule is likely to matter more over the next four months than anything in the income statement.
Company filings. Q2-2026 results: Form 8-K and Exhibit 99.1, SEC accession 0001628280-26-052515, filed 4 August 2026 (all Q2 revenue, segment, adjusted-EBITDA, capex and cash-flow figures; period ended 30 June 2026). Quarterly report: Form 10-Q, accession 0001628280-26-052535, filed 4 August 2026 (balance sheet, share counts from the cover page, Note 18 Segments, and the Related Party Transactions note supplying the Valor and Tesla Megapack figures). Lock-up terms: Form 424B4 final prospectus, accession 0001628280-26-042639, filed 12 June 2026 — “Shares Eligible for Future Sale” and the Underwriting lock-up provisions, from which the release ladder, the 20% early-release clause, the 30%-of-offering-price condition, the 366-day founder lock-up and the extended lock-up pool are taken verbatim. June valuation inputs: SpaceX Form S-1/A, accession 0001628280-26-040364, filed 3 June 2026. All filings retrieved from SEC EDGAR on 5 August 2026.
Earnings call. Q2-2026 earnings call, 4 August 2026, 16:30 ET. Quotations attributed to CFO Bret Johnsen (capital-expenditure phasing, the ARR trajectory) and CEO Elon Musk (the $100B ARR remarks, the 20GW/15GW capacity comments, the $1 trillion revenue date, Starship Flight 13) are taken from published transcripts of that call retrieved 5 August 2026, and were cross-checked against two independent transcript publishers (Investing.com and MarketBeat) before use, because two of the five gauge verdicts in this note turn on them. None of these statements appears in any filed document: readers who verify only against the 8-K will not find them. The call is a secondary source relative to the filings; every figure here that carries a valuation consequence is sourced to a filing, and call-sourced statements are identified as such in the text.
Market data. SPCX $115.59 and TSLA $324.64 at 14:28 UTC (10:28 ET) on 5 August 2026; official closes of $125.33 (4 August) and $114.53 (3 August); the after-hours high of $130.99 and low of $114.33 on 4 August; the ten official closes of 22 July–4 August used to test the lock-up price condition; and the 8/7 $112 straddle marks used in the resolution chart ($18.15 on 3 Aug ~11:00 ET; $22.83 at the 4 Aug close, call $18.33 + put $4.50; $10.48 on 5 Aug ~11:13 ET, call $7.03 + put $3.45, with implied volatilities of ~189% and ~143% at the two endpoints) — all from the brokerage consolidated market-data feed, pulled 5 August 2026. Float of 646.0 million shares and 30-day average volume of 74.4 million shares from the same feed, same date. Note that the feed’s shares-outstanding field (14.02 billion) differs from every filed count; we use 13.176 billion (7,607M Class A + 5,569M Class B) for every per-share calculation. That is the issued and outstanding count on the 30 June 2026 balance sheet, chosen to match the balance-sheet date of the net-cash figure it is paired with — not the 10-Q cover page, which reports 13.182 billion as of 28 July 2026. The 0.04% difference is immaterial to every figure here.
Formulas and conventions. Enterprise value = filed share count × market price − net cash, where net cash = cash and equivalents $93,522M + short-term investments $6,487M − debt and finance leases current $2,525M − non-current $36,839M = $60,645M. Implied steady-state free cash flow = EV × (WACC − g) with WACC 10% and g 4%, the same constants used in the June initiation. Annualized free cash flow multiplies the reported quarter (operating cash flow $2,419M less capital expenditure $18,369M = −$15,950M) by four; it is a run-rate illustration, not a forecast. Quarterly figures are derived by differencing six-month from three-month disclosures where the company reports on a year-to-date basis. Trailing-twelve-month revenue = FY2025 $18,674M − H1-2025 $8,138M + H1-2026 $12,508M = $23,044M. Sum-of-the-parts values are our estimates, not company disclosures, and the scenario probabilities (25/50/25) are unchanged judgment carried from the June initiation. Cumulative-float figures in the lock-up table assume every scheduled tranche releases in full, no further conditional releases, and no sales by founder or extended-lock-up holders; they are ceilings on supply, not predictions of selling.
Prior TON618 research relied upon. The First Print: What $111 Still Has to Believe (3 August 2026), which pre-registered the five gauges scored here and supplies the June comparison figures; and SpaceX (SPCX) Tech-Equity Valuation — Initiating Coverage (12 June 2026), which supplies the original sum-of-the-parts, the $66.4966 probability-weighted fair value, and the $79.5B net-cash estimate corrected in this note.
Limitations. The $100B December ARR figure is a management target stated on a call, not a filed projection, and is treated in this note as a claim to be tested rather than an input. The “consensus” framing of pre-print expectations is deliberately omitted: we could not verify a primary-source estimate. Today’s price decline cannot be cleanly attributed between the earnings reaction and anticipation of tomorrow’s unlock, and we do not attempt to separate them.
Information only. TON618 Capital. This report is for information purposes only. Nothing here is an offer to sell or a solicitation of an offer to buy any security, fund interest, or digital asset, and nothing here is personalized investment advice or a recommendation regarding any instrument.
Publisher's exclusion. All research is published solely as general, impersonal information of regular circulation. It is not tailored to the objectives or circumstances of any individual and is not issued in connection with compensation from any client. The Fund has no clients and distributes all research free of charge. On that basis it publishes in reliance on the publisher's exclusion from the definition of "investment adviser" under the Investment Advisers Act of 1940 (§202(a)(11)(D); cf. Lowe v. SEC, 472 U.S. 181 (1985)).
Registration & conflicts. TON618 Capital is not registered as an investment adviser or broker-dealer in any capacity. The Fund is a Bitcoin fund and may hold or transact in the securities or digital assets it discusses; material conflicts are disclosed where they exist. Ownership: the Fund holds no position in SPCX or TSLA as of the report date, and neither is a Fund holding or a BTC-correlated instrument. Compensation: the Fund received no compensation from any party in connection with this report and charges nothing for it.
Use of AI. Artificial intelligence is used in the creation of this research. All methodology and data integrity are reviewed and approved before publication by TON618 Capital's Chief Investment Officer, Keyth Beck; errors may nonetheless occur, and readers should verify independently.
CFA. This report was prepared to align with CFA Institute analytical standards (methodology only). CFA® and Chartered Financial Analyst® are registered trademarks owned by CFA Institute. That reference describes the analytical framework applied; it does not imply the report was prepared, reviewed, or authored by a CFA charterholder, and the report is not issued, reviewed, endorsed, certified, or approved by — nor affiliated with — CFA Institute.
Risk & feedback. Past performance is not indicative of future results. Digital assets and equities are volatile and may result in total loss of capital. Corrections and feedback are welcome — please direct them to CIO Keyth Beck at keyth@ton618capital.com. Version 1.0 · analyst: TON618 Equity Research.