SPCX · SpaceX (Space Exploration Technologies Corp.) · Stock research

Current Thesis

Two dated unknowns that had defined this name since listing were removed inside three sessions.

The first quarterly disclosure in company history landed after the close on 2026-08-04. Revenue $7.81B, up 92% year on year, against roughly $6.93B consensus. Diluted loss $0.09 versus a $0.26 expected loss. Adjusted EBITDA $3.54B, up 191%, at a 45.3% margin. Connectivity $4.291B with $1.656B of segment operating income and 12.0 million Starlink subscribers at $66/month ARPU; AI $2.561B, up 247%, carrying a $1.257B operating loss on 1.4GW of nameplate compute against 0.4GW a year earlier; Space $962M, up 29%, at a $542M operating loss.

What moved the tape was the spend line. Capex was $18.37B for the quarter, $15.83B of it AI-related, against a ~$13.2B consensus, with management indicating similar quarterly outlays for the next two quarters. First-half free cash flow works out near -$25.0B on $3.47B of operating cash flow. Shares had run 9.43% to $125.33 into the print, then closed 2026-08-05 at $108.27.

The supply event followed on 2026-08-06: 911.5 million shares, roughly $101B, taking public float from 4.9% to 11.8% of 13.18 billion shares outstanding. Price dipped to $105.11 in the first hour and closed $114.92, up 6.14%, at a $1.51T capitalisation. The conditional 455.8 million-share tranche stayed locked — its $175.50 condition was never within reach — and rolls into December.

The leg on offer now is narrower than the one that existed before the print. It is a business compounding revenue at 92% with $100B of cash against $39.4B of debt, whose largest first-tranche unlock in market history was absorbed at a low 28 cents above the all-time low, set against a capital plan that consumes that balance sheet faster than any pre-print estimate had it.

Life-cycle: SATURATED. The forced passive bid already came and went — Nasdaq-100 inclusion took effect 2026-07-07, and price has travelled from the $225.64 high to $114.92 since. Direxion listed LOFD, a daily 2X bear fund on SPCX, on 2026-08-06, the same session as the unlock, completing bull-and-bear leveraged pairs across both Musk listings. On that one day Whitney Tilson called it the most overvalued large-cap of all time at 92x trailing revenue, Peter Diamandis floated a $10T valuation, and Jim Cramer compared it to 100-year railroad bonds. Sell-side targets span $62 to $800 around a ~$223 average. Eight weeks after listing, the index membership, the leveraged packaging on both sides, and the two-sided celebrity commentary all exist; the marginal uncommitted participant is hard to locate. The label stays off DEAD because the operating record is intact: $47.5B backlog, $14.1B of cloud-services agreements signed in the quarter, and the 92% growth rate itself.

Bullish and bearish views on SpaceX (Space Exploration Technologies Corp.)

The model's bull view on SpaceX (Space Exploration Technologies Corp.) (SPCX), in brief: The print beat on both lines it could beat on. The bear view: Capex overshot consensus by 39%. $18.37B against ~$13.2B expected, with similar quarterly spend indicated for Q3 and Q4 — an implied full-year pace near $65B. First-half free cash flow near -$25.0B. The AI segment loses money at the operating line while it scales. $1.257B… Both cases follow in full.

Bull Case

• The print beat on both lines it could beat on. 2026-08-04: revenue $7.81B versus ~$6.93B consensus, a 13% beat; diluted loss $0.09 versus $0.26 expected; adjusted EBITDA $3.54B, +191% YoY, 45.3% margin; operating margin improved roughly 2,200 basis points to -1.8%.
• The largest first-tranche unlock on record did not produce a wall. 2026-08-06: low $105.11, close $114.92, +6.14%, against 911.5 million newly eligible shares. JPMorgan's Doug Anmuth attributed the absence of pressure to pre-positioning ahead of the date; Mizuho's Brett Linzey noted eligibility is not the same as offering.
• AI revenue is compounding off a real installed base. $2.561B in Q2, +247% YoY and +213% sequentially, on 1.4GW of nameplate compute versus 0.4GW a year earlier, with $14.1B of cloud-services agreements signed and a stated target of roughly 15GW by end-2027.
• Sell-side raised into the drawdown. JPMorgan reiterated Overweight and lifted its target from $225 to $240 on 2026-08-06, 122% above the prior close of $108.27. Morgan Stanley's Adam Jonas held Overweight and a $300 mid-2027 target, with more than half of it attributed to the AI segment. 27 of 29 ratings are buy-equivalent.
• Balance sheet absorbs the plan for now. $100B of cash and securities against $39.4B of total debt at 2026-06-30, up from $24.7B of cash at the end of Q1 after roughly $85.7B of net IPO proceeds and a $25B investment-grade issue.

Bear Case

• Capex overshot consensus by 39%. $18.37B against ~$13.2B expected, with similar quarterly spend indicated for Q3 and Q4 — an implied full-year pace near $65B. First-half free cash flow near -$25.0B.
• The AI segment loses money at the operating line while it scales. $1.257B operating loss on $2.561B of revenue in Q2; the launch segment lost $542M on $962M with $1.076B of R&D.
• Terafab adds a second capital sink. Confirmed for Grimes County, Texas on 2026-08-06: more than 100 million square feet, a $16.8B first phase with Tesla, Intel's 14A process, at least 3,000 jobs. A regulatory filing published in May had put the proposed initial investment at $55B; the March announcement carried $25B. The number keeps moving.
• Supply is a calendar, not an event. Third-party trackers place time-based tranches near 300 million shares each at the 70/90/105/120/135-day milestones through October, roughly 1.3 billion tied to the Q3 print, and the full 180-day block resolving 2026-12-08. The 455.8 million conditional shares that failed their August trigger join December.
• Valuation carries no index support. Tilson's 2026-08-06 arithmetic puts the stock at 92x trailing revenue; S&P 500 eligibility requires four profitable quarters, placing it no earlier than 2027, while Nasdaq-100 membership is already in the price.
• The short book is large and now cheaper to run. S3 Partners had 219.3 million shares short at 2026-07-29, about 34% of a ~640 million-share float and $24.6B — larger in dollar terms than Tesla's. Unlocked holders can lend, which eases the borrow scarcity that had been squeezing that position.

Setup & Price Structure

Four sessions define everything currently visible.

2026-08-03: record low $104.83, close $114.53, +5.68% on 70.4 million shares — an outside-day reversal into the print. 2026-08-04: close $125.33, +9.43%, results after the close, shares down roughly 7.5% by 18:05 ET. 2026-08-05: close $108.27, giving back the entire pre-print advance. 2026-08-06: low $105.11, close $114.92, +6.14%, on the day 911.5 million shares became sellable.

That leaves a two-point shelf at $104.83 and $105.11, with the second low set under maximum supply pressure — a higher low by 28 cents. The 52-week range runs $104.83 to $225.64; the $135 IPO price sits 17% above the last close, and the stock has spent every session since 2026-08-03 beneath it. There is no rising moving average to measure distance from; the name has trended down since July inclusion, and the entire base is three sessions old.

Positioning observables, stated as observables: short interest at 34% of float on the most recent S3 read, against a float that just more than doubled; a 2X bear ETF listed on the unlock date; open interest concentrating in $111 calls into 2026-08-06 per Benzinga's options coverage; weekly options into the 2026-08-07 expiry carried 163% implied volatility and a 14.32% implied move when priced on 2026-08-03. Insider distribution reports on a lag — Form 4s post within two business days of a sale, and no post-unlock filings appear in the record as of 2026-08-06.

Catalyst Calendar (next 30 days)

• ~2026-08-12 (est.) — second time-based lock-up tranche, roughly 300 million shares on third-party estimates at the 70-day milestone. Exact tranche dates are not in publicly circulated schedules. First test of whether the 2026-08-06 absorption repeats when supply arrives without headlines.
• 2026-08-14 — Q2 2026 13F deadline, the first institutional ownership snapshot that includes SPCX after the 2026-06-12 IPO. Checks the pre-positioning argument against the actual holder list.
• ~2026-08-26 (est.) — next tranche on the roughly two-week cadence running through October.
• ~2026-09-06 — no confirmed company event inside the window beyond the supply schedule; Q3 results are not expected before late October.

Elapsed catalysts

• 2026-08-07 — weekly options expiry carrying the 163% implied volatility and 14.32% implied move priced on 2026-08-03. Both events that volatility was written against have passed; the unwind lands in a float that only just doubled. (passed 2d ago)

What Would Change Our Mind

The structure is two lows 28 cents apart, set three sessions apart, and it is the only thing between the current quote and open air below the record low. Losing it says the 2026-08-06 absorption was pre-positioned inventory rather than fresh demand, and that the unlock resolved nothing. A weekly close below $105 is the gradeable form of that break.

The fundamental version is a Q3 capex print at or above $18.4B without combined connectivity and AI revenue clearing roughly $8.5B, which would show the spend decoupling from the ramp it is meant to fund. A funding event — an equity follow-on, or debt beyond the $25B already raised — reads the same way against $100B of stated cash.

On the other side, the case strengthens if the ~2026-08-12 tranche and the one after it pass with the $105 shelf intact on volume below 2026-08-06's, and if the 2026-08-14 13F filings show institutional size consistent with the pre-positioning explanation rather than a shareholder base that never formed. A close back above $125.33 would mean the post-print reversal has been fully retraced with the supply calendar still running — a different regime from the one described here.

Correlation Notes

• Tesla. The 2026-08-06 Terafab confirmation is a joint SpaceX/Tesla facility on Intel's 14A process; Musk's stated split for its output is roughly 25% Tesla Optimus and 75% AI spacecraft. Leveraged products already treat the two as one exposure — Leverage Shares' ELOL holds 100% TSLA plus 100% SPCX daily exposure, and Direxion's LOFF/LOFD now sit beside TSLL/TSLS.
• Nasdaq-100 complex. Membership since 2026-07-07 at a sub-1% weight ties the name to QQQ flows. The mirror products — Nasdaq-100 Ex-Elon (QQNE) and S&P 500 Ex-Elon (SPNE) — exist specifically to strip this correlation, which measures how concentrated the exposure has become.
• AI capex complex. The $18.37B quarterly print, $15.83B of it AI, lands in the same tape as hyperscaler spending commentary; the read-across runs to the compute supply chain rather than to aerospace peers. Coverage on 2026-08-06 framed the AI build as roughly 3x the telecom boom in scale.
• Anysphere/Cursor. The $60B all-stock merger expected to close in Q3 2026 sets its exchange ratio on a seven-day average share price at closing, so a depressed quote mechanically raises the share count issued.
• Intel. Named as the Terafab process partner in April 2026, which makes SPCX capital-plan headlines a second-order input to INTC foundry expectations.