
SpaceX
Three segments: Connectivity (Starlink subscription broadband, ~$66/mo ARPU, 10.3M subs), Space (commercial/government launch on Falcon 9/Heavy/Starship), and AI (Colossus data-center buildout). Highly vertically integrated; reusable rockets are the structural cost moat.
Sources — 22 figures with citations
- Q2 FY2026 revenue and year-over-year growthfiled2026-06-30$7,814M, +92% (vs $4,071M in Q2 2025)s21.q4cdn.com — Consolidated Statements of Operations, quarter ended June 30, 2026, in the company's Q2 2026 results release issued August 4, 2026.
- Segment revenue splitfiled2026-06-30Connectivity $4,291M / AI $2,561M / Space $962Ms21.q4cdn.com — Second Quarter Financial Highlights table. Mix works out to 54.9% Connectivity, 32.8% AI, 12.3% Space of the $7,814M total.
- Segment operating income (loss)filed2026-06-30Connectivity +$1,656M / AI ($1,257M) / Space ($542M); total ($143M)s21.q4cdn.com — Income (Loss) from Operations table. Connectivity operating margin = 1,656 / 4,291 = 38.6%; consolidated = -143 / 7,814 = -1.8%.
- Gross marginderived2026-06-30~55.3%s21.q4cdn.com — Revenue $7,814M less cost of revenue $3,495M = gross profit $4,319M; 4,319 / 7,814 = 55.3%. SpaceX does not report a gross profit subtotal, so this excludes R&D ($3,548M) and SG&A ($912M).
- Net loss and EPSfiled2026-06-30($541M); ($0.09) basic and diluted on 5,864M weighted average sharess21.q4cdn.com — Improved from a ($1,008M) net loss and ($0.34) per share in the year-ago quarter.
- Adjusted EBITDAfiled2026-06-30$3,538M, +191% year over years21.q4cdn.com — Reconciliation of Net Loss to Adjusted EBITDA. Non-GAAP; adds back $2,848M of D&A and $831M of share-based compensation among other items. Margin = 3,538 / 7,814 = 45.3%.
- Capital expenditure and capex intensityderived2026-06-30$18,369M in the quarter (AI $15,828M), equal to 235% of revenues21.q4cdn.com — Capex table gives $18,369M total ($15,828M AI, $1,367M Connectivity, $1,174M Space); 18,369 / 7,814 = 235.1%. First-half capex was $28,476M.
- Free cash flow, first half 2026derived2026-06-30approximately ($25.0B), about -200% of first-half revenues21.q4cdn.com — Selected Cash Flow Information: six-month operating cash flow $3,466M less six-month capex $28,476M = ($25,010M); divided by six-month revenue of $12,508M = -200%. SpaceX discloses cash flow only for the six-month period, so a quarter-only figure cannot be computed.
- Net cash positionderived2026-06-30~$60.6B net cashs21.q4cdn.com — Consolidated Balance Sheet: cash and equivalents $93,522M plus marketable securities $6,487M = $100,009M; less debt and finance leases of $2,525M current plus $36,839M non-current = $39,364M; net = $60,645M. Related-party debt within that total is $2,039M current plus $11,290M non-current = $13,329M.
- Starlink subscribers and ARPUfiled2026-06-3012.0M subscribers (double a year ago, +1.7M sequentially); ARPU $66 per months21.q4cdn.com — Connectivity Segment Operating and Financial Data. ARPU flat versus the March quarter's $66 and down from $85 in the June 2025 quarter. 10.2K satellites in orbit across 167 countries.
- Launch cadence and mass to orbitfiled2026-06-3038 launches in the quarter (10 customer, 28 internal); 78 launches and 1,041 metric tons in the first halfs21.q4cdn.com — Space Segment Operating and Financial Data. Down from 46 launches in the June 2025 quarter and 40 in March 2026; first-half comparison is 78 launches / 1,041 t against 84 launches / 1,102 t a year earlier.
- Starship V3 flight progressfiled2026-08-04Flight 12 in May, Flight 13 in July (20 production V3 satellites deployed, in-space Raptor relight, intact heatshield)s21.q4cdn.com — Space Business Highlights. Flight 13 occurred after quarter end and is disclosed as a subsequent event; the release states two successful Starship V3 flight tests in the past 90 days.
- AI compute capacity and contracted cloud servicesfiled2026-06-301.4 GW nameplate compute; $14.1B of cloud services agreements signed, $1.6B of incremental AI infrastructure revenue in the quarters21.q4cdn.com — AI Segment Operating and Financial Data plus AI Business Highlights. Compute rose from 1.0 GW in March 2026 and 0.4 GW in June 2025, with Colossus II still building out.
- Cloud services counterpartiesfiled2026-08-04Anthropic and Google; a further $6.7B under contract ramping from Octobertechcrunch.com — Counterparties named in reporting on the earnings call; the $6.7B figure is CFO Bret Johnsen's remark on the call. SpaceX's own release discloses the $14.1B aggregate without naming customers or per-customer amounts.
- Form 10-Q for the quarter ended June 30, 2026filed2026-08-04Filed August 4, 2026 (CIK 1181412); backlog $47,461M including $14,286M deferred revenue; 18,712 bitcoin at $1,098M fair value; $354M litigation accrualsec.gov — SpaceX's quarterly report on Form 10-Q, filed the same day as the results release.
- Share price closemarket2026-08-07$133.11 on August 7, 2026stockanalysis.com — Official Nasdaq close, 4:00 PM ET Friday August 7, 2026. Independently confirmed at the same value by MarketBeat (https://www.marketbeat.com/stocks/NASDAQ/SPCX/). Not an intraday or 52-week figure.
- Market capitalisationmarket2026-08-07~$1.75Tstockanalysis.com — Reported at $1.75T by StockAnalysis and $1.74T by MarketBeat on the August 7 close. Cross-checks against the balance sheet: 7,607M Class A plus 5,569M Class B = 13,176M shares outstanding at June 30; 13,176M x $133.11 = $1.754T. Implies ~56x the annualised June-quarter revenue ($7,814M x 4 = $31.3B).
- Consensus versus reportedmarket2026-08-04Revenue $7.814B vs ~$6.81B expected; EPS ($0.09) vs ($0.26) expected; capex $18.4B vs ~$13B expectedmarketbeat.com — EPS consensus of ($0.26) per MarketBeat; the ~$6.81B revenue and ~$13B capex expectations are FactSet-polled analyst figures reported by Quartz (https://qz.com/spacex-earnings-revenue-q2-2026-080426) and CNN (https://www.cnn.com/2026/08/04/business/spacex-earnings-q2-2026). These are analyst polls, not company or exchange disclosures.
- Market reaction to the printmarket2026-08-05Fell as much as 8% after hours on August 4; closed $109.27 on August 5, down about 13% from $126.12stockanalysis.com — Daily closes: $115.17 (Aug 3), $126.12 (Aug 4, results after the close), $109.27 (Aug 5), $115.92 (Aug 6), $133.11 (Aug 7). Aug 5 decline derived as 109.27 / 126.12 - 1 = -13.4%. After-hours move reported by CNN.
- Terafab semiconductor plantmarket2026-08-06$16.8B first phase announced August 6, 2026 with Tesla, in Grimes County, Texas; total phases indicated up to $119Btechcrunch.com — Announced two days after the results. SpaceX's IPO filing characterised Terafab as a non-binding general framework leaving either party free to withdraw. A principal driver of the August 7 share price move, alongside analyst upgrades and the passing of the first lock-up expiry.
- Capital structure events in the quarterfiled2026-06-30IPO closed June 15, 2026 (638,888,888 Class A shares at $135.00, net proceeds ~$85.7B); $25B bond issuance closed June 26 at a 5.855% weighted average rates21.q4cdn.com — Other Key Business Developments. Notes issued in five tranches maturing 2031, 2033, 2036, 2046 and 2056 at coupons between 5.35% and 6.65%. Trading began on Nasdaq June 12, 2026.
- Forward operational targets from the earnings callmarket2026-08-04Capex in the next two quarters 'very similar' to $18.4B; 2 GW compute by end-2026, ~10 GW by end-2027; Starship Flight 14 late Augustnotateslaapp.com — Management remarks on the August 4 call. These are operational targets, not formal financial guidance; SpaceX issued no revenue or EPS guidance range.
The thesis on this name
Space / Launch Economics
Space Exploration Technologies (SPCX) — the world's dominant launch provider (Falcon 9 / Starship) and operator of Starlink LEO broadband, now also running a large 'AI' data-center segment (Colossus); freshly public on Nasdaq as of 12 Jun 2026.
State of the AI Cloud
SpaceX (SPCX) — now public, with the AI segment filing $2.561B in the June 2026 quarter (~$10.2B annualised). The valuation leg has substantially paid: $133.11 at the 7 Aug 2026 close, ~$1.75T, down ~28% from the ~$185 that framed the entry, re-basing the multiple to ~56x annualised June-quarter revenue from ~87x.
State of the AI Cloud
Now public (largest IPO in history) with the AI segment filing $2.561B in the June quarter (~$10.2B annualised) — the valuation leg has largely paid at ~56x annualised revenue, leaving cash burn, not the multiple, as the short.
State of the AI Cloud
The ~87x sales multiple that framed the call has re-based to ~56x after a ~28% fall; a tiny valuation-short/avoid now carried for the cash-burn and cancellable-backlog risk, marked to the 7 Aug 2026 close.
Earnings, margins, COGS & capex
FY2025 revenue $18.7B (+33%) with a consolidated operating loss of ($2.6B) and net loss of ~$4.9B. Starlink (Connectivity) is the profit engine — $11.4B revenue (+50%), ~$7.2B segment EBITDA — but it is more than offset by the AI segment's ($6.4B) operating loss and ~$20.7B of total capex. Q1 FY26 deepened the loss: revenue $4.69B (+15.4%), operating loss ($1.94B) vs a $27M profit a year prior, net loss $4.28B, EPS ($1.27). The story is a cash-generative broadband business funding two enormous capital sinks (Starship + AI data centers).
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~45¢ is cost of goods and ~55¢ operating expense, and the remainder is an operating loss.
Revenue trend
Margins
improving — segment EBITDA +86% YoY, op income +120% YoY
worsening — Q1 FY26 op loss $1.94B vs $27M profit prior-year quarter
positive at adj-EBITDA level but masks AI capex burn
COGS structure
Not disclosed at line-item level. Launch COGS driven by Falcon 9 reusability (marginal cost far below expendable peers); Starlink COGS dominated by satellite manufacture + user-terminal subsidy (ARPU fell from ~$81 in 2024 to ~$66 in Q1 FY26 as the mix broadened); AI segment COGS dominated by data-center power + GPU depreciation.
Capex
FY25 total ~$20.7B (AI $12.7B / Connectivity $4.2B / Space $3.8B). Q1 FY26 $7.7B–$10.1B in a single quarter (~3/4 to AI). Annualized AI burn implied north of $30B. This is the single most important number in the story — it is why a profitable Starlink does not produce a profitable SpaceX.
Latest earnings
A large beat on the income statement and a large miss on the spending line. Revenue of $7.814B came in roughly $1.0B above the ~$6.81B analysts had modelled, and every segment cleared its bar: Space $962M against $835M expected, Connectivity $4.291B against $3.83B, AI $2.561B against $2.18B. The loss per share of $0.09 beat the $0.26 loss consensus by $0.17, and Connectivity's 38.6% operating margin beat the 35.9% expected. But capital spending of $18.369B landed roughly 40% above the ~$13B modelled, and that is what the market traded on: shares fell as much as 8% after hours and closed the next session at $109.27, down about 13% from $126.12. The stock then recovered across August 6-7 to $133.11 on the Terafab chip-fab announcement, analyst upgrades, and relief that the first insider lock-up expiry passed without incident.
No formal revenue or earnings guidance range was issued. Management gave operational targets instead: capital spending in each of the next two quarters will be 'very similar' to the June quarter's $18.4B; nameplate compute reaches 2 GW by the end of 2026 and around 10 GW by the end of 2027; a further $6.7B of cloud services revenue is already contracted and begins ramping in October. Starship Flight 14 is set for late August, targeting the first operational deployment of V3 Starlink satellites plus an upper-stage catch attempt. On Starlink, roughly 1,000 V3 satellites are needed to unlock the next service tier, which management put around the second quarter of 2027, with standalone Starlink Mobile service targeted for end-2027. Elon Musk framed a $100B annualised revenue run-rate by December as achievable and pointed to an internal ambition of $1T of annual revenue by 2030 - aspirations voiced on the call, not committed guidance.
- Revenue
- $7.814B, +92% year over year
- Net loss / EPS
- ($541M) and ($0.09) per share, versus ($1.008B) and ($0.34) a year earlier
- Adjusted EBITDA
- $3.538B, +191% year over year, a 45.3% margin
- Connectivity (Starlink) revenue
- $4.291B, +66% year over year, 55% of total revenue
- Connectivity operating income
- $1.656B, +79% year over year, a 38.6% operating margin - the only profitable segment
- AI segment revenue
- $2.561B, +247% year over year and +213% sequentially, on new cloud services agreements
- AI segment adjusted EBITDA
- +$1.146B, positive for the first time, versus ($609M) in the March quarter
- Space (launch) revenue
- $962M, +29% year over year, but an operating loss of ($542M)
- Capital expenditure
- $18.369B in the quarter, of which $15.828B was AI infrastructure
- Starlink subscribers
- 12.0M at quarter end, double a year ago and up 1.7M sequentially, across 167 countries
- Starlink ARPU
- $66 per month, flat sequentially, down from $85 a year earlier
- Satellites in orbit
- 10.2K
- Launches
- 38 in the quarter (10 customer, 28 internal); 78 in the first half carrying 1,041 metric tons to orbit
- Nameplate compute
- 1.4 GW, up from 1.0 GW in the March quarter and 0.4 GW a year ago
- Contracted cloud services
- $14.1B of contracted sales signed in the quarter, contributing $1.6B of incremental AI infrastructure revenue
- Backlog
- $47.5B, including $14.3B of deferred revenue
- Cash and liquidity
- $100.0B of cash, equivalents and marketable securities at June 30, 2026
Growth drivers
- Starlink subscriber growth (10.3M subs, +>100% YoY, 164 countries) and direct-to-cell / enterprise / maritime / government expansion
- Starship reaching operational cadence — unlocks Starlink V3 mass-to-orbit and a new launch cost curve
- Government/defense backlog (NASA Artemis HLS, Space Force PWSA Tranche 1, Rocket Cargo)
- Third-party launch demand (incl. launching competitor Amazon Leo/Kuiper)
- AI segment (Colossus / Colossus II) compute monetization — currently a loss leader
Bull & bear
A company approaching $31B of annualised revenue grew 92% and all but eliminated its operating loss in the same quarter - and the segment that was the largest cash sink a year ago just turned adjusted-EBITDA positive. Starlink is now a genuine profit machine, the compute business has found paying customers at scale, and a $100B cash position funds the build-out without an immediate return to the market.
- Growth accelerated at scale rather than decaying. Revenue rose 92% year over year to $7.814B, up from $4.694B just one quarter earlier. Very few companies at a $30B run-rate post triple-digit-adjacent growth, and it came from all three segments at once.
- Operating losses have nearly closed. The operating loss was $143M on $7.814B of revenue, versus $970M on $4.071B a year ago - a swing from -23.8% to -1.8% of revenue. Adjusted EBITDA of $3.538B was up 191%, a 45.3% margin.
- Starlink is a high-margin, compounding franchise. Connectivity delivered $4.291B of revenue (+66%), $1.656B of operating income (+79%) and a 38.6% operating margin, with subscribers doubling to 12.0M and a record 1.7M net additions in the quarter. The higher-value enterprise and government line grew 108% year over year to $1.806B - a mix shift toward the better revenue.
- The AI segment flipped from cash sink to contributor in two quarters. AI adjusted EBITDA came in at +$1.146B against ($609M) in the March quarter and ($276M) a year ago, as capacity was redirected from training in-house models to renting compute to outside customers - a change management described as carrying high incremental margins.
- Demand is contracted, not hoped for. The quarter added $14.1B of cloud services agreements (Anthropic and Google among the counterparties), with a further $6.7B under contract that begins ramping in October, plus more than $6B of multi-year Starshield awards from the US government. Total backlog stands at $47.5B.
- The balance sheet can fund the plan. $100.0B of cash and marketable securities against $39.4B of debt leaves roughly $60.6B of net cash, assembled from $85.7B of IPO proceeds and a $25B investment-grade bond issued in June at a 5.855% weighted average rate - long-dated money out to 2056 at pre-AI-boom pricing.
- Starship is clearing real technical gates, not slipping. Flight 12 in May achieved a precision landing of the upper stage; Flight 13 in July hit every objective, deploying 20 production V3 satellites, relighting a Raptor engine in space, and returning an intact heatshield - the specific milestone that rapid reusability depends on. Flight 14, targeted for late August, attempts the first operational V3 deployment plus an upper-stage catch.
- The direct-to-cell position is becoming hard to attack. FCC approval of the EchoStar transfer brings 65 MHz of US spectrum plus global mobile satellite licences, layered on top of new agreements with American Airlines, Southwest, Virgin Atlantic, Iberia and Aer Lingus, and carrier partnerships with SoftBank, NTT Docomo and Spark NZ.
The improving income statement and the cash reality are moving in opposite directions. Capital spending hit 235% of revenue, free cash flow was around -$25B for the half, and the narrowed loss is flattered by depreciation that has not yet arrived. Meanwhile the launch business the company is named for is shrinking and losing money, Starlink's price per subscriber is falling, and capital is being committed through an unusually related web of counterparties - all against a valuation near 56 times the annualised run-rate.
- The spending is the story, and it is accelerating. Capex was $18.369B in a single quarter - 235% of revenue, and roughly 40% above what analysts modelled. Management guided the next two quarters to be 'very similar', which points to something near $65B for the full year against roughly $30B of revenue. The 13% one-day drop after the print was a verdict on exactly this.
- Free cash flow is deeply negative and the reported loss understates the burn. First-half operating cash flow of $3.466B against $28.476B of capex implies roughly -$25.0B of free cash flow, about -200% of first-half revenue. A $541M net loss and a $25B cash outflow describe the same three months.
- A depreciation wave is queued behind the current numbers. Depreciation and amortisation was $2.848B in the quarter, already up 87% year over year, while net property, plant and equipment grew $23.1B in six months to $65.736B. Much of that capital is not yet in service; as it is placed in service, the charge steps up against revenue that has to grow into it.
- The launch business is now the smallest, shrinking and least profitable segment. First-half Space revenue of $1.581B was down from $1.611B a year earlier, on 78 launches versus 84 and 1,041 metric tons to orbit versus 1,102. Its first-half operating loss widened to $1.204B from $439M. Launch is 12% of revenue and getting smaller.
- Starlink's growth is being bought at a lower price. ARPU is $66 per month, flat sequentially and down 22% from $85 a year ago. Subscribers doubled while Connectivity revenue grew 66% - the gap is the discount required to add the next cohort, and it compounds as the base shifts to cheaper tiers and geographies.
- AI revenue is concentrated and its durability is untested. Of $2.561B in AI revenue, $2.194B is solutions and infrastructure driven by a handful of very large compute tenants. Advertising, the older and more diversified line, actually declined year over year to $367M from $426M. Two quarters of compute-rental revenue is not yet a proven annuity, and these customers are building their own capacity.
- Capital is being committed through related parties at very large scale. The balance sheet carries $13.3B of related-party debt with $327M of related-party interest expense in the quarter alone. On top of that sit a $60B agreement to acquire Cursor - roughly 7.7 times the entire quarter's revenue, expected to close in the third quarter - and the $16.8B first phase of the Terafab chip plant announced jointly with Tesla on August 6, a project SpaceX's own IPO filing described as a non-binding general framework that either party may walk away from. Total Terafab phases have been indicated at up to $119B.
- The valuation leaves no room for a stumble. At the August 7 close of $133.11, the roughly $1.75T market capitalisation is about 56 times the annualised June-quarter revenue, for a business that is still loss-making at the operating line, has two quarters of public reporting history, and is guiding to spend more cash in the next six months than it will earn in revenue.
What it is worth
Reverse-DCF / multiple sanity-check on a ~$2.0T market cap (price ~$153, Jun 2026). At ~$19.3B LTM revenue the implied P/S is ~104 (Bloomberg) to ~131 (recent), vs a sub-30 durable ceiling even for game-changing companies — so the price already discounts Starlink scaling globally AND Starship/AI succeeding. To justify ~$2T at a normalized ~6-8x sales, revenue must reach ~$250-330B, implying SpaceX must hold ~40%+ revenue CAGR for a decade while turning the AI/Starship capex into FCF. The reverse-DCF demands both durable Starlink margin expansion (~63% segment EBITDA holding as ARPU compresses) and the AI segment flipping from a ($6.4B) loss to material profit.
~$62 (low analyst)
AI capex keeps destroying value, Starship slips, rivals compress Starlink pricing, and P/S mean-reverts toward sub-30 — a multiple de-rate of 50-60%+ from spot.
~$188 (consensus mean
Buy): Starlink carries the story, launch dominance persists, AI burn moderates; supports today's ~$2T zip-code with volatility.
~$310 (high analyst)
Starship operational, Starlink compounding 40%+ with margin intact, AI compute monetizes — three platforms re-rate together.
The market is pricing near-flawless execution across three capital-intensive verticals simultaneously; analyst targets span $62 (low) to $310 (high) around a ~$188 mean, the widest dispersion of any mega-cap — a signal the outcome is genuinely undetermined. Near-term, the thin ~4% float + staggered insider lockup (up to 20% unlocking late Jul/Aug) make price a poor read on intrinsic value.
SWOT
Strengths
- Structural launch monopoly — Falcon 9 reusability gives a cost-per-kg moat no Western rival matches; ~140 launches/yr planned, more than the rest of the world combined
- Starlink is the only at-scale, profitable LEO broadband network (10.3M subs, ~63% segment EBITDA margin) with a multi-year deployment lead
- Deep vertical integration (engines, avionics, satellites, terminals in-house) compresses cost and supplier dependency
- Entrenched government/defense relationships (NASA Artemis HLS, Space Force PWSA, classified launch) create a sticky, high-barrier backlog
- Brand + talent gravity and ~$15.9B cash to fund the buildout
Weaknesses
- Consolidated GAAP losses — FY25 net loss ~$4.9B, Q1 FY26 ($4.28B); profitability is years out
- Capex intensity >110% of revenue and ~-$9.1B quarterly FCF — the model depends on continuous capital raises
- AI segment burns ~$30B/yr annualized with an unproven monetization path — it converts a profitable Starlink into a loss-making whole
- Space/launch revenue is near a mature run-rate and fell ~28% YoY in Q1 FY26 as assets shift to Starship
- Starlink ARPU compression (~$81→$66) as the mix broadens
- Heavy reliance on Elon Musk; key-person and governance/dilution risk flagged in the S-1
Opportunities
- Starship operational cadence unlocking Starlink V3 economics and a step-change in $/kg
- Direct-to-cell, enterprise, maritime, aviation and sovereign-connectivity expansion of Starlink
- Defense/space-architecture spend (PWSA, Rocket Cargo, lunar/Mars logistics)
- Monetizing AI/Colossus compute if utilization and pricing materialize
- Launching third-party constellations (incl. Amazon Leo) — revenue from competitors
Threats
- Well-capitalized constellation rivals — Amazon Leo/Kuiper (367 sats, FCC half-deploy deadline Jul 2026), Blue Origin TeraWave (5,408 sats from Q4 2027)
- China's Spacesail/SatNet state-backed mega-constellations (15,000 sats by 2030) — geopolitical + spectrum competition
- Valuation air-pocket — P/S ~104-131 is historically unsustainable; staggered lockup unlocks (up to 20% of insider shares from late Jul/Aug) with only ~4% float
- Starship technical risk (5 of 12 flights failed as of May 2026) — schedule slips ripple to Artemis and Starlink V3
- Regulatory/spectrum, orbital-debris and antitrust scrutiny as a dominant launch+broadband incumbent
- Key-person risk on Musk; capital-markets risk if rates/sentiment turn against cash-burning growth
Moats, dependencies & bottlenecks
Moats
High (multi-year) No Western rival matches cost-per-kg or cadence; ~140 launches/yr planned, more than the rest of the world combined.
10.3M subs and a deployed constellation give years of lead, but Amazon Leo / Blue Origin / China are now deploying.
In-house engines, avionics, satellites, terminals compress cost and reduce supplier leverage.
NASA Artemis HLS, Space Force PWSA Tranche 1, Rocket Cargo, classified launch — high switching barriers.
Lets it outspend rivals, but depends on continued capital-markets appetite for the burn.
Dependencies
~-$9.1B quarterly FCF and ~$20.7B FY25 capex require ongoing issuance; S-1 explicitly flags debt/equity for AI + M&A (dilution).
Technical/Execution 5 of 12 flights failed (May 2026); Artemis HLS, Starlink V3 economics, and the bull case all ride on operational cadence.
Customer concentration NASA + DoD are anchor customers; appropriations or policy shifts move the backlog.
Vertically integrated but still sources custom chips/RF/interconnects (e.g. STMicroelectronics co-designed Starlink silicon).
Governance/Regulatory Key-person concentration plus FCC/ITU spectrum and orbital-debris regimes.
Advantages
- Lowest cost-per-kg to orbit and highest launch cadence in the world
- Only profitable at-scale LEO broadband network
- Deep vertical integration
- Sticky government/defense backlog
- $15.9B cash + capital-markets access to fund the buildout
Weaknesses
- Consolidated GAAP and FCF losses driven by AI capex
- Valuation (P/S ~104-131) historically unsustainable
- Thin ~4% float + staggered insider lockup → volatility/dilution
- Key-person concentration on Musk
- Eroding broadband moat as rivals deploy
Bottlenecks
- Starship reliability + cadence ramp
- Free-cash-flow burn vs capital-markets access
- Starlink ARPU compression as the subscriber mix broadens
- Launch revenue at a mature run-rate until Starship scales
- AI segment monetization (utilization/pricing of Colossus compute)
Top signals & trends
Top signals
Costs +64% YoY; profitability of the whole entity remains distant.
The cash engine is compounding and globally under-penetrated.
Bear (valuation) · Enormous dispersion = the market has not priced the AI/Starship outcome.
Bear (technical) · Supply overhang into thin liquidity is a near-term volatility catalyst.
Progressing but not yet operational; binary near-term risk to the thesis.
Trends
Validates the market but erodes Starlink's pricing/share moat over time; spectrum + geopolitics intensify.
SpaceX is the driver and primary beneficiary; widens its lead and expands addressable demand.
SpaceX's AI/Colossus bet rides the wave but is the source of the loss; outcome hinges on compute monetization.
PWSA, Rocket Cargo, resilient comms — durable government demand.
Expands Starlink TAM into mobile/handset connectivity.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Decade-long partner; co-designs custom chips for Starlink satellites, terminals, gateways — billions of components shipped. (Non-US/EU-domiciled; named as analysis, not a recommendation.)
Analog/embedded semiconductors used in space-grade electronics.
Interconnects / RF connectors for satellite and launch hardware.
Connectors and sensors for aerospace/satellite systems.
Avionics, GNC sensors, and aerospace components.
Advanced carbon-fiber/composite materials for aerospace structures.
GPUs for xAI's Colossus AI data centers (Colossus 1 = 220k+ GPUs) — the compute SpaceX leases to Anthropic, Google and Reflection.
Artemis Human Landing System (Starship HLS), crew/cargo to ISS, science launch — anchor customer.
PWSA Tranche 1 Transport Layer launches; $102M Rocket Cargo program; national-security launch.
Bought SpaceX launches to deploy its own competing constellation — a competitor that is also a customer.
Starlink end-customers (consumer/enterprise/maritime/aviation/government) 10.3M subscribers across 164 countries; the core recurring-revenue base.
2nd-largest pure-play space co (~$68.6B mkt cap, Jun 2026); Electron flying, reusable Neutron in development; LTM revenue ~$680M vs SpaceX ~$19.3B. P/S ~123.
Direct-to-cell LEO; ~$34B mkt cap; pre-revenue scale, P/S ~409 — direct competitor to Starlink direct-to-cell.
367 production sats by Jun 2026; FCC half-deploy deadline Jul 30 2026. Best-capitalized Starlink challenger — and ironically a SpaceX launch customer.
~$34B mkt cap; satellite + spectrum holdings; legacy/MSS connectivity competitor.
GEO/MEO satellite broadband incumbent losing share to LEO; competes for enterprise/government connectivity.
Established LEO narrowband/IoT/voice; niche vs Starlink broadband but overlaps in IoT + government.
Bezos-funded; New Glenn launch + 5,408-sat TeraWave constellation from Q4 2027; deep pockets, behind on schedule.
Boeing/Lockheed JV; Vulcan launch; primarily a launch provider, key alternative for national-security launch.