# SPCX earnings call intelligence LOPJLB CallCard / temporal rollup · freemium · [stock page](https://www.lopjlb.com/stock/SPCX) · [Earnings tab](https://www.lopjlb.com/stock/SPCX?tab=earnings) Updated: 2026-10-20T05:17:56 Quarters analyzed: 1 ## Cross-quarter narrative SpaceX's Q2 2026 call revealed dramatic financial and operational scaling. Revenue surged 92% year-over-year to $7.8 billion, while adjusted EBITDA nearly tripled (+191%) to $3.5 billion, demonstrating strong operating leverage as the net loss narrowed. Starship V3 achieved 13 of 14 successful flights, with stage catch (booster and ship) targeted for year-end — a critical milestone for the 10x launch cost reduction thesis. Starlink added 1.7 million subscribers in its best quarter yet, reaching 167 markets with $66 ARPU, while mobile partnerships were delivered. The company is aggressively building AI compute capacity, targeting 2 GW by year-end and 10 GW by 2027, anchored by an exclusive NVIDIA Vera Rubin commitment. Grok model iterations (4.5 through 5) are on a rapid cadence, though only ~10% of compute is allocated to training. The Cursor acquisition remains pending regulatory closure. Capital intensity remains high across Starship infrastructure, Starlink V3 satellites, and AI clusters, though the balance sheet is described as fortified. Key uncertainties persist around Starship human-rating safety (targeting end-2027), Starlink Mobile terrestrial CapEx opacity, Cursor regulatory risk, Grok monetization, and single-source NVIDIA dependency. ## Latest CallCard · Q2 SpaceX Q2 2026: Revenue $7.8B (+92% YoY), adj. EBITDA $3.5B (+191%), net loss narrowed; Starship V3 flights 13/14 successful, targeting catch this year; Starlink added 1.7M subs (best quarter), ARPU $66, 167 markets; AI compute scaling to 2GW by year-end, 10GW by 2027; Grok 4.5/4.6/4.7/5 cadence; Cu **Guidance:** vague — No formal financial guidance provided; management outlined operational targets: Starship catch attempts this year, V3 Starlink deployment, 2GW compute by end-2026, ~10GW by end-2027, Grok 5 by year-end, Cursor close soon, Starlink Mobile service end of 2027. **Tone:** mgmt 0.8 · Q&A pressure 0.4 · divergence 0.4 Prepared remarks from Elon Musk and Gwynne Shotwell express high confidence in Starship reusability, Starlink V3 bandwidth scaling, AI compute expansion, and Grok model improvements, using superlatives like 'ridiculously profound difference' and 'extraordinary momentum'. ### Demand visibility Strong demand across Starlink consumer/enterprise/government, launch manifest, and AI compute from external customers. Starlink added 1.7M net subscribers in Q2 (best quarter), ARPU stable at $66, service in 167 markets; enterprise/gov revenue potential to exceed consumer; major airline deals (American, Southwest, Virgin Atlantic, Iberia, Aer Lingus); mobile partnerships with SoftBank, NTT Docomo, Spark; $6B+ US government contracts in Q2; AI compute demand from Google, Anthropic, and internal Grok training/inference. ### Margins / costs Adjusted EBITDA grew 191% YoY vs revenue +92%, showing operating leverage; Starship targets 10x launch cost reduction. Space segment revenue +29% YoY to $962M; segment costs rose $389M YoY with accelerated Starship R&D; Starship aims to quadruple payload capacity and reduce launch costs 10x vs Falcon 9; compute deployment efficiency claimed highest in industry; balance sheet strengthened via IPO and investment-grade bond offering. ### Capital allocation Investing heavily in Starship infrastructure, Starlink V3 satellites, AI compute (NVIDIA Vera Rubin), and Cursor acquisition; balance sheet fortified. Accelerating Raptor and launch vehicle production, Gigabay build-out, Starship launch infrastructure for thousands of launches/year; deploying V3 Starlink satellites on Starship; building 2GW compute by end-2026, targeting ~10GW by end-2027 exclusively on NVIDIA Vera Rubin; Cursor acquisition closing soon; terrestrial build-out for Starlink Mobile using EchoStar spectrum. ### Milestones - **Starship V3 Flight 13** [delivered]: Demonstrated core capabilities for orbital mission and return to Starbase for catch. - **Starship V3 Flight 14** [on_track]: Planned to fly Version 3 Starlink satellites to operational orbit. - **Starship stage catch (booster and ship)** [on_track]: Attempt to catch first and second stages this year; ship catch possible as early as next flight. - **Starlink V3 satellite operational deployment** [on_track]: Intend to deploy V3 satellites on upcoming Starship missions after successful laser link test. - **Starlink subscriber growth** [delivered]: Added net 1.7M subscribers in Q2, best quarter to date; ARPU stable at $66; 167 markets active. - **Starlink Mobile partnerships** [delivered]: Launched with SoftBank, NTT Docomo, Spark New Zealand; EchoStar spectrum transfer FCC approved. - **Grok 4.5 release** [delivered]: Released last month; enterprise feedback positive, token consumption tripled. - **Grok 4.6 release** [on_track]: Expected next week. ### Fears / risks - **Starship human-rating timeline**: Need super reliability before crewed flights; Elon targets end of 2027 for safety level, but execution risk remains. - **Starlink Mobile terrestrial CapEx uncertainty**: Gwynne Shotwell declined to disclose CapEx for terrestrial build-out; novel femtocell approach unproven at scale. - **Cursor acquisition regulatory risk**: Elon Musk cautioned about jumping the gun on regulatory closures; acquisition not yet closed. - **Grok compute allocation and monetization**: Elon expects only ~10% of compute for Grok training, rest for inference/rental; monetization signals early. - **Dependence on NVIDIA Vera Rubin supply**: Exclusive commitment to NVIDIA architecture creates single-source dependency for AI compute scaling. - **Starship production scaling execution**: Accelerating Raptor production, Gigabay build-out, and launch infrastructure for thousands of launches/year carries execution risk. - **Government contract concentration**: Significant revenue from US Space Force programs; policy or budget changes could impact. - **Competitive response in launch and satellite internet**: Elon encourages competitors but acknowledges Starship difficult to copy; market dynamics could shift. ### Key quotes > “We're making great progress developing a fully and rapidly reusable, reliable rocket with Starship.” — Elon Musk > “We go from 2,500 tons a year to 1 million, potentially 10 million tons per year. It's really a ridiculously profound difference.” > “I would say that we do not see any technical obstacles at this point to achieving full and rapid reusability.” > “The spectrum that we purchased from EchoStar does have terrestrial components. So we definitely intend to build out the terrestrial component.” — Gwynne Shotwell ## Quarter one-liners - **2026 Q2:** SpaceX Q2 2026: Revenue $7.8B (+92% YoY), adj. EBITDA $3.5B (+191%), net loss narrowed; Starship V3 flights 13/14 successful, targeting catch this year; Starlink added 1.7M subs (best quarter), ARPU $66, 167 markets; AI compute scaling to 2GW by year-end, 10GW by 2027; Grok 4.5/4.6/4.7/5 cadence; Cu ## Theme arcs - **Revenue and profitability growth** (improving): Revenue +92% YoY, adj. EBITDA +191% YoY, net loss narrowing - **Starship flight cadence and reusability** (improving): 13/14 V3 flights successful; stage catch targeted for 2026 - **Starlink subscriber and market expansion** (improving): 1.7M net adds (best quarter), 167 markets, $66 ARPU, mobile partnerships delivered - **AI compute infrastructure scaling** (new): Targeting 2 GW by year-end, 10 GW by 2027; exclusive NVIDIA Vera Rubin commitment - **Grok model development cadence** (new): Grok 4.5/4.6/4.7/5 sequence underway; ~10% compute for training - **Cursor acquisition integration** (new): Acquisition announced but not closed; regulatory closure risk noted - **Starship human-rating timeline** (new): Super reliability required before crewed flights; end-2027 safety target with execution risk - **Starlink Mobile terrestrial CapEx uncertainty** (new): CapEx for femtocell build-out undisclosed; novel approach unproven at scale - **NVIDIA supply dependency for AI compute** (new): Exclusive commitment to Vera Rubin creates single-source risk ## Fear persistence - **Starship human-rating timeline** [new]: Need super reliability before crewed flights; end-2027 safety target with execution risk - **Starlink Mobile terrestrial CapEx uncertainty** [new]: CapEx for femtocell build-out undisclosed; novel approach unproven at scale - **Cursor acquisition regulatory risk** [new]: Acquisition not yet closed; regulatory closure risk cautioned - **Grok compute allocation and monetization** [new]: Only ~10% compute for training; monetization signals early - **Dependence on NVIDIA Vera Rubin supply** [new]: Exclusive commitment creates single-source dependency for AI compute scaling ## Guidance path 2026 Q2:vague --- Research context only. Not personalized investment advice. API: `GET /bff/api/bigfive/earnings-intel/SPCX`