HeyEric

The FAA opened the paperwork for nearly-daily Starship. Cap is still 25. On Tuesday the FAA filed a Notice of Intent to prepare a Programmatic Environmental Impact Statement (PEIS) on SpaceX's request to raise Starship–Super Heavy operations at Starbase (Boca Chica, TX). Federal Register document 2026-20963, docket FAA-2026-12808, scheduled to publish Oct 14. Public scoping comments close Nov 30. Meetings: Brownsville Nov 17, virtual Nov 18, Port Isabel Nov 19. The numbers in the notice: Today's authorized ceiling: 25 orbital launches and 50 landings a year at Starbase (25 ship + 25 booster), under the 2025 tiered EA / mitigated FONSI. What the PEIS will study: up to 456 launches originating from the site and up to 912 landings there (landings can follow launches from elsewhere). Phase 1: up to 100 launches / 200 landings per year, any day, any hour. Phase 2: the full 456 / 912, "within several years" as the vehicle matures — any hour, weekdays only. Starship economics are not a "rocket" story. They are a cadence × reuse × payload story. Cost per useful ton falls when: 1. the same hardware flies again quickly (reuse), 2. the pad and range can clear another flight soon after (cadence), 3. each flight carries enough mass (payload). Hardware reuse without regulatory cadence is a museum piece. Regulatory cadence without reusable hardware is hot air. SpaceX has been closing the hardware side in public (booster catches, ship splashdowns, Ship 40 home for inspection, B23 in cryo). The missing throttle for Starbase itself has been the environmental authorization that sets how many times a year the site may launch and land. A PEIS is not a license amendment. Under NEPA it is the multi-year impact study that must finish before the FAA can lawfully authorize a much higher operating envelope. Phase 1 at 100/year is roughly a flight every ~3–4 days if you actually flew the cap. Phase 2 at 456 is roughly a weekday factory rhythm (~1.8/day on weekdays). That is the difference between a test site and an industrial launch machine. Why "programmatic": one review covering a stepped growth path and expanded landing/expenditure areas, instead of fighting a new EA every time cadence wants to climb. Why Phase 2 is weekday-only: noise, wildlife, and community load at near-daily rates — the FAA is already baking a social constraint into the maximum case. The real meter moves from "will they ask?" to "will the PEIS survive intact?" Scoping → draft PEIS → comments → ROD / license mod. Watch Phase 1 (100) more than the headline 456; 100 is the near-term industrial step. A company that files for ~18× the authorized cadence is arguing it intends to industrialize the site. The government putting that ask into a PEIS process is arguing it will study it seriously — not that it already said yes. What to watch next: 1. Federal Register publish Oct 14 — confirm final text matches the public-inspection numbers (456 / 912, Phase 1 100/200, Phase 2 weekday-only). 2. SpaceX first-party language — do they own the ask on @SpaceX, or leave it to the docket? 3. Flight 15 prep — ship-catch hardware vs NET Nov 8; any slip that uncouples reuse proof from the PEIS calendar. 4. Scoping heat — early Brownsville / Port Isabel reaction before the Nov 17–19 meetings. 5. Draft PEIS timing — any FAA hint on when the draft lands (the real gate after scoping). 6. Parallel: FSD 14.3.12 unify + Arrival Options — owned-car stack converging on Robotaxi behaviors; separate deep if utilization board needs it after earnings Oct 21.

Europe got the plinth. Austin got the miles. On Monday morning at Mondial de l'Auto (press day), Tesla staff peeled a grey cover off a gold production Cybercab on stand C25, Hall 7.1, Paris Expo Porte de Versailles. The car came from Tesla's U.S. Robotaxi fleet — same hardware that takes paid rides in Austin: no steering wheel, no pedals, solid disc covers, Cybercab script on the door. Optimus Gen 2 sat in a glass case a few steps away. Red production cars lined up behind. Visitors can walk around the plinth. They cannot ride it. What Europe can try this week is separate: six Model 3s for free FSD (Supervised) sessions — branded "Tesla Conduite Assistée" in France — 30–45 minutes each, public days Oct 13–18. That is supervised driving in a wheel-and-pedals car, not Cybercab. French outlet Turbo and show reporting say no European Cybercab service is planned at this stage. FSD (Supervised) is still not authorized for customer cars in France; Germany's push and a TCMV path aim at a broader EU decision as early as December. Same day, U.S. ops kept compounding: Texas registry prints 319 Cybercabs (~43% of Tesla's Texas robotaxi fleet after the Oct 9 batch), official @robotaxi lengthened Austin hours to 6am–11pm and claimed shorter waits from supply + orchestration, Dallas riders reported more cars in the second city, and NHTSA opened a defect-petition review (DP26010) on school-zone speeding with ADS engaged — review only, not a formal investigation; separate from the Oct 30 Cybercab FMVSS answers. Three different products sat on one stand. Confusing them is how investors get hurt. 1. Purpose-built robotaxi hardware — Cybercab: wheel-less two-seater designed for unsupervised paid miles. 2. Supervised ADAS product — FSD (Supervised) / Conduite Assistée: driver-in-loop on Model 3/Y. 3. Humanoid theater — Optimus in a case: adjacent thesis, not a ride. Paris put (1) on a plinth and sold (2) as a free demo. A plinth answers a perception and lobbying problem: European consumers, dealers, and regulators can stand next to a real U.S. fleet car instead of a render. It does not answer the permission problem. Wheel-less commercial service needs type approval and local operating authority. EU small-series rules for fully automated vehicles are narrow. France has not cleared customer FSD (Supervised), let alone Cybercab rides. So Tesla shipped a physical proof of manufacturing across the Atlantic while the legal path for miles stayed home. Austin answers a different equation: density × hours × wait SLA. 300+ Cybercabs serving, hours stretched into late night, Dallas supply rising — that is utilization. Paris is visibility. Visibility can soften regulators and seed demand. Utilization compounds cash and training data. Only utilization pays the path. Tonight measures whether Europe is next for miles or next for photos. The press-day evidence says photos — with FSD demos as the regulatory soft-sell. What this is not: EU L4, a France launch date, cabin Bot on the Paris plinth, or proof that Conduite Assistée rides convert into Cybercab law. FSD (Supervised) demos are the EU wedge, not Cybercab. Six Model 3s teaching Conduite Assistée is how Tesla builds political and consumer familiarity while type approval lags. Watch December TCMV / EU path harder than stand tweets. What to watch next week 1. Any French or EU official language at the show that moves Cybercab or FSD (Supervised) from "display" to "timeline." 2. Public-day report-outs (Oct 13–18) — are Conduite Assistée slots filling, and does anyone get cabin Bot or a Cybercab ride? (Expect no.) 3. Earnings Oct 21 — Europe Robotaxi language vs U.S. utilization / software-attach language. 4. Dallas Cybercab — in-app booking for purpose-built cabs, not only Model Y. 5. NHTSA Oct 30 Cybercab FMVSS answers + DP26010 school-zone review path (~120 days from Sept 23 petition). 6. Density board — Austin waits under longer hours; Texas registry vs cars actually serving.