Gonçalo Esteves / building capable systems Field Notes →

The Vault / Companies

Company

SpaceX

the launch and satellite company the Moonshots with Peter Diamandis panel cites in 75 of 88 indexed episodes and had no page for. Starship's first orbital flight is on Moonshots 298 — Why AI Leaders Have Changed Their Minds About AI Safety, Elon on UHI, Anthropic's IPO.

status · researched added 6 Oct 2026 updated 16 Nov 2026 from Moonshots #298 · Moonshots #304 · Moonshots #305 · Moonshots #306 · Moonshots #308 · Moonshots #309 · Moonshots #310 · Moonshots #311 organizationspaceoperator-life see it in the graph →

What it is (one line): the launch and satellite company the Moonshots with Peter Diamandis panel cites in 75 of 88 indexed episodes and had no page for. Starship's first orbital flight is on Moonshots #298.

Why it is in this archive

Gate: cited in 75 episodes. Promoted 2026-10-06 off the EP298 decomposition, where it is the whole space segment (1:35:49–1:44:30, 38 mentions in the episode).

Facts (EP298, checked 2026-10-06)

claim on air on tape against the record
Flight 14 reached orbit "Monday" 1:35:49 ✅ launched 2026-09-28 from Starbase, the first orbital Starship flight (CNN live)
Ship 41 "lost one of its Raptor engines on the way up, it finished its orbital burn anyway", two orbits, soft landing in the North Pacific near Hawaii 1:36:00 Engine loss ✅ (an RVac shut down at T+7:19; launch control polled and went for orbit). "Two orbits" and "near Hawaii" not re-checked (Wikipedia, secondary)
Deployed "all 26 of its operational V3 Starlink satellites" 1:36:11 ✅ 26 V3 satellites, 262×275 km, 30.5° inclination (same sources)
"each Starlink satellite carries more than a terabit per second of down link" → "60 of these" per launch = "20 Falcon 9 launches" 1:37:10 26 × 1 Tb/s = 26 Tb/s matches the reported "26 terabits per second" (Tesla Oracle, secondary). "60 per launch" and "20 Falcon 9s" are Peter's projection, not checked
"8 petabytes per second in orbit by 2028" on "150 Starship launches" = "10 times all of Starlink's bandwidth today" = "all global internet traffic in 2025" 1:37:30–1:37:41 ⚠️ One analyst's estimate (Aaron Burnett, caption "Bernett"), "conservative" is his word. Not checked. Do not repeat as a SpaceX plan
Elon: "increasingly probable that Starlink will carry a majority of Earth's IP traffic long-term" 1:37:51 ⚠️ his post, an opinion; post not opened
"next time we're going to catch Starship at Starbase" 1:36:49 Peter's "if the prediction holds"; no date given

What the panel said (not facts)

  • Alexander Wissner-Gross: Starship now "delivering economic value to low Earth orbit" for the first time (1:38:58); the capacity is "an over supply in search of an under demand" unless orbital data centres fill it (1:39:43); "go help Elon and SpaceX saturate all of their capacity" (1:40:35). Also publicly shames Delta for not adopting Starlink (1:38:13).
  • Emad Mostaque: not bullish on orbital data centres (1:41:30); a Starlink backbone as the resilient layer against agent-swarm attacks (1:41:50).
  • Q&A (2:17:53, Alex): "Should we be worried about Elon having a monopoly of an entire civilizational stack?" — "Absolutely not"; the worry is "not enough people following Elon's example". Peter agrees, "I could not be more proud of him" (2:19:06).

What an operator can take from it

  • The orbit Ship 41 deployed into (262×275 km) sits in the thermosphere drag regime, where density sets lifetime — relevant to anyone working on orbital-density measurement. ⚠️ My inference from the figures above; not something the panel said.

Open

  • SpaceX's own page for Flight 14 still not readable (spacex.com mission pages return a title only, re-tried 17 Oct). "Two orbits" is still unconfirmed; the landing site has a trade-press source only.
  • Flight 15 (flew 13 Nov; ship caught 14 Nov, booster lost): the root cause of the booster's short boostback is still under investigation (SpaceX, 14 Nov). Still not found: SpaceX's own mission page (renders empty), the FAA licence on the record (FAA page 403), and any FAA word on whether the booster loss is a mishap. No Flight 16 date.
  • Starship V4 (a larger ship, "about 70 m, nine engines"): one finance-site report of a Musk statement on 16 Nov; the statement itself not opened. Not carried as a fact.
  • The Apollo GPU lease: watch for pricing (the fixed rate) and the close, both expected before year end per the 26 Oct 8-K. Not closed by 10 Nov.
  • The board seat Donald Harrison vacated (8-K, 6 Nov): watch for a successor filing.
  • The Boost Mobile talks (WaPo via CNBC, 2 Nov) and the Qualcomm deal's terms: both unquantified; watch for an 8-K or either company's release.
  • Q3 results date: settled by the 20 Oct 8-K (read); the 16 Oct investor release itself was never opened.
  • Pages for Starlink, Starship and Starbase are not built; promote only if cited with new content.

How this page stays current

Weekly. New Moonshots episode → add the dated tape block, with each number checked against SpaceX's own flight page or a regulator before it is called a fact.

This week (checked 2026-11-16)

What changed this week: Starship Flight 15 flew, and SpaceX caught the ship but not the booster. Ship 42 reached orbit, released 26 V3 Starlink satellites and was caught by the Starbase tower on Sat 14 Nov, about 22.5 hours after liftoff. It was the first time a Starship upper stage was caught. Booster 22 called off its own catch and was destroyed just above the Gulf. The SEC week was quiet: no 8-K, only two shareholder filings. The Apollo lease has not closed. One more classified Falcon 9 flew from Vandenberg.

  • Flight 15, the launch (trade press, secondary): liftoff Fri 13 Nov, 5:47 p.m. ET, Starbase, Texas (CBS News, William Harwood, updated 13 Nov). This is a day later than the 12 Nov planning date and the first attempt to catch both stages on one flight. The EP311 panel, recording live that afternoon, calls it "a few hours from now" (0:00:01–3:13).
  • Booster 22, lost (secondary; SpaceX's words via its webcast): SpaceX said the tower and flight controllers were "go". On Friday, SpaceX's Dan Huot said "the booster aborted itself for a GNC issue that it detected, and that put it into its contingency landing offshore." On Saturday he corrected himself: "It was not a GNC issue that caused us not to come back. GNC triggered that abort; it was doing what it was supposed to do." Kate Tice (SpaceX) said: "Early indications are that the booster ended its boostback [burn] early due to a low fuel mass alarm… The team is still investigating the root cause" (Space.com, Tariq Malik, 14 Nov). One outer-ring engine was out for the boostback, and Musk posted "Slight underperformance of engines on this flight is the reason" (SpaceQ, Marc Boucher, 16 Nov; the post itself not opened). The booster's flight termination system fired just above the water, about 6.5 minutes after liftoff (Space.com; CBS: "exploding in a spectacular fireball moments before hitting the water"). Safe line: "the booster aborted its catch after a short boostback; root cause under investigation."
  • Ship 42, caught (secondary): the ship reached orbit, released 26 V3 Starlink satellites and "circled Earth about 14 times". Three of the Starlinks carried cameras to photograph its heat-shield tiles. The tower's chopsticks caught it at about 3:20 p.m. CT on Sat 14 Nov, "for the first time" for a Starship upper stage (SpaceQ). The 22.5 hours checks against the CBS liftoff time. One search summary says SpaceX confirmed contact with all 26 satellites; no page read says so. Not carried. Say "first tower catch of a Starship upper stage", not "first full recovery": the booster was lost.
  • What is still missing: SpaceX's own Flight 15 page renders empty. The FAA Starship page returned 403 again, so the licence that this flight must have had is still not on our record, and there is no FAA statement on whether the booster loss counts as a mishap. There is no Flight 16 date.
  • SEC (primary, EDGAR company list read 16 Nov): no 8-K filed 7–16 Nov; the newest 8-K is still the 6 Nov director resignation. Two Schedule 13G/A amendments, both from shareholders and both as of 30 Sep: Founders Fund II, LP and 30 related reporting persons (filed 13 Nov, acc. 0001426012-26-000007): "Collectively… 3.8% of the Issuer's outstanding Class A common stock", 314,357,058 shares. Antonio J. Gracias (filed 16 Nov, acc. 0001213900-26-120715): 457,614,684 shares, 5.5% of Class A. Neither filing mentions the Apollo lease or a new director. The Apollo lease is still "a $40B lease commitment, not yet closed, rate not set", and the Harrison seat still has no successor filing.
  • Falcon 9 (launch schedule site, secondary): USSF-627 flew from Vandenberg on 12 Nov, 04:36 UTC, listed as a success (spacenexus.us). That fits last week's scheduled 9 Nov 10:03 p.m. EST only if it slipped about two days. No trade-press story was read. Next listed are USSF-317 and USSF-266 (18 Nov) and NROL-136 (21 Nov), all from Vandenberg. Starlink 15-21 has conflicting dates and is not confirmed flown. Counts still not cited.
  • Boost Mobile: nothing new found; still "in talks, reported".
  • Not carried: "Starship V4, about 70 m and nine engines", reported 16 Nov by a finance site (GuruFocus) from a Musk statement that was not opened. It waits for SpaceX's own words.
  • Moonshots: EP311 (13 Nov, live) has a SpaceX-for-Nvidia equity-swap riff and the line that SpaceX runs its own space situational awareness. Dated block below. EP310's block (12 Nov) stands.
  • Open programmes: nothing new from SpaceX (searched 16 Nov). The FAA Starbase PEIS scoping meetings start tomorrow: 17 Nov in Brownsville, 18 Nov virtual, 19 Nov in Port Isabel. Comments are due 30 Nov 2026. The SpaceX Super Intelligence Partner Network is still rolling, with no deadline published. Both are already in the pipeline, so no new row.

Week of 2026-11-10 (kept)

What changed that week: a SpaceX executive gave Flight 15 a date for the first time — NET Fri 13 Nov — but there is still no FAA licence on the record. One SEC filing: a Class B director, Donald Harrison, resigned (no disagreement, no successor named). The Apollo lease has not closed, and the Boost Mobile talks are still talks. One Falcon 9 flew in the window, a classified Space Force mission.

  • SEC (primary, EDGAR company list and filing read 10 Nov): one filing in 5–10 Nov, an 8-K, Item 5.02, date of report 4 Nov, filed 6 Nov (acc. 0001628280-26-071689, signed by CFO Bret Johnsen). Verbatim: "On November 4, 2026, Donald Harrison, a Class B Director… notified the Company of his intention to resign as a member of the Board of Directors… effective immediately. Mr. Harrison's decision to resign was not due to any disagreement with the Company on any matter relating to its operations, policies, or practices." No successor, no committee detail, no exhibits. Nothing else filed 27 Oct–10 Nov besides the 26 Oct 8-K. No close of the Apollo GPU lease filed or announced: still "a $40B lease commitment, not yet closed, rate not set."
  • Flight 15: NET 13 Nov, from SpaceX's VP of Launch; still not licensed (secondary). Kiko Dontchev gave the date on stage at UP.Summit on Thu 5 Nov, per Ars Technica's Eric Berger as reported by Teslarati, 5 Nov — "the first launch target to come directly from a SpaceX executive"; Teslarati: "The FAA's Starship page does not yet list a Flight 15 license" and "SpaceX has not posted an official mission page yet." Astronomy.com (9 Nov) lists the window opening 5:45 p.m. EST from Starbase Pad 2. The FAA Starship page and NASASpaceflight's "awaits FAA" piece both returned 403; spacex.com/launches renders no content. Teslarati (citing NSF) also settles last week's open static-fire question at the secondary level: Ship 42 single-engine burn 26 Oct, all six engines for 60 s on 27 Oct; Booster 22 33-engine fire 25 Oct. Payload still unconfirmed. Safe line: "NET 13 Nov per a SpaceX executive, first catch attempt for both stages, not yet licensed." Last week's 12 Nov was an FAA planning-slide date, now superseded.
  • Falcon 9 (trade press, secondary): USSF-432 ("R-4"), 6 Nov, 4:48 p.m. PST, SLC-4E Vandenberg, a classified Space Force payload (catalog shows 23 objects; "Starshield-based" is outside analysts' reading, not confirmed), booster B1095 on its 8th flight, drone-ship landing on Of Course I Still Love You; delayed from 1 Nov; deployment not confirmed by SpaceX or the Space Force (Spaceflight Now, 6 Nov). Scheduled, not confirmed flown at time of check: USSF-627, 9 Nov 10:03 p.m. EST, and USSF-317, 12 Nov 11:23 p.m. EST, both Vandenberg (Astronomy.com). Whether Starlink 15-18 (30 Oct) flew is still not confirmed in what was read. Counts still not cited.
  • Boost Mobile (reported, unchanged in substance): no deal announced or priced; no company has commented. Trade analysis notes Boost's MVNO agreements with AT&T and T-Mobile would need carrier consent (RCR Wireless, 4 Nov, not opened). Still "in talks, reported".
  • Not carried: a search summary attributes to Dontchev that Falcon 9 is "on the road to retirement" and SpaceX is "fundamentally an infrastructure company" — no page read carries the quote. A report that Starlink Mobile service moved to early 2028 (StockTitan headline) — not opened. Both wait for a primary source.
  • Moonshots: two new episodes. EP308 (4 Nov) has SpaceX in the Artemis segment and a Starlink thin-client riff; EP309 (7 Nov) has the 3.6 GW / 13 GW compute line and Starlink as the "baby Dyson swarm". Dated blocks below.
  • Open programmes: nothing new from SpaceX (search 10 Nov: no SpaceX/Starlink/xAI call, prize or grant dated this month). The two standing rows are unchanged and already in the pipeline: FAA Starbase PEIS scoping — public meetings 17 Nov (Brownsville), 18 Nov (virtual), 19 Nov (Port Isabel), comments due 30 Nov 2026 — and the SpaceX Super Intelligence Partner Network (rolling, no deadline published). No new row.

Week of 2026-11-04 (kept)

What changed that week: SpaceX named Qualcomm as the supplier for the ground half of Starlink Mobile — the first named cell-site partner — and the SEC week was silent. Flight 15 now has a reported NET of 12 Nov, still with no licence found. The Apollo lease has not closed. Falcon 9 passed its 700th launch.

  • SEC (primary, EDGAR filing list read 4 Nov): no filing 29 Oct–4 Nov; the newest remains the 26 Oct 8-K (the Apollo commitment, below). The 26 Oct commitment also has releases on both sides — SpaceX's IR release ("SpaceX Secures $40 Billion in the Largest GPU Financing Ever") and Apollo's press page, which repeats "expected to close before year-end"; trade press names the equipment as Nvidia Vera Rubin GPUs (Morningstar, secondary). No close announced by 4 Nov: still "a $40B lease commitment, not yet closed, rate not set."
  • Qualcomm, the ground network (4 Nov, from Qualcomm's statement; SpaceX itself said nothing): SpaceX struck a multi-year deal to use Qualcomm's Dragonwing cellular-infrastructure hardware and software in Starlink Mobile base stations, rollout "to begin next year" (2027); Durga Malladi (EVP, Qualcomm): "Together with SpaceX, we are building world-leading cellular infrastructure in the United States." Financial terms not disclosed, and no site count is on the record from either company — "tens of thousands of new cell sites" is CNBC's own line with no named source (CNBC, 4 Nov, Lora Kolodny; also Axios). This answers the 23 Oct "80,000 cell sites" item the same way: no number from SpaceX. CNBC also cites a Washington Post report (2 Nov, not opened) that SpaceX is in talks to acquire EchoStar's Boost Mobile — carried as reported only.
  • Flight 15: NET 12 Nov, reported; still unlicensed. Wikipedia's Flight 15 page (edited 3 Nov) lists NET 12 Nov — its citation is a CANSO/CADENA air-traffic operations-planning listing, not opened — and Starlust says 12 Nov "pending regulatory approval"; one tracker still shows NET 30 Nov, so the dates disagree. No FAA licence modification found this week either. Ship 42's static fire is still not on the record (Wikipedia lists none; last week's Techgenyz claim stands unconfirmed), and the payload is unconfirmed (Wikipedia's "Likely Starlink Group 31-2" is uncited; a video claim of production V3 Starlinks is not carried). Safe line: "NET 12 Nov reported, first catch attempt for both stages, not yet licensed."
  • Falcon 9 (trade press, secondary): Starlink 15-18, 30 Oct, 9:04 p.m. PDT, SLC-4E Vandenberg, 27 V2 Mini Optimized satellites — Spaceflight Now's preview headline calls it the "700th Falcon 9 launch to date" (category page, read 4 Nov; the launch itself not confirmed in what was read). Nothing flew 31 Oct–4 Nov per the same page; next listed is Starlink 15-21, NET 5 Nov (rocketlaunch.live). Year-to-date counts still disagree across outlets; do not cite one — "700th Falcon 9 ever" is Spaceflight Now's count, not SpaceX's.
  • "Largest customer", partly resolved (via Anthropic, 4 Nov; 10-Q not re-read here): the 28 Oct Bloomberg item now has a primary anchor — SpaceX's own Q3 10-Q shows a "Customer B" at 34.5% of revenue (~$5.0B); Bloomberg says that customer is Anthropic, and Anthropic's 2 Nov S-1/A reportedly puts the buy at ~$3.1B/month at full delivery (Dec 2026), cancel windows Jan/Mar 2027. The percentage is primary; the name is still reported.
  • Moonshots: no episode after EP307 (watcher 4 Nov, 109 on disk); EP307's block below stands. SPCX traded down ~2–3.6% in the week's tech sell-off; share price is not tracked here.
  • Open programmes: nothing new from SpaceX — the Qualcomm deal is a supplier contract, not a call. The two standing rows are unchanged and already in the pipeline: FAA Starbase PEIS scoping (comments due 30 Nov 2026; the scoping meetings 17–19 Nov are now under two weeks out) and the SpaceX Super Intelligence Partner Network (rolling, no deadline published). No new row.

Week of 2026-10-29 (kept)

What changed that week: one SEC filing, and it is a $40B GPU lease commitment with an Apollo-managed entity, not closed debt. Flight 15 hardware moved (booster static fire), but there is still no licence and no date. Starshield service in Ukraine is reported restored.

  • SEC (primary, read at EDGAR 29 Oct): one filing in 23–29 Oct, an 8-K, Item 8.01, date of report 24 Oct, filed 26 Oct (acc. 0001104659-26-118614, signed by CFO Bret Johnsen). On 24 Oct SpaceX "entered into a binding commitment letter" for a master finance lease "with an entity managed by Apollo Global Management, Inc.": the lessor buys "graphics processing unit pods and related equipment" and leases them in tranches of "approximately five to five and one-quarter years"; rent quarterly in advance, about one-twentieth of each tranche's cost plus a fixed rate set "at pricing"; aggregate principal "approximately $40 billion"; "expected to close prior to year end"; schedules "expected to commence during the first half of 2027 upon delivery of the GPUs"; it "may not close… or at all". No exhibits. Nothing filed 27–29 Oct. Say "a $40B lease commitment, not yet closed, rate not set". Musk's 25 Oct "closed $40B of debt" post is on Elon Musk.
  • Flight 15: still November, still unlicensed (secondary). Booster 22 fired all 33 engines on 25 Oct; SpaceX announced on 28 Oct a full-duration six-engine static fire of Ship 42 (Techgenyz, citing a SpaceX X post not opened; the firing's own date not given). One tracker still shows Ship 42's static fire as pending after an 18 Oct return for cryo-proofing, so the two disagree. No FAA licence modification found (FAA Starship page returned 403; trade press says none published). Unofficial trackers estimate 8–21 Nov. Safe line unchanged: "November, with a first catch attempt for both stages, not yet licensed."
  • Falcon 9 (trade press, secondary): Starlink 15-16, 27 Oct, 12:25 a.m. PDT, SLC-4E Vandenberg, 27 satellites, booster B1082 on its 26th flight, drone-ship landing (Spaceflight Now; its UTC time is misprinted and deployment is not confirmed on the page). Year-to-date counts still disagree across outlets; do not cite one.
  • A customer contract (company press release, posted 26 Oct on SpaceNews): Astranis upgraded its fourth launch to a dedicated Falcon 9, an amendment to a late-2024 contract, carrying six MicroGEO satellites direct to GEO on an Impulse Helios upper stage, launch 2028. No price given (SpaceNews).
  • Ukraine (reported): Starshield "appears to be working normally again", per Ukraine's defence minister via the WSJ (TechCrunch, 24 Oct); the military says it got no explanation, and SpaceX has not commented. The ~10-day restriction and Musk's own account of it (24 Oct post) are on Elon Musk. This closes the 23 Oct item below as "restored, reported".
  • Reported, not confirmed: Anthropic is SpaceX's largest single customer, more than a third of the AI segment's revenue (Bloomberg via a post, 28 Oct; article not opened; see Anthropic for the S-1 figures). Bernstein raised its price target to $267 (post, not opened). Neither is carried as a fact.
  • Moonshots: EP306 (28 Oct) has SpaceX in finance only; dated block below.
  • Open programmes: nothing new from SpaceX. Two calls stay open and are already in the pipeline: the FAA Starbase PEIS scoping (comments due 30 Nov 2026; meetings 17–19 Nov), and the SpaceX Super Intelligence Partner Network (announced 22 Oct, rolling, no deadline published; reported, the x.ai pages return 403). No new row.

Week of 2026-10-23 (kept)

What changed that week: the first full quarter as a public company is filed, the company itself now says November and a ship catch for Flight 15, and the Tesla merger appears in a filing for the first time, as an "evaluation".

  • Q3 filed at the SEC (primary, read): 8-K, items 2.02/9.01, filed 20 Oct; 10-Q for the quarter to 30 Sep, filed 22 Oct (accession 0001628280-26-066880); an S-8 (employee plan shares), filed 22 Oct. From the 8-K (EX-99.1): revenue $14,575M (+161% YoY), net income $4,404M (a $1,054M loss a year earlier); segments Space $1,767M, Connectivity $5,248M, Super Intelligence $7,560M; capex $19,593M, of which SI $16,444M; no guidance. From the 10-Q: backlog $62,959M at 30 Sep, "approximately 67%" expected within a year; Consumer revenue $3,121M for the quarter; spectrum payments of $856M owed under the Spectrum Credit Agreement (transfer closed 22 May 2026). The 10-Q does not give a Starlink subscriber count, an FAA licence or an HLS status in the part read.
  • The merger, in writing: the 10-Q's only mention is a forward-looking risk: "the evaluation of a potential combination with Tesla, Inc., including whether and when any discussions will continue." No special committee is named in the filing; the committees come from Musk on the 20 Oct call (reported). Tesla's own 8-K and 10-Q do not mention it, and Tesla's 3 Dec proxy has no merger item (see Elon Musk). Not a deal; an evaluation.
  • Flight 15: November, now from the company's own call (reported, no transcript read): on the 20 Oct webcast SpaceX played an ~80-second ship-catch animation under "Up Next: Ship Catch"; Gwynne Shotwell said SpaceX expects the Flight 15 licence "to include return to launch site and tower catch for both stages"; Musk put the catch at "more likely than not" (Teslarati, 20 Oct, 21 Oct, secondary). Range documents still say NET 8 Nov (NSF, secondary). No FAA licence modification found (FAA licence page returned 403). Safe line still: "November, with a first ship-catch attempt, not yet licensed."
  • Falcon 9 (trade press, secondary): Bandwagon-5 rideshare, 20 Oct 07:04 EDT, SLC-40 Cape Canaveral, a day late for recovery-zone weather; 31 customer payloads plus the second flight of SpaceX's Starfall return capsule (first was 23 Jun); booster B1085 on its 19th flight, drone-ship landing (Spaceflight Now, Space.com). Payload deployment not confirmed in what was read. A Vandenberg Starlink mission (27 satellites) was stood down repeatedly and retargeted 23 Oct, 3:28 pm PDT (Spaceflight Now); not confirmed launched at time of check.
  • Artemis (NASA, 23 Oct; see NASA): NASA's acceleration briefing kept Artemis III as a crewed low-Earth-orbit mission in 2027 and surface return in 2028. The ASAP (22 Oct) put SpaceX's cryogenic-transfer demo in early 2027 and an uncrewed landing in late 2027 (as logged on the NASA page; not re-read here).
  • Reported, not confirmed: Ukraine says Starlink/Starshield speed limits since 14 Oct persist (CNBC, 23 Oct; page 403, SpaceX's answer not read). SpaceX seeking access to "80,000" US cell sites: no source found for the number; the nearest on record is Bloomberg, 5 Aug 2026, "thousands of cell towers and small cell nodes". Not carried as a fact.
  • Moonshots: no episode after EP304 (watcher 23 Oct, 106 held). EP304's block below stands.
  • Open programmes: none from SpaceX. The one public call stays the FAA Starbase PEIS scoping (comments due 30 Nov 2026; meetings 17–19 Nov), already in the pipeline as gov-faa-starbase-peis-scoping and in the operator guide; no new row.

Week of 2026-10-17 (kept)

  • Flight 15 slipped to November, with a ship catch planned (reported, not SpaceX's page): Elon Musk at the X Takeover event, Giga Texas, 10 Oct: Flight 15 is "four to six weeks away", "probably launches a little bit before Thanksgiving, is my guess", and "We're going to try to catch the ship… the first time that there's been full recovery of an orbital class rocket" (Teslarati, 12 Oct, secondary; the event video not opened). Range and airspace planning documents list NET 8 Nov (NASASpaceflight, 14 Oct, secondary). Last week's 19 Oct is dead. No SpaceX mission page and no FAA licence modification found. Do not say a date; "November, reported" is the safe line.
  • FAA opened the Starbase expansion review (primary, read at the Federal Register): Notice of Intent published 14 Oct 2026, FR Doc 2026-20963, docket FAA-2026-12808 (Federal Register): a Programmatic EIS for "up to 456 annual launches" and "up to 912 annual landings at the Boca Chica Launch Site" (Phase 1 up to 100 launches / 200 landings; today's licence is "up to 25 annual launches and 50 total annual landings"). Public scoping meetings 17 Nov (Brownsville, Texas Southmost College), 18 Nov (virtual), 19 Nov (Port Isabel); scoping comments due 30 Nov 2026 at regulations.gov. A Draft PEIS "at a later date". Flight 15 flies under the existing 25.
  • Falcon 9, three launches in the window (trade press, secondary; SpaceX's launch page returned no content): Starlink 15-25, 12 Oct 17:28 PDT from Vandenberg, 27 V2 Mini satellites, the first Falcon 9 Starlink launch since 19 Sep (Space.com, NSF); CRS-35, 13 Oct 06:33 EDT, NASA cargo to the ISS (Spaceflight Now headline); USSF-481 / "TH-2", 15 Oct 17:01 PDT from Vandenberg, a classified Space Force payload (Spaceflight Now). Year-to-date counts disagree across outlets (114th Falcon 9 vs 120th/121st SpaceX mission): do not cite a count.
  • Q3 2026 results Tue 20 Oct, after the close, audio webcast 3:30 pm CT on ir.spacex.com and X (reported from a 16 Oct investor release by several outlets; the release itself not opened, the IR site returned no content). First full quarter reported as a public company (Nasdaq: SPCX). SEC EDGAR, primary, read 17 Oct: CIK 0001181412, no filing between 1 and 17 Oct; the latest is a Form 4/A dated 30 Sep.
  • TIME Best Inventions 2026, 15 Oct (primary, read): the Starlink V5 terminal is listed, "about 35% smaller and 52% lighter than the V4 terminal", "more than 12 million people" subscribe, "11,000 Starlinks in orbit" (TIME). TIME's figures, not SpaceX's.
  • Not this week, despite the search dates: the Space Force awards that surfaced ($4.16B SB-AMTI, $2.29B Space Data Network Backbone) are from 29 May and 26 May 2026; the listing pages were re-dated by the crawler. Not new.
  • Moonshots: the 15 Oct episode's SpaceX content is the "quadrillion" post only (transcript swept 17 Oct: no Starship, Flight 15, Starbase or Starlink mention); its block below is unchanged. No newer episode on disk or in the feed to 17 Oct.
  • Open programmes: none from SpaceX. One public call: the FAA scoping comment period above, logged in the submissions pipeline as gov-faa-starbase-peis-scoping (fit none, pass). The supplier-portal row is unchanged (page still unreadable).

Week of 2026-10-11 (kept)

  • Flight 14 landing site, partly confirmed: Spaceflight Now headline, 28 Sep: "Starship returns to Earth; rocket splashes down north of Hawaii after three-hour flight" (headline only, tag page). "Near Hawaii" now has a trade-press source; "two orbits" still not confirmed, and SpaceX's own flight page not opened (spacex.com returned no content).
  • Falcon 9, 9 Oct: Spaceflight Now headline: SpaceX launched 21 data satellites for the Space Development Agency (headline only; payload not opened).
  • Flight 15: unconfirmed. A user-generated prediction-market page cites a NET of 19 Oct with a first ship catch attempt; trackers list it TBD; no FAA licence or mishap statement found. Do not repeat the date.
  • Open programmes: none found. No SpaceX call, prize or grant open to outsiders. The Texas Space Commission $7.5M award is a state grant to SpaceX (undated in results); the UCLA/Dragon research call closed 15 Mar 2025. Supplier registration is spacex.com/supplier (third-party description only).

Updated: 2026-11-16.

Related

Moonshots #298 · Alexander Wissner-Gross · Emad Mostaque · NASA

2026-10-15 episode (Moonshots #unnumbered): the "quadrillion" post

Checked 16 Oct, secondary sources only (the post itself not opened). The panel treats it as its last story, at 1:38:20, 82% of the way in.

  • The post: Elon Musk on X: "This will sound super crazy, but I see a path to SpaceX being worth orders of magnitudes more than the current human economy" (tape; reports quote "Earth economy"). In a follow-up he said a path is not extreme confidence and the probability is not zero (reported). He said a similar thing in July (reported).
  • The arithmetic on tape (1:38:40): "$120 trillion per year… 1 to 10 quadrillion dollars." Reported anchors: IMF 2026 world output about $126 trillion; SpaceX's market value about $2.2 trillion at the close before the post (reported). Two orders of magnitude above $126T is about $12.6 quadrillion. His reasoning is energy (the Kardashev scale; orbital solar and data centres), per coverage.
  • Panel: Dave Blundin: "I see a path. It's not a forward-looking statement because it doesn't make a specific prediction" (1:39:09). Alexander Wissner-Gross: do the same work "100 or 1,000 or 10,000 times more quickly". Peter Diamandis: "the first quadrillionaires will be made there" (1:42:00). No timeline is claimed by any of them.

EP299 (2026-10-03, Moonshots #299)

  • Cold-open clause: SpaceX "launched the next batch of astronauts [to] the ISS, and also launched Google's TPUs into orbit for Project Suncatcher" (0:02:01). Suncatcher ✅ flew on Transporter-18, 1 Oct, four TPUs, built with Planet (Google); never discussed in the episode.
  • Musk co-leads Project Meridian; Peter notes SpaceX and Anduril hold multi-billion defence contracts (2:02:40).

EP304 (Moonshots #304, aired 2026-10-21): a share for every child, and the merger committees

  • The pledge: "Next year, I will give one of my SpaceX shares to every kid in America. This will be the first of many." (Musk on X, 19 Oct, read at 0:22:39). At ~$146–148 a share × ~72M children ≈ $10.5B (CNBC, NYT, 19 Oct). Trump says the shares go through Trump accounts. Not said on tape: Gwynne Shotwell pledged a share per child to ~2M Trump-account children in July (CNN, 19 Oct).
  • The host's own arithmetic says it is a gesture: "Can Trump accounts… actually move the needle for the average American? And I think the answer is probably no" (Peter 0:24:28).
  • The merger: "Musk says Tesla and SpaceX have formed special committees to discuss the merger" (0:31:03). Record: SpaceX Q3 call, 20 Oct; Musk: assuming it is certain "would be overstating it"; Tesla needs a majority of all shares. Peter's Moonshots #303 prediction (announcement by 31 Dec 2026) is open, not met.
  • Peter's SpaceX holding ("my largest financial bet", Moonshots #284) is not restated in this segment (Moonshots: the disclosures ledger).

EP305 (Moonshots #305, aired 2026-10-23): the Q3 call, retold; checked 2026-10-24

  • The headline number is the AI segment, not the company. Peter: "830%. That's how fast SpaceX's business grew in just one year. AI, not rockets, is now SpaceX's largest business" (0:05:59). ✅ Super Intelligence revenue $7.56B against ~$813M a year earlier (+830%, secondary: several outlets quoting the shareholder materials; the $7,560M is the 8-K figure above). Company revenue grew 161%. "Rockets is third in terms of revenue" (0:11:23) ✅: Space $1,767M, Connectivity $5,248M, SI $7,560M. Say "the AI segment grew 830%".
  • Capex and payback: "almost 20 billion per quarter mostly on AI data centers" (0:15:08) ✅ $19,593M, of which SI $16,444M. "Each new data center pays for itself in under a year" (0:15:19) ⚠️ the CFO's line was "a payback of less than a year" at current pricing (call transcript, earningswhispers).
  • Compute, in Musk's words (call transcript): "fairly confident of being able to bring an incremental 10 gigawatts of compute online in 2027"; "we're actually going to aim for an incremental 20 gigawatts of power cooling infrastructure next year"; "a non-zero possibility… of over 20 gigawatts of super intelligence compute online next year"; he would be surprised by under 13 GW total at end-2027; 440,000 Vera Rubin for the Q2 training cluster. All terrestrial (Peter, 0:16:09: "It's all terrestrial right now", correcting Dave's "in space"). See Orbital compute (data centres in space).
  • Starlink: 13.6M subscribers, +1.6M in the quarter ✅; 800 MHz of spectrum bought and an FCC-approved 15,000-satellite mobile constellation ✅ (XTB). Musk, "about a 100 times the bandwidth of the current Starlink constellation" within about three years ✅ said, a target. The show's own EP298 number for the same horizon was 10× (one analyst's "conservative projection", 1:37:30, above); the two are not reconciled on tape.
  • Flight 15: "we expect to happen next month… attempt to catch both the booster and the ship" ✅ said; Musk put it "more likely than not"; still no FAA licence modification found (see 2026-10-23 block above).
  • $100B: Peter, "their guidance, a 100 billion of annual revenue run rate by December" (0:15:39). ⚠️ Reported as SpaceX saying it "remains on track to reach $100 billion in annualized revenue by the end of 2026" (Broadband Breakfast); the 8-K has no guidance, and the call transcript was not read for the sentence. Say "SpaceX says it expects". Musk also floated "$500 billion and a trillion" of recurring revenue by end-2027 (Tesla Oracle, secondary), which Peter does not repeat.
  • Merger, on tape: special committees, Tesla needs 50% of all shares, and Alex's "pre-registered prediction" that the merged company is called XCorp (0:40:57). Peter's EP303 date (announcement by 31 Dec 2026) remains open. The "29% to 56%" market jump (0:37:11) is not reproducible: Kalshi's "before 1 May 2027" market was ~56% in one snapshot; Polymarket's year-end market was 19% in another.
  • Disclosure: Peter restated his holding at 0:11:02 ("I'm an early SpaceX investor. I my have a huge position personally in SpaceX"), the first time on tape since Moonshots #285.
  • Open for the space lane: the FAA Starbase PEIS comment period closes 30 Nov 2026 (above); nobody on the panel mentions it.

EP306 (Moonshots #306, aired 2026-10-28): SpaceX as a borrower and a compute vendor; checked 2026-10-29

  • No space segment. SpaceX appears only in finance: Elon's 25 Oct post says SpaceX "closed $40 billion of debt to buy GPUs" because "We're investment grade and have Starlink and launch revenue" (read by Peter, 0:27:50–0:28:12). ⚠️ "Closed" is Musk's word. Reuters said it was unclear the deal had closed, and Bloomberg said pricing was still being negotiated. It is reported as private bonds from an off-balance-sheet vehicle, and SpaceX CDS hit a record 266 bp. Peter: "launch business has zero real competitors" (0:28:23), an opinion.
  • Anthropic's S-1 counterparty: "Enthropic… just agreed to buy $120.5 billion of compute from SpaceX through 2029" (0:28:43). ⚠️ A maximum, via an analyst post. Earlier reports said $84.5B, and the non-cancellable portions end Jan/Mar 2027. See Anthropic.
  • Dave: Elon borrows "at 8%" vs neoclouds at 15% and resells compute to Anthropic, "a license to print infinite money" (0:30:41–0:30:53). ❓ The 8% was not found. Also Dave: "you can do that via SpaceX too, but you're buying into a lot of other things" (0:36:47), SpaceX as the other way into AI.
  • Disclosure: Peter's holding (EP284, restated EP305) is not restated here while he prices SpaceX's moat (Moonshots: the disclosures ledger).
  • Open for the space lane: Starship Flight 15 (the "next month" catch from the Q3 call) gets no mention. The FAA PEIS window (30 Nov) is still unmentioned.

EP307 (Moonshots #307, aired 2026-10-31): "a data center hyperscaler company", and the Moon as a regulatory exit

  • No space segment, for the second episode running. SpaceX comes up inside other stories.
  • Alex, the NASA line: Elon "maybe doesn't care as much say about NASA contracts as he used to. Why? Because SpaceX is now a data center hyperscaler company and Anthropic is its largest customer" (1:29:22–1:29:45). ⚠️ "Largest customer" is an inference from the Anthropic deal the show read on Moonshots #306. That deal is a maximum (up to $120.5B), firm only into early 2027. Whether NASA work has slipped is checkable against Artemis milestones; the tape gives none. Earlier, Alex said SpaceX "is no longer a rocketry company and is now a Neo hyperscaler" (0:07:48, [attr?]) and that Elon "retcons the moon into SpaceX's plans" (0:41:11).
  • The Moon: on Altman's self-replicating data centres Alex says they want to be "orbital or even better lunar because they're free of regulatory constraints", and that Altman avoids saying so because it hands "the ball to Elon and SpaceX and their virtual monopoly in the short term on heavy launch" (1:04:32–1:04:54). Peter predicts OpenAI buys Rocket Lab or Relativity Space and that "regulation's going to go to the moon" (1:05:26). Relativity was Peter's own early investment (Moonshots #266), not restated. See Orbital compute (data centres in space).
  • Cybercab as Starlink infrastructure (Peter, 1:30:51): a "repeater for Starlink" with "20 kilowatt hours of power"; Alex says production priority (Optimus over Cybercab) argues against it.
  • The merger, taken for granted: Alex uses "Tesla SpaceX merge" (1:29:45) as a premise and makes Dave's scenario his own: they merge so that Optimus builds SpaceX's data centres (1:21:56).
  • Disclosure: Peter's holding (EP284, restated EP305) is not restated (Moonshots: the disclosures ledger).

EP308 (Moonshots #308, aired 2026-11-04): Artemis III with both landers, and the Starlink thin client; checked 2026-11-10

Caption quotes below are from the auto-captions on disk, which read as a re-translation; wording is approximate, timestamps are exact.

  • Artemis III docking with both landers (Peter, 1:26:43–1:27:34): Isaacman's plan, "next summer, 2027", four astronauts ("three Americans and… one Italian"), "a two-week mission in Earth orbit" to "practice docking the Orion capsule with both lunar landers", Blue Origin's and "Starship SpaceX", "three launches from Cape Canaveral"; Artemis IV lands in 2028; NASA "will wait as long as possible before choos[ing] between SpaceX and Blue Origin" to land the crew. Against the record: LEO mission 2027 and surface return 2028 match what NASA already holds (NASA's 23 Oct briefing). The crew nationality, the "two weeks", the three-launch count and the deferred lander choice are not re-checked here.
  • Starlink as the thin-client network (Dave and Alex, 0:10:59–0:13:17): phones as thin clients served "due to Starlink directly" from a data centre; "You hear us, Elon. We want a phone Starlink and phone service, and… a new Starlink laptop"; Starlink "can be a direct competitor in the consumer market" for RAM. Opinion, no number. See Orbital compute (data centres in space).
  • Starship over-supply (Alex, answering a viewer, 1:46:07–1:47:00): if Starships are built "in the expected and publicly predicted quantity" without new uses, "many of these starships… will eventually come out empty"; he calls that unlikely because orbital data centres will absorb the lift. Opinion.
  • Disclosure: Peter's holding is not restated in the Artemis segment (Moonshots: the disclosures ledger).

EP309 (Moonshots #309, aired 2026-11-07): compute on the ground, and the "baby Dyson swarm"; checked 2026-11-10

  • "spacex is building like 3.6 gigawatts here on earth… targeting 13 gigawatts by the end of next year" (Salim, 0:19:41–0:19:49), said unattributed. Per the episode's own check: 3.6 GW under construction is SemiAnalysis's estimate (3 Nov), 13 GW by end-2027 is Musk's (20 Oct call), and SpaceX's 10-Q shows 1.7 GW installed at 30 Sep. Carried with that attribution; not re-read here.
  • "The baby Dyson swarm is already here" (Alex, 1:35:52): "between starlink and soon all of elon's ai satellites and star cloud and lonestar… We have a financial interest"; the disclosure covers Starcloud and Lonestar, not SpaceX. Opinion. See Orbital compute (data centres in space).

EP310 (Moonshots #310, aired 2026-11-11, recorded 11-10): Baron, disclosed; "no longer catering to NASA" vs CLPS 2; checked 2026-11-12

  • Peter disclosed his position before the claim (1:08:25); see Moonshots: the disclosures ledger.
  • Baron (6 Nov), as said: a ~$100B run-rate in December 2026; December 2027 "quite likely… exceed 500 billion", "not a promise" (✅ run-rate on December revenue). Terafab "a trillion dollars" is Baron's figure; Musk said it would get there over time, in phases (⚠️). Robots: "probably rent them" (✅).
  • 🔴 Alex: SpaceX "no longer is catering to the government business, no longer catering to NASA" (1:13:14). It is a CLPS 2 awardee (7 Nov) and the HLS provider, and it will fly cislunar comms satellites on Starship in 2028 (NASA).
  • 🔴 Peter: SpaceX "got started with the commercial crew contracts… December 2008" (1:58:01). It was the CRS cargo award ($1.6B, 23 Dec 2008), which came after Falcon 1's fourth flight succeeded (Sep 2008).
  • Compute location split: Emad and Dave call Terafab plus space data centres a "head fake"; Alex says data centres and the Dyson swarm (Orbital compute (data centres in space)).

EP311 (Moonshots #311, recorded live 2026-11-13): an equity swap with Nvidia, and SpaceX's own space tracking; checked 2026-11-16

  • Flight 15 as a curtain-raiser: in the cold open the panel says the flight with the booster and ship catch is "a few hours from now" (0:00:01–3:13). The flight itself is covered in "This week" above. The episode was recorded before either catch attempt.
  • SpaceX for Nvidia, a riff (Emad Mostaque, 0:25:13): "have a 100red billion of SpaceX equity for 100 billion of Nvidia equity" (the auto-caption's wording), offered during the Jensen Huang resignation segment. It is a suggestion. No talks were reported, and SpaceX has filed nothing (the EDGAR check above).
  • Space situational awareness (Alexander Wissner-Gross, 1:41:19): low-orbit tracking is "abysmally poor" and "SpaceX had to spin up their own". Opinion, not re-checked. Safe as "SpaceX runs its own conjunction screening", not "had to".
  • Disclosure: Peter's SpaceX holding (EP284, restated EP305 and EP310) is not restated while the panel discusses SpaceX buying into Nvidia (Moonshots: the disclosures ledger).
Reading rule. Anything marked unverified or on-air is a claim made on the podcast and has not been checked against a primary source yet. It stays marked until it has. Sourced numbers link to where they came from.