SpaceX's laser-link interface opened to any builder by end of 2028
25% · Scenario modelResolves 2028-12-31
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimBy 11:59 PM ET, Dec 31, 2028, the interface used by SpaceX's optical inter-satellite links (the waveform and protocols of Starlink or Starshield laser terminals) is made available for any qualified company to build compatible terminals without buying SpaceX hardware: published or openly licensed by SpaceX, submitted to a standards body, or adopted by the U.S. Space Force as an interface standard released to industry. The release must be publicly announced; the specification itself may be export-controlled.
FalsifierNo such release is publicly announced by 2028-12-31 23:59 ET. Not counted: SpaceX selling its laser terminals (Plug & Plaser, mini laser); a specification shared only with one contractor or kept classified; SpaceX adopting the SDA OCT standard or flying third-party SDA-standard terminals; a government-owned interface standard that is not SpaceX's laser interface. The announcement date governs.
Scenario modelA. Space Force adopts SpaceX's laser interface as the SDN standard and releases it to other vendors — weight 0.20, 70% clears; B. SpaceX opens its interface itself, NACS-style, or submits it to a standards body — weight 0.10, 95% clears; C. Interop solved without opening it: SDA-standard or other third-party terminals on SpaceX backbone satellites, or a government-owned SDN interface — weight 0.45, 2% clears; D. Status quo: SpaceX sells terminals, keeps the interface closed; translators bridge the gap — weight 0.25, 0% clears. Weighted total 0.244.
FCC grants all 100,000 Gen3 Starlink satellites in one order by end of 2027
20% · Monte CarloResolves 2027-12-31
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimThe FCC authorizes SpaceX to construct, deploy and operate all 100,000 satellites requested in its Gen3 NGSO application (filed July 6, 2026) in a single order adopted on or before 11:59 PM ET, Dec 31, 2027.
FalsifierNo FCC order adopted by 2027-12-31 23:59 ET authorizes all 100,000 Gen3 satellites. An order that grants fewer than 100,000 or defers any part of the satellite count resolves No, even if the rest is granted later. A full-count grant that defers some frequency bands or adds conditions still counts. The adoption date on the order governs.
Monte Carlo · inputs400,000 trials, seed 17, clock from filing 2026-07-06; deadline 17.8 months. Time to first action lognormal, median 18 months, sigma 0.45 (Gen1 16 months; Kuiper ~13; Gen2 partial ~30). 60% processed with the Part 100 speedup (adopted 2026-07-22, 'default to yes'; effective date pending), duration x U(0.6, 0.9). P(full count | first action) Beta mean 0.20 under old practice (Gen2: 7,500 of 29,988, then 7,500 more, 14,988 still deferred, 'untested on orbit'), 0.30 under the new rules. 5% a second order grants the rest by the deadline. AI factor share 10-30%, effect -19% to +31%. Result 0.196; P(any FCC action by the deadline) 0.64.
First SpaceX Starmind satellite reaches orbit in 2027
30% · Monte CarloResolves 2027-12-31
Disclosure: I hold SpaceX and NVIDIA.
Claim, falsifier, and reasoning
ClaimA satellite SpaceX identifies as a Starmind satellite (AI1 or a prototype of it) reaches Earth orbit on or before 11:59 PM ET, Dec 31, 2027.
FalsifierNo satellite that SpaceX identifies as Starmind or AI1 reaches orbit by the deadline. A Starlink satellite carrying a compute test payload does not count unless SpaceX identifies it as Starmind or AI1. A satellite destroyed before reaching orbit does not count. A satellite in orbit that fails afterwards still counts. The launch date in UTC governs.
Monte Carlo · inputs400,000 trials, seed 13, clock from 2026-11-07. SpaceX target: first Starmind satellites in Q4 2027, moved up from 2028 in the IPO prospectus (Musk on X, reported 2026-08-24/25). Center mid-Nov 2027. Earlier reports of two AI1 prototypes in early 2027 (Space.com; 'people familiar'): 15% chance a prototype flies early, months 3-10. AI1 needs Starship (Engadget); Starship's first orbital flight 2026-09-28 (Flight 14, 26 Starlink V3); 85% Starship ready in time, else +3 to +9 months. Slip lognormal median 4 months, sigma 0.8, with 10% on time or early (reference: Starship V3 debut, April target, flew May 22-23 2026). AI factor share 10-30%, effect -19% to +31%. Result by Dec 31 2027 0.288; by Sep 30 2027 0.133; by Jun 30 2028 0.767.
Stoke's Nova Pathfinder reaches orbit by June 30, 2027
10% · Monte CarloResolves 2027-06-30
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimStoke Space's Nova Pathfinder (or any Stoke Nova vehicle) places its upper stage or a payload into Earth orbit on or before 11:59 PM ET, Jun 30, 2027.
FalsifierNo Stoke launch reaches orbit by the deadline. Orbit means the upper stage or a payload completes at least one orbit, as confirmed by Stoke, the FAA, or an object catalogued by U.S. Space Force tracking (Space-Track). A launch that reaches space but not orbit does not count. A failure of reuse, landing or recovery after orbit is reached does not affect resolution. The launch date in UTC governs.
Monte Carlo · inputs400,000 trials, seed 11, clock from 2026-11-07. Target: first orbital launch 'early 2027' (Stoke, 2026-09-08), NET Mar 2027; center mid-March. As of Sep-Nov 2026 the stages had not yet been fired together at the Cape (TechCrunch; Lapsa via Runtimewire). Stoke's own targets: orbit in 2025 (Oct 2023), early 2026 (Jul 2025), late 2026, then early 2027. Slip mixture: 35% a 0-3 month slip, else 4-15 months (reference: Neutron, Vulcan, New Glenn). First-flight success Beta mean 0.35 (Seradata: 50% maiden loss since 1990, 30% since 2000; Blue Origin first new space company to orbit on a first launch, Jan 2025). Second attempt median 7 months later, 60% success. AI factor share 10-30%, effect -19% to +31%. Result 0.124; P(first launch by Jun 30) 0.35.
Space Force fixes its laser-link gap both ways by end of 2028
20% · Monte CarloResolves 2028-12-31
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimBy 11:59 PM ET, Dec 31, 2028, both of the following are public: (a) the U.S. Space Force or SpaceX states that Space Data Network Backbone satellites carry, or will carry, optical terminals that link directly with satellites built to the Space Development Agency's Optical Communications Terminal (OCT) standard; and (b) the Space Force awards a contract, order or agreement to build and fly government-owned Space Exchange Point (translator) satellites, beyond Phase A planning, to a prime contractor other than SpaceX, covering 20 or fewer satellites.
FalsifierEither leg is not public by 2028-12-31 23:59 ET; the date of the public statement or award announcement governs. Leg (a) does not count if the link runs through a translator satellite, or if it is only a study with no commitment to fly. Leg (b) does not count for Phase A or other planning and study awards, for translator satellites with SpaceX as prime, for an award covering more than 20 translator satellites, or for ground-gateway routing. An award made in secret counts only if it is public by the deadline.
Monte Carlo · inputsTwo legs, 400,000 trials, seed 7, clock from 2026-11-07. Funding: CR through Dec 11 with no new starts (GovExec, Sep 2026); assumed 20% enacted by Dec 11, 45% by Mar 31 2027, else a full-year CR to Oct 1 2027; 30% chance the translator work proceeds on prior-year or OTA funds. Leg (b): SEP Phase A on York's DO3, announced 2026-10-12, ending 5.0-8.2 months out; 65% the Space Force pursues government translators, cut by 25% if a SpaceX fix is public first; award lag lognormal, median 3 months; 80% the buy is 20 or fewer, 85% the prime is not SpaceX, 85% disclosed. Leg (a): 65% a SpaceX forward fix exists by end-2028 (SpaceX holds architectural design authority, Sep 2026; prototype due end-2027); confirmation time triangular 4/12/26 months; 80% disclosed. AI factor: durations scaled by 1 - share x effect, effect -19% to +31% (METR 2025 RCT; Echoes of AI), share 10-30%. Result 0.193 (a 0.52, b 0.37); by Sep 30 2027 0.03.
FCC opens part of 482 MHz to satellite-to-phone by end of 2027
75% · Monte CarloResolves 2027-12-31
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimThe FCC adopts a Report and Order in the "Unleashing Licensed Spectrum for Direct-to-Device" proceeding (GN Docket Nos. 26-281 and 23-65; proposal adopted Oct 29, 2026) that authorizes satellite direct-to-device use of at least one band from the proposed 482 MHz, on or before 11:59 PM ET, Dec 31, 2027.
FalsifierNo such Report and Order has an adoption date on or before 2027-12-31 23:59 ET. The adoption date on the order governs, not Federal Register publication or the effective date. Any band from the proposal counts, even if the others move to a further notice. A further notice of proposed rulemaking, a public notice or a declaratory ruling alone does not count.
Monte Carlo · inputsSatellite rulemakings under this FCC took 8–12 months from proposal to final rules: Upper C-band about 8, the licensing overhaul about 9, satellite spectrum sharing about 12; the 2023–24 satellite-to-phone (SCS) framework 12. The model adds comment extensions, a 20% stall risk, a 60% chance the contested 2.5 and 3.45 GHz bands are split off, shutdown risk and meeting timing. Clock from NPRM adoption 2026-10-29. Federal Register publication median 30 days; comments +30 and replies +60 days after publication; 40% chance of a 15–30 day extension; stall risk drawn from Beta(2,8) (mean 20%), a stalled item taking a median 540 days; otherwise replies to order median 165 days; 60% chance the contested 2.5 and 3.45 GHz bands are split off, else 30–120 extra days; 30% chance of a shutdown adding a median 20 days; up to 30 days to the next open meeting or circulation vote. 400,000 trials, seed 42.
How I got 75%
ModelOrder date = NPRM adoption (2026-10-29) + Federal Register publication + comment cycle (+ extension) + replies-to-order time (stall or priority track, plus band-split and shutdown delays) + wait for the next vote. P(YES) = share of 400,000 trials in which the order is adopted by 2027-12-31.
Federal Register publication, days after adoption (median; clipped 10–120) = 30.00The NPRM is published in the Federal Register a median 30 days after adoption.Evidence. A judgement.
Comment cycle, days after publication = 60.00Replies close 60 days after publication (comments at 30); a 40% chance of a 15–30 day extension on a contested item.Evidence. The FCC's circulated draft (fact sheet, 2026-10-08) sets comments at 60 and replies at 90 days after publication; the model uses 30 and 60.
Stall risk (Beta 2, 8; mean) = 0.20One time in five the item languishes, taking a median 540 days from replies to an order.Evidence. A judgement.
Replies to order, days (median, priority track) = 165.00On the priority track, an order follows replies in a median 165 days.Evidence. Kevin's reference class: satellite rulemakings under this FCC took 8–12 months from proposal to final rules.
Contested 2.5 and 3.45 GHz bands split off = 0.6060% of the time the FCC moves the contested bands to a further notice and acts on the rest; otherwise interference analysis adds 30–120 days.Evidence. CTIA's members asked the FCC to drop 2.5 and 3.45 GHz, 290 MHz of the 482 (press, 2026-10-29). The split probability is a judgement.
Shutdown during the window = 0.30A 30% chance of a lapse in appropriations, adding a median 20 days (clipped 3–70).Evidence. A judgement.
Wait for the next open meeting or circulation vote, days (uniform) = 15.00Up to 30 days, uniform, from a finished draft to the vote.Evidence. A judgement.
ResultBy 2027-12-31: 0.738, logged as 0.74. By 2027-10-31: 0.582; by 2028-06-30: 0.874. Median order date 2027-10-08 (p10 2027-06-28, p90 2028-09-27).
CorrelationThe stages are added, and each draw is independent. A single cause that slows every stage at once (a change in FCC leadership or priorities) is only partly carried by the stall branch.
SensitivityOne input at a time, the rest as stated (P by 2027-12-31): stall risk ~10% → 81.4%, ~30% → 66.2%; replies-to-order median 4.5 months → 79.1%, 7.5 months → 57.1%; always split → 79.7%, never split → 65.1%; no shutdown → 75.1%, 60% shutdown → 72.4%.
Signals to watch
Up
- Federal Register publication within about a month, with no extension of the comment dates.
- The 2.5 and 3.45 GHz bands moved to a further notice.
- A draft Report and Order on an open-meeting agenda.
Down
- An extension of the comment or reply dates.
- Carrier opposition widening beyond 2.5 and 3.45 GHz.
- A lapse in appropriations, or a change in FCC leadership or priorities.
Not counted as evidenceA further notice of proposed rulemaking, a public notice, a declaratory ruling, or a waiver or special temporary authority for one operator. None of them is a final rule.
Sources
The model, runnable →
No new robotics or machinery tariff by Oct 31, 2027
80% · Reference classResolves 2027-10-31
Disclosure: I hold Tesla.
Claim, falsifier, and reasoning
ClaimNo new Section 232 tariff, duty, quota or fee on imports of robotics, industrial machinery or their components is proclaimed by 11:59 PM ET, Oct 31, 2027.
FalsifierA proclamation under the 2026-10-28 Section 232 action on robotics, industrial machinery and their parts and components, signed by 2027-10-31 23:59 ET, imposes any tariff, duty, quota or fee on those imports, including a country-specific measure such as one on China only. The signing date governs, not the effective date. Not counted: existing steel, aluminum or copper derivative tariffs that already cover machinery; tariffs under other authorities (Section 301, IEEPA); negotiated agreements that impose no measure; an investment incentive program.
Reference classTalks-first Section 232 actions, which rarely became tariffs within a year: autos (2019), titanium sponge (2020) and processed critical minerals (2026; no tariff three months past its July 13 update). Adjusted up for the 2026 tariff record (drones, semiconductors, pharmaceuticals, polysilicon) and the proclamation's finding on certain foreign entities, which makes a China-only measure the likeliest NO path; adjusted down because tariffs on machine tools raise the cost of the domestic capacity the policy aims to build.
2026-09-004 · logged 2026-10-29 · Space
Starlink's segment tops $19B in 2026
20% · Monte CarloResolves 2027-04-30
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimSpaceX reports Connectivity segment revenue of at least $19,000 million for the fiscal year ended 2026-12-31 in its Form 10-K.
FalsifierThe Form 10-K for the year ended 2026-12-31 filed by Space Exploration Technologies Corp. (CIK 0001181412) on SEC EDGAR reports Connectivity segment revenue below $19,000 million in its segment note.
Monte Carlo · inputsQ1–Q3 2026 Connectivity revenue $3,257M, $4,291M, $5,248M (10-Qs); Q3 consumer $3,121M and enterprise and government $2,127M (Q3 10-Q, Note 3); subscribers 12.0M and 13.6M at 30 June and 30 September, Q3 ARPU $69; judged: Q4 net adds 1.4–2.5M, ARPU $65–72, enterprise and government growth +12% (or +35% in a large-booking quarter, 25% likely), hardware residual two ways. Model: scripts/forecasts/2026-09-004_mc.py.
How I got 20%
ModelQ4 Connectivity revenue = consumer service (average subscribers × ARPU × 3) + consumer hardware residual + enterprise and government. P(YES) = share of 1,000,000 trials in which Q4 ≥ $6,204M, blended 50/50 over two hardware cases.
Q4 net adds, millions (triangular 1.4 / 1.9 / 2.5) = 1.90Starlink adds 1.4 to 2.5 million consumer subscribers in Q4, most likely 1.9 million.Evidence. Subscribers rose from 12.0M at 30 June to 13.6M at 30 September (Q2 and Q3 10-Qs). The range is a judgement.
ARPU, $ a month (triangular 65 / 68.5 / 72) = 68.50Consumer ARPU in Q4 is $65 to $72.Evidence. Q3 ARPU was $69, Q2 $66 (10-Qs). The range is a judgement.
Large enterprise and government booking quarter, probability = 0.25One chance in four that Q4 carries a large carrier or government booking (+35% ± 12% on Q3's $2,127M); otherwise +12% ± 8%.Evidence. Q3 enterprise and government was $2,127M (Q3 10-Q, Note 3). The probability is a judgement.
Hardware residual = 471.00Q3 consumer revenue less subscribers × ARPU leaves $471M of hardware and other; in Q4 it either grows with consumer or falls back after a Q3 retail sell-in, 50/50.Evidence. Inferred, not filed: $3,121M − 12.8M × $69 × 3.
ResultBlended P(FY ≥ $19.0B) = 0.183 (hardware grows: 0.221; falls back: 0.144). FY p10/p50/p90 $18,373M / $18,683M / $19,176M. Logged at 0.18.
CorrelationThe draws are independent. A single cause that lifts adds, ARPU and bookings at once (India going live, say) would widen the upper tail beyond what the model shows.
SensitivityBooking-quarter probability 10% → 9%, 40% → 27%. Threshold $18,500M → 77%, $18,800M → 35%, $19,500M → 2%.
Signals to watch
Up
- A large carrier or government agreement booked in Q4.
- Starlink India going live before December.
- A year-end consumer count above 15.5 million.
Down
- A year-end count near 14.5 million.
- ARPU slipping back toward $66.
Not counted as evidenceSI (formerly AI) segment revenue and Space segment revenue: the claim is the Connectivity segment line only.
Sources
The model, runnable →
Big Iron attempts orbit by Dec 31, 2028
15% · Reference classResolves 2028-12-31
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimCowboy Space Corporation's Big Iron rocket makes an orbital launch attempt on or before 2028-12-31.
FalsifierNo Big Iron lifts off on an orbital trajectory by 2028-12-31 23:59 UTC. Static fires, hops and suborbital tests do not count; a liftoff on an orbital attempt counts whatever the outcome.
Reference classNew orbital launch vehicles' first flights against their first announced target dates (New Glenn, Neutron), which typically came 1 to 3 years late.
Tesla reports 1 million Optimus robots deployed by end-2031
30% · Scenario modelResolves 2032-04-30
Disclosure: I hold Tesla.
Claim, falsifier, and reasoning
ClaimTesla reports at least 1,000,000 cumulative Optimus robots deployed, inside and outside Tesla, on or before 2031-12-31.
FalsifierTesla's reported cumulative Optimus deployments, inside and outside Tesla, are below 1,000,000 as of 2031-12-31, per its Form 10-K for fiscal 2031 or another Tesla report published by 2032-04-30. If Tesla reports cumulative production but not deployments, cumulative production counts in place of deployments. No qualifying figure of at least 1,000,000 by 2032-04-30 resolves the claim false.
Scenario modelThree scenarios: works and scales, 18% (all clear 1M); works but slower, 32% (40% clear); struggles, 50% (none clear). Total 30.8%, posted at 30%.
OSC lends to a second launch company by Oct 31, 2027
40% · Reference classResolves 2027-10-31
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimThe Department of War's Office of Strategic Capital publicly announces a loan or conditional loan commitment to a launch company other than Stoke Space between 2026-10-26 and 2027-10-31.
FalsifierNo Office of Strategic Capital announcement, between 2026-10-26 and 2027-10-31 23:59 ET, of a loan or conditional loan commitment to a launch company other than Stoke Space. A launch company is one whose primary business includes developing or operating orbital launch vehicles; the loan's purpose does not matter.
Reference classOSC's announced loans to date (MP Materials, August 2025; Energy Fuels, conditional, June 2026; Stoke Space, conditional, 26 October 2026), with judgment for the 1 October 2028 expiry of its authority to make new loans.
First Starship ship-to-ship propellant transfer by Jun 30, 2027
50% · Monte CarloResolves 2027-06-30
Disclosure: I hold SpaceX.
2026-10-23 Source added: NASA, "Artemis Acceleration" (NASA Update on Artemis Program), Kennedy Space Center, 2026-10-23, 2:00 pm ET; on-screen graphic at about 57 minutes: the propellant transfer demo's first vehicle is in build and integration. Also on YouTube (youtube.com/live/pAKqJxuxDgk); event listing at SpacePolicyOnline.
source
Claim, falsifier, and reasoning
ClaimPropellant is transferred between two Starship vehicles in orbit on or before 2027-06-30 (UTC), as confirmed by SpaceX or NASA.
FalsifierNo SpaceX or NASA confirmation, published by 2027-07-07 23:59 UTC, that propellant was transferred between two Starship vehicles in orbit on or before 2027-06-30 23:59 UTC. A transfer between tanks of a single vehicle (as on Flight 3 in 2024), a docking without a confirmed transfer, or a ground test does not count.
Monte Carlo · inputsFirst demo vehicle launch-ready median 4 mo (p90 7.5); second vehicle median 1 mo (p90 3); 3 intervening flights, 25% mishap each, stand-down median 1.5 mo (p90 4); first attempt succeeds 70%, retry 80%; 'early 2027' reported target with a 2.0x remaining-time slip, from this demo's two prior ~12-month slips (NASA OIG IG-26-004). 1M trials; blend of bottom-up and slip-factor models.
How I got 50%
ModelP(YES) = 0.5 × P_A + 0.5 × P_B. Model A builds the date from its parts: first vehicle ready + second vehicle + stand-downs after mishaps on the flights in between + a retry if the first attempt fails, and asks whether a transfer succeeds by 2027-06-30. Model B takes the reported early-2027 target and stretches the remaining time by a slip factor drawn from this demo's own history.
A1 first vehicle ready, months (median; p90 7.5) = 4.00The first demo vehicle, in build and integration on 2026-10-23, is launch-ready in a median of 4 months.Evidence. NASA's Artemis update of 2026-10-23 put the first vehicle in build and integration [UNVERIFIED: livestream, no fetched transcript]. The duration is a judgement.
A2 second vehicle, months after the first (median; p90 3) = 1.00The second vehicle is ready and launched within a median of one month of the first.Evidence. A judgement: the demo needs two vehicles in orbit within the first one's loiter time.
A3 mishap chance on each of 3 intervening flights = 0.25Each of the three flights before the demo has a 25% chance of a mishap, each costing a stand-down of a median 1.5 months (p90 4).Evidence. Judged from the Starship flight record, 2023–2026, which includes failures and stand-downs.
A4 first attempt achieves a transfer = 0.70The first attempt transfers propellant; if not, a retry follows after a median 2.5 months (p90 5) and succeeds 80% of the time. A third attempt is treated as too late.Evidence. A judgement for a first-of-kind cryogenic transfer between two vehicles.
B1 months to the reported target (midpoint of 15 Jan – 15 Apr 2027) = 4.20SpaceX's reported target is early 2027, drawn uniformly between 15 January and 15 April.Evidence. eeNews Europe, 2026-10-23: SpaceX plans an orbital propellant-transfer demonstration early in 2027.
B2 slip factor on remaining time (median; p90 3.1) = 2.00The time left to the target stretches by a median factor of 2.0, never less than 1.0.Evidence. This demo's own history: planned for March 2025, then March 2026, unscheduled as of September 2026 (Wikipedia; NASA OIG IG-26-004). A generic SpaceX milestone slip, median 1.7, gives about 0.59 instead.
ResultModel A (bottom-up): 0.542. Model B (slip factor): 0.487. Blend: 0.514, logged as 0.51 and shown as 50%.
CorrelationThe two models are averaged, not chained. Inside Model A the draws are independent, so a single cause that delays everything at once (a long stand-down that also holds the demo vehicles) is not modelled; that makes Model A somewhat optimistic.
SensitivityOne input at a time, the rest as stated: slip median 1.7 → 59%, 2.4 → 43%; first vehicle median 5.5 months → 44%; mishap 0.40 per flight → 47%; first attempt 0.55 → 48%, 0.85 → 54%.
Signals to watch
Up
- Both demo vehicles shown complete or rolled to the pad.
- A clean Flight 15 with no stand-down.
- A NASA or SpaceX date for the demo inside the first quarter of 2027.
Down
- A mishap on any intervening flight, with an FAA-required investigation.
- The demo moved to the second half of 2027 in a NASA or SpaceX statement.
- The first vehicle reported back in rework.
Not counted as evidenceFlight 3's 2024 transfer between two tanks of one ship, a docking without a confirmed transfer, and ground tests. None of them is a ship-to-ship transfer in orbit.
Sources
The model, runnable →
Updates since 2026-10-23
- 2026-10-23: A1 evidence from [UNVERIFIED] livestream, no link to linked: NASA "Artemis Acceleration", 2026-10-23. The source for the first vehicle's status is now linked (on-screen graphic at about 57 minutes). No input or number changes. New result: 50%. sources
New Glenn flies again by Mar 31, 2027
65% · Reference classResolves 2027-03-31
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimA Blue Origin New Glenn rocket lifts off on or before March 31, 2027.
FalsifierNo New Glenn lifts off between logging and 2027-03-31 23:59 UTC. Static fires, hotfires and scrubs without liftoff do not count; a liftoff that fails does count.
Reference classLaunch vehicles returning to flight after a ground-test anomaly that destroyed the vehicle and damaged its pad, including Falcon 9 after AMOS-6 (2016).
New Glenn flies again by Dec 31, 2026
20% · Reference classResolves 2026-12-31
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimA Blue Origin New Glenn rocket lifts off on or before December 31, 2026.
FalsifierNo New Glenn lifts off between logging and 2026-12-31 23:59 UTC. Static fires, hotfires and scrubs without liftoff do not count; a liftoff that fails does count.
Reference classLaunch vehicles returning to flight after a ground-test anomaly that destroyed the vehicle and damaged its pad, including Falcon 9 after AMOS-6 (2016).
A second Vikram-1 flies by Jul 18, 2027
80% · Reference classResolves 2027-07-18
Claim, falsifier, and reasoning
ClaimA second Skyroot Vikram-1 flight lifts off on or before July 18, 2027.
FalsifierNo second Vikram-1 lifts off by 2027-07-18 23:59 UTC. A scrub without liftoff, a static fire or a ground test does not count; a liftoff that fails to reach orbit does count as a flight. 'By its one-year anniversary', as posted, is read as on or before 2027-07-18, one year after Vikram-1's first orbit on 2026-07-18.
Reference classTime from first to second flight for new small launch vehicles.
A second Vikram-1 flies by Dec 31, 2026
35% · Reference classResolves 2026-12-31
Claim, falsifier, and reasoning
ClaimA second Skyroot Vikram-1 flight lifts off on or before December 31, 2026.
FalsifierNo second Vikram-1 lifts off by 2026-12-31 23:59 UTC. A scrub without liftoff, a static fire or a ground test does not count; a liftoff that fails to reach orbit does count as a flight.
Reference classTime from first to second flight for new small launch vehicles.
2026-09-003 · logged 2026-10-15 · Space
SpaceX lists a per-kg Starship price to Earth orbit on spacex.com before 2028
20% · Monte CarloResolves 2028-01-31
2026-10-23 Shown as 20% from 23 Oct 2026, when every entry moved to the nearest 5%. W1 published it as 19%; the stored number, 0.19, is unchanged and is the one graded.
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimBefore 2028-01-01, SpaceX lists on its own website (spacex.com, including the rideshare page) a Starship launch price per kilogram to Earth orbit, or a Starship rideshare price with a per-kg rate to Earth orbit.
FalsifierAs of 2028-01-01, spacex.com (including the rideshare page) lists no Starship launch price per kilogram to Earth orbit and no Starship rideshare price with a per-kg rate to Earth orbit, assessed against the live site and Internet Archive snapshots of it. Not counted: prices to the Moon, Mars or other destinations beyond Earth orbit (including the Mars cargo rate); per-kg figures computed from a launch price and a payload mass; customer regulatory filings; executive statements or posts; cost targets; press reports; private contract prices.
Monte Carlo · inputsFour judged factors in the model P(YES) = R × K × L + (1 − R × K × L) × O, drawn as correlated Beta distributions: Starship credible enough to sell 2029 seats (R), rideshare moved to Starship (K), a per-kg price listed in time (L), any other route (O). The reasoning, the sources and the runnable model are published with the entry.
How I got 20%
ModelP(YES) = R × K × L + (1 − R × K × L) × O
R = 0.45By mid-to-late 2027, Starship is credible enough that SpaceX will sell 2029 seats to outside customers.Evidence. First orbital flight on 28 Sep 2026 (Flight 14; 26 Starlink V3 satellites deployed, early return after an engine shutdown). A ship catch attempt is planned for Flight 15, not yet flown. Musk's condition for winding down Falcon is Starship "flying reliably several times per week" (X post of 22 Aug 2026), far above the current few flights a year. My judgement: his timelines have historically slipped two to three times.
K = 0.50SpaceX keeps rideshare and moves it to Starship, rather than dropping it or leaving it on the few remaining Falcon 9 missions.Evidence. SpaceX reportedly is not booking Falcon 9 beyond 2028 (single-source report, Sep 2026). Industry reports say it may stop all but a few Falcon 9 missions from 2028. Rideshare is small next to Starlink, but SpaceX dominates it.
L = 0.65Given R and K, SpaceX lists a per-kg price on spacex.com before 2028-01-01.Evidence. SpaceX already publishes Falcon 9 rideshare pricing: $350,000 for up to 50 kg to sun-synchronous orbit, then $7,000/kg for 2026 (SpaceX's own price feed, read 11 Oct 2026). The feed's ladder runs to 2027 and does not mention Starship. Timing risk: rideshare sells 12 to 24 months ahead, so a listing for a late-2029 flight could land in early 2028, after the deadline.
O = 0.04Any other route: an early listing for marketing, or a response to competitors.Evidence. Seven years of Starship development with no listing so far.
ResultWith independent factors: 0.45 × 0.50 × 0.65 = 0.146, and 0.146 + 0.854 × 0.04 = 0.180.
CorrelationThe factors move together: if Starship works, the rest become more likely. A Monte Carlo with a moderate correlation (0.3 between R and K; 0.15 between each of them and L) gives a mean of 0.187, so 19%. 80% range: 8% to 31%.
SensitivityEach factor moved 0.15 up and down, the others fixed: R 13% to 23%; K 14% to 22%; L 15% to 21%.
Signals to watch
Up
- A ship catch on Flight 15.
- A reflight of a recovered ship.
- The first Starship launch from LC-39A.
- Falcon 9 Transporter missions disappearing from the 2029-and-later rideshare manifest.
- A 2028 or later row in SpaceX's rideshare price feed that names Starship.
- Any Starship option appearing on spacex.com/rideshare. That resolves the question yes.
Down
- Repeated Starship failures or a long stand-down.
- SpaceX confirming that Transporter missions continue on Falcon 9 into 2029 and later.
- A 2028 row in the price feed that stays on Falcon 9.
- Rideshare quietly discontinued.
Not counted as evidenceA $90M price for a dedicated Starship launch on a 2029 mission appeared in a customer's regulatory filing (reported Mar 2026). It is a customer filing, per launch and not per kilogram, and not on spacex.com, so it does not resolve this question.
Sources
- Starship Flight 14 live updates · Space.com · 2026-09-28 · press · archived
- Starship Reaches Orbit on 14th Test Flight · Payload · 2026-09-28 · press · archived
- Goodbye, Falcon 9: SpaceX planning to retire workhorse launcher when Starship is up and running · Space.com, citing Musk's X post of 22 Aug 2026 · 2026-08-24 · press · archived
- SpaceX has stopped taking Falcon 9 bookings past 2028, betting everything on Starship · TechSpot; secondary, single-source report · 2026-09-24 · secondary report · archived
- SpaceX rideshare pricing feed · SpaceX; read 11 Oct 2026: $7,000/kg for 2026, 50 kg minimum · 2026-10-11 · primary · archived
- SpaceX rideshare pricing as of February 2026 · New Space Economy, citing spacex.com/rideshare · 2026-02-27 · secondary report · archived
- SpaceX Starship Launch Price Set at $90 Million for 2029 Mission · IndexBox, citing a customer's annual filing · 2026-03 · secondary report · archived
- Starship, baseline for resolution · SpaceX; the page as it stood when the forecast was made · 2026-10-11 · primary · archived
- Rideshare, baseline for resolution · SpaceX; the page as it stood when the forecast was made · 2026-10-11 · primary · archived
The model, runnable →
2026-10-028 · logged 2026-10-02 · Defense
Congress passes a full-year Defense spending bill by March 31, 2027
60% · Reference classResolves 2027-03-31
Claim, falsifier, and reasoning
ClaimA regular FY2027 Department of Defense appropriations act, standalone or as a division of an omnibus or minibus, is signed into law on or before March 31, 2027.
FalsifierNo regular FY2027 Department of Defense appropriations act is signed into law by 2027-03-31 23:59 ET, per Congress.gov's public-law record. A full-year continuing resolution, or any continuing resolution extending Defense funding past March 31, 2027 without a regular act enacted by that date, resolves this false.
Reference classRegular DoD appropriations acts for FY2017-FY2026: 8 of 10 enacted by March 31 of the fiscal year (misses FY2017, enacted May 5, 2017, and FY2025, a full-year CR); adjusted down because the midterm election falls inside the current CR window.
2026-09-027 · logged 2026-09-24 · Space
Space Epoch's Yuanxingzhe-1 reaches orbit before the end of 2026
20% · Reference classResolves 2026-12-31
Claim, falsifier, and reasoning
ClaimSpace Epoch's Yuanxingzhe-1 reaches Earth orbit on or before 2026-12-31.
FalsifierNo Yuanxingzhe-1 launch places a payload or upper stage into Earth orbit completing at least one full revolution by 2026-12-31 23:59 UTC, counted from the public orbital catalog. Every attempt failing to reach orbit resolves this false. First-stage recovery is not required. Suborbital hops, static fires and transatmospheric trajectories that fall short of one revolution do not count.
Reference classFirst flights of new orbital launch vehicles, 2006-2026: 4 of 27 flew in the year first announced, and 9 of 23 maiden attempts reached orbit; adjusted up for a near-horizon restatement and for Chinese commercial debuts running above the global rate, adjusted down because the end-2026 date is already a slip from 2025 and has stood seven months without a flight.
2026-09-026 · logged 2026-09-21 · Space
Rocket Lab closes its Iridium acquisition by mid-2027
35% · Reference classResolves 2027-06-30
Claim, falsifier, and reasoning
ClaimRocket Lab completes its acquisition of Iridium Communications on or before June 30, 2027.
FalsifierNo completed closing by June 30, 2027, including termination of the merger agreement or a closing after that date.
Reference classLarge U.S. space and telecom acquisitions requiring shareholder approval, HSR clearance, and FCC license transfer, measured against their announced closing targets; adjusted up for cash component already financed and Rocket Lab's three completed 2026 acquisitions.
2026-09-025 · logged 2026-09-21 · Cross-sector
Katalyst beams power from orbit to a ground receiver by 2029
5% · Reference classResolves 2029-12-31
Claim, falsifier, and reasoning
ClaimA Katalyst spacecraft under the ASGARD program transmits power from orbit that is received and converted to electricity by a ground receiver on or before December 31, 2029.
FalsifierNo public confirmation by Dec 31, 2029 that an ASGARD spacecraft beamed power from orbit to a ground receiver that converted it to electricity. Does NOT count: launch or on-orbit assembly without ground reception, spacecraft-to-spacecraft beaming, ISS or lab demonstrations, ground-to-ground tests, or other programs (SSPIDR/Arachne, Caltech, DARPA).
Reference classFirst-of-kind in-space robotic assembly and space-to-ground power beaming demonstrations; adjusted down for Katalyst's abandoned Swift boost attempt (LINK, launched July 2026, attitude control problems) and for no published ASGARD schedule.
2026-09-024 · logged 2026-09-21 · Space
True Anomaly's GHOST-R satellite enters RG-XX service by 2029
15% · Reference classResolves 2029-12-31
Claim, falsifier, and reasoning
ClaimA True Anomaly-built GHOST-R spacecraft is accepted into operational RG-XX service on or before December 31, 2029.
FalsifierNo official confirmation that a True Anomaly GHOST-R vehicle entered operational RG-XX service by Dec 31, 2029. Launch alone, demonstrations, follow-on prototype awards, or a Northrop-only transition do not count.
Reference classDIU prototype-to-program-of-record transitions; adjusted up for explicit 'direct transition' language, down for a 2028 launch leaving about one year for demonstration and acceptance.
2026-09-023 · logged 2026-09-21 · Space
Starship deploys operational Starlink V3s in orbit by Oct 31
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimA Starship flight reaches orbit and deploys Starlink V3 satellites intended for operational use on or before October 31, 2026.
FalsifierNo confirmed orbital insertion and deployment of operational Starlink V3 satellites from Starship by Oct 31, 2026. Suborbital flights, mass simulators, and non-operational test satellites do not count.
Reference classStarship flight outcomes 2023–2026 and first-of-type operational payload deployments; adjusted for a stated near-term launch target.
2026-09-022 · logged 2026-09-21 · Space
Starship lifts off with operational Starlink V3s by Sep 30
Disclosure: I hold SpaceX.
Claim, falsifier, and reasoning
ClaimA Starship flight carrying Starlink V3 satellites intended for operational use lifts off on or before September 30, 2026.
FalsifierNo liftoff of a Starship carrying Starlink V3 satellites by Sep 30, 2026 (UTC). A flight without V3 satellites, or a scrub without liftoff, resolves false.
Reference classStarship launch dates publicly stated by Elon Musk or SpaceX within ~7 days of the target, 2023–2026; adjusted for the booster already rolled to the pad.
2026-09-021 · logged 2026-09-21 · Energy
Radiant finishes all five DOME test phases in 2026
45% · Reference classResolves 2026-12-31
Disclosure: I hold a position in a Radiant competitor.
Amendment · 2026-10-08 · FalsifierNo public confirmation that all five phases of the Kaleidos DOME test campaign, including the 150-hour continuous run without operator intervention, were completed on or before Dec 31, 2026.The claim requires all five phases; the falsifier as logged checked only the 150-hour run. The record as logged is unchanged and shown below. This entry is graded on the amended falsifier.
Claim, falsifier, and reasoning
ClaimRadiant completes all five phases of its Kaleidos DOME test campaign, including the 150-hour continuous run without operator intervention, on or before December 31, 2026.
FalsifierNo public confirmation that the 150-hour continuous run finished on or before Dec 31, 2026.
Reference classFirst-of-a-kind reactor power ascension schedules; adjusted for a stated Q3 2026 target, a missed July 4 criticality target, and one quarter of slip allowance.
2026-09-020 · logged 2026-09-21 · Energy
Radiant's Kaleidos goes zero-power critical by Sep 30
Disclosure: I hold a position in a Radiant competitor.
Claim, falsifier, and reasoning
ClaimRadiant's Kaleidos microreactor achieves zero-power criticality at INL's DOME facility on or before September 30, 2026.
FalsifierNo public confirmation by Radiant, INL, or DOE that criticality occurred on or before Sep 30, 2026.
Reference classDOE-authorized reactor startups under the 2025-26 pilot and authorization pathway, time from facility arrival to criticality; adjusted for Radiant's missed July 4 target and Aug 25 DOME arrival.
2026-09-019 · logged 2026-09-21 · Space
A U.S. nuclear thermal rocket fires in space by 2032
5% · Reference classResolves 2032-12-31
Claim, falsifier, and reasoning
ClaimA nuclear thermal propulsion engine built by or for a U.S. government agency or U.S. company operates in space, with the reactor critical and heated propellant producing thrust, on or before December 31, 2032.
FalsifierNo public confirmation of an in-space nuclear thermal engine firing by Dec 31, 2032. Does NOT count: ground tests, nuclear electric propulsion (including SR-1 Freedom), surface or orbital power reactors, or a launch in which the engine never fires.
Reference classU.S. space nuclear propulsion programs 1955–2025 (Rover/NERVA, SP-100, Timberwind, Prometheus, DRACO, JETSON): none reached flight. Adjusted for current federal momentum directed to nuclear electric propulsion and surface power rather than nuclear thermal.
2026-09-018 · logged 2026-09-21 · Space
SR-1 Freedom nuclear spacecraft launches by end of 2028
10% · Reference classResolves 2028-12-31
Claim, falsifier, and reasoning
ClaimNASA's Space Reactor-1 Freedom (or a renamed successor that is NASA's fission-powered nuclear electric propulsion spacecraft to Mars) lifts off on or before December 31, 2028.
FalsifierNo liftoff with the reactor aboard by Dec 31, 2028. A launch without the reactor, a descoped non-nuclear mission, or a slip to the 2031 Mars window resolves false. Reactor startup after launch is not required.
Reference classNASA first-of-kind technology demonstrations tied to fixed planetary launch windows; adjusted down for the first U.S. space reactor since SNAP-10A (1965) with the reactor as the primary new system, adjusted up for scope limited to existing hardware (repurposed Gateway PPE).
2026-09-017 · logged 2026-09-21 · Defense
Golden Dome declared initially operational by end of 2028
60% · Reference classResolves 2028-12-31
Claim, falsifier, and reasoning
ClaimThe Department of Defense, the Office of Golden Dome for America, or U.S. Northern Command publicly declares initial operational capability (or an explicitly equivalent 'initial capability operational' status) for the Golden Dome homeland missile defense architecture on or before December 31, 2028.
FalsifierNo official declaration of IOC or explicit equivalent for the Golden Dome architecture as a whole by Dec 31, 2028. Does NOT count: activation of individual operational sites (including Fort Story, VA, announced Aug 2026), component tests, demonstrations, contract awards, or relabeling of existing systems without an architecture-level declaration.
Reference classMajor DoD programs with politically fixed IOC dates; adjusted up because Phase 1 integrates systems already in production and the declaration is administrative, adjusted down for interceptor delivery lead times and program-office delays through 2026.
2026-09-016 · logged 2026-09-21 · Cross-sector
Congress preempts state AI laws by Jan 3, 2027
10% · Reference classResolves 2027-01-03
Claim, falsifier, and reasoning
ClaimA federal statute expressly preempting, or imposing a moratorium on, state laws that regulate AI models or systems generally is signed into law before the 119th Congress ends on January 3, 2027.
FalsifierNo such statute signed by Jan 3, 2027. Does NOT count: executive orders (including EO 14365), agency rules or FCC/FTC actions, court rulings, discussion drafts, bills passing one chamber only, or sector-specific preemption (e.g., child online safety or privacy bills) that does not reach state AI laws generally.
Reference classCongressional attempts to preempt state AI laws via standalone or must-pass bills, 2025–26 (Senate stripped the OBBBA moratorium 99-1; omitted from the FY2026 NDAA); adjusted for the Obernolte–Trahan draft remaining unintroduced as of Sep 2026 and a lame-duck FY2027 NDAA as the remaining vehicle.
