Rocket Report: Some Canadians don’t want a spaceport; Falcon hits 600 landings

"If it doesn’t rely on a solid, there’s no reason why we can’t launch."

See full article...

 

[dennis.kext]

I hope it's not true that SpaceX will buy Cursor. It was easy to compartmentalize SpaceX when it was just "rockets and satellites", but now they are adding AI slop to their product mix.

AI yai yai!

 

[MST2.021K]

SpaceX estimates that its total addressable market—the maximum revenue a company could generate if it captured every customer in a particular market—could be as much as $28.5 trillion.

Hey, look at me! I can make up fake numbers too! "We aim to launch 69 sextillion rockets this century."

 

[peterford]

Astrobotic tests rotating detonation engine … During testing, each engine produced more than 4,000 pounds of thrust

Draco - 90 lbf
SuperDraco - 16,400 lbf (1)
RL-10 - 24,700 lbf (2)

 

[Lexomatic]

SpaceX estimates that its total addressable market—the maximum revenue a company could generate if it captured every customer in a particular market—could be as much as $28.5 trillion. [... of which] $26.5 trillion, to stem from the AI sector [of which] $22.7 trillion, could come from AI for businesses.

Some context to quantify why that sounds bonkers: Total world GDP in 2025 was around $120 trillion, of which the services sector was $100 trillion. (The conventional three-sector breakdown of economic activities is primary: agriculture; secondary/industry: mining, manufacture, utilities, construction; tertiary/services: trade, transportation, ..., government.) I surmise SpaceX's premise is "AI-software will replace the output attributable to humans in all information, communication, financial, insurance, real estate, professional and administrative roles," or "AI-driven robots will replace all physical activities" or "AI will increase world GDP, even if they replace only a fraction of human labor" -- which is unlikely any time soon, but is consistent with Musk's recent futurist statements (Business Insider Nov 2025, Forbes.com Nov 2025, Fortune Dec 2025 (paywalled), AOL Apr 2026) and insistence that Optimus will soon be the major product of Tesla.

(Source: I work in international macroeconomics.)

Edit: Added supporting links, a third possible premise, phrasing.

 

[fenris_uy]

I hope it's not true that SpaceX will buy Cursor. It was easy to compartmentalize SpaceX when it was just "rockets and satellites", but now they are adding AI slop to their product mix.

AI yai yai!

SpaceX already bought Twitter and xAI. So they are already a AI Slop company.

 

[schnackenpfefferhausen]

SpaceX expects more than 90 percent of that market, or $26.5 trillion, to stem from the AI sector.

View: https://m.youtube.com/watch?v=Fkk9DI-8el4&t=26s&pp=2AEakAIB

 

[fancysunrise]

Hey, look at me! I can make up fake numbers too! "We aim to launch 69 sextillion rockets this century."

Fake it till people give you so much money that you can throw it in enough directions until something makes it, or not, and repeat

 

[syl1979]

Thanks a lot for all your reports and articles.
I was very happy to see you on french/german TV on this documentary about Artemis 2

 

[jason8957]

I can't help but just think that Musk is just smearing "AI" all over SpaceX in an attempt to boost the valuation at the IPO. He is mainly hype and this is his style. And it seems like any company mentions "AI" and the dumb market automatically boosts their value.

 

[MilanKraft]

Hey, look at me! I can make up fake numbers too! "We aim to launch 69 sextillion rockets this century."

I have never trusted musk entirely, but I officially lost all hope in SpaceX at: "1,000,000 unit AI data center in space, to spread consciousness to the stars". Also, Starship is a steaming turd of a design, and I have rocket engineering kind in my circle of contacts to inform this opinion. Musk basically uses the name, bigness factor, and the crane capture gimmick to capture the public's attention, but in terms of practicality and achieving real milestones in space exploration, it's more gimmick than practical solution...

...but back to the larger crux: at some point it becomes insulting to listen to this bullcrap. A million units (napkin math will show would take 100+ years)? Consciousness (brainless pattern recognition bot good ≠ conscious, understanding entity)? "To the stars" (AFAIK, we've barely left our own solar system — from systems launched decades ago — much less reached other nearby star systems)? Please fuck off already, you grifting sociopath. How about we not talk "to the stars" until we've actually invented "to the stars" type technology (radically new propulsion systems, etc)? Or is that just too much to ask?

The level of con-artistry and russian doll business model subterfuge is so bad that nobody should be rooting for Space X at this point, unless they simply don't care how businesses behave / subscribe to a "by any means necessary" ethic. There is no separating the engineering and "space is cool" aspects from the man anymore, IMHO.

I get there are space geeks who will say there are still ordinary, non-political rocket engineering types working at SpaceX in order to make launches and recoveries happen, so it's OK to root for them, but IMHO it doesn't matter anymore. I simply cannot root for them on any level. I will be rooting instead for them to fail because principles matter / how we get where we're going... matters. Those "ordinary people" have made a choice to stick with this despicable, conartist asshole — and unless they're blind, they now know for sure this is who he is — so they have also chosen to be stained by his actions. They are now his enablers.

For my part, bring on the "unscheduled, rapid disassemblies" en masse. Explosions can be entertaining and educational. The sooner NASA rules them out for manned flight, the better.

(Apologies to the downvoters for mixing rockets and politics, but sometimes it is impossible to separate them — this is one of those times.)

 

[Lexomatic]

Incremental summary of rocket launches by SpaceX and others

• Time frame: 2026 YTD through Thu 4/23 (day 113 of 365, week 17 of 52, 0.31 year)
• Increment: 4/10 to 4/23 (2 weeks)
• By SpaceX: 49 (+3 w/w), of which 49 by Falcon 9, of which 40 Starlink (82%)
• By others: 50 (+4 w/w) by 18 providers of 7 nationalities [1], of which 25 by China (50%)

Spoiler: Launch performance by SpaceX

Launches completed, during increment and YTD

• Total launches: +7=49
• By customer/contract type:
  • Internal: +5=40 (Starlink)
  • External/paying: +2=9 (ISS: +1=2, other U.S. gov: +1=2, other: 5) [2]
  • Test: 0 (Starship-Super Heavy)
• Payload upmass:
  • Potential: >866 ths. kg [11]
  • Known and estimated: >640 ths. kg (~1.6 ISS-masses) [3,4,11]
  • Known excl. Starlink: >27 ths. kg
  • Satellite count: >1,230
• By launch site: CCSFS: 25, KSC: 0, VSFB: 24, Starbase: 0
• By landing site:
  • Florida: ASDS ASOG: 12, JRTI: 11, RTLS LZ-1: 0, LZ-2: 0, LZ-40: 2, expend: 0 [5]
  • VSFB: ASDS OCISLY: 21, RTLS LZ-4: 3
  • Controlled splashdown of Starship (success/attempt): 0/0

Performance goals (F9 and FH), during 2026
Launch rate:

• In five prior years: 31, 61, 83, 138, 165
• Goal for this year: None stated [12]
• Achieved: 49
• Turnaround of core stages, in days: *** [7]

On-time statistics for external/paying launches:

• Count of launches: 9
• Days of delay, mean: 2.2
• Days of delay, spectrum (days x count): 0x3, 1x2, 2x2, 6x2, 8
• Exceptional delay(s): None

On-time statistics for internal/Starlink launches:

• Count of launches: 40
• Days of delay, mean: 4.6
• Days of delay, spectrum: 0x12, 1x4, 2x9, 3x2, 4x3, 5x3, 6, 9, 16x2, 18, 62

Spectating a SpaceX launch, during 2026 (probabilities)
At Florida's Space Coast:

• No more than one day after initial schedule: 0.44 (=11/25) [8]
• On any given day: 0.22
• Within any given week: 0.79
• Weeks without a launch: 3/17
• During dawn, daylight or dusk hours: 0.48 (=12/25)
• RTLS: 0.00 (=0/25) [9,10]

At VSFB:

• Within any given week: 0.77
• Weeks without a launch: 1/17
• RTLS: 0.13 (=3/24) [9,10]

Spoiler: Fleet utilization by SpaceX

F9 and FH cores, annual summary:

Year	2023	2024	2025	2026 YTD
Added	***	2	8	2
Expended	4	6	2	0
Lost	***	1	1	0
Reuse, mean (flights per core)	***	5.3	6.6	2.3
Reuse, range	***	1 to 10	1 to 11	1 to 4
Turnaround (days)	***	35.2	44.7	***
Falcon Heavy launches	5	2	0	0

Detail by tail number

• 2024
  • Added - 2: B1072, 85
  • Expended - 6: B1060, 61, 64(s), 65(s), 87(c), 89(c)
  • Lost - 1: B1062
• 2025
  • Added - 8: B1091, 92, 93, 94, 95, 96, 97, 1100
  • Expended - 2: B1073, 76
  • Lost - 1: B1086
• 2026
  • Added - 2: B1101, 1103
  • Expended - 0
  • Lost - 0

Fleet, by location (surviving/total):

• East coast: 14/17 - B1067, 69, 73[-], 76[-], 77, 78, 80, 83, 85, 86[-], 90, 91, 92, 94, 95, 96, 1101
• West coast: 10/10 - B1063, 71, 75, 81, 82, 88, 93, 97, 1100, 1103

Flights, over comparative periods

• By model, during 2024-2026: F9 327, FH 2, SS-SH 9
• By core number, during 2026 only:
  • 1 flight: 4 - B1092, 94, 96, 1103
  • 2 flights: 8 - B1063, 67, 69, 71, 77, 88, 90, 1101
  • 3 flights: 7 - B1078, 80, 82, 81, 85, 93, 95
  • 4 flights: 2 - B1097, 1100
• By core number, over full lifetime:
  • 1 flight: B1072(s), 87(c), 89(c), 1103
  • 2 flights: B1101
  • 3 flights: B1091
  • 5 flights: B1086(s), 95, 1100
  • 6 flights: B1064(s), 65(s)
  • 7 flights: B1094, 95
  • 8 flights: B1097
  • 10 flights: B1092
  • 11 flights: B1090
  • 12 flights: B1093
  • 14 flights: B1088
  • 15 flights: B1085 [6]
  • 16 flights: B1083
  • 21 flights: B1082, 75
  • 22 flights: B1076[-]
  • 23 flights: B1081
  • 26 flights: B1080
  • 27 flights: B1077, 78
  • 30 flights: B1069
  • 32 flights: B1063, 71
  • 34 flights: B1067

Missions by Dragon capsules, over full lifetime

• Test vehicles and Cargo Dragon 1 (flown 2010 to 2020) - 20 missions by 11 capsules:
  • C103: 1 (CRS-1)
  • C104: 1 (CRS-2)
  • C105: 1 (CRS-3)
  • C106: 3 (CRS-4, 11, 19)
  • C107: 1 (CRS-5)
  • C108: 3 (CRS-6, 13, 18)
  • C109: 1 (CRS-7)
  • C110: 2 (CRS-8, 14)
  • C111: 2 (CRS-9, 15)
  • C112: 3 (CRS-10, 16, 20)
  • C113: 2 (CRS-12, 17)
• Test vehicles and Cargo Dragon 2 (flown 2019 to present) - 15 missions by 8 capsules:
  • C201: Ground prototype, status: retired
  • C202: Ground prototype, status: retired
  • C203: Ground test article, status: active
  • C204: 1 (Demo-1) status: destroyed
  • C205: 1 (In-flight abort test) status: retired
  • C208: 5 (CRS-21, 23, 25, 28, 31)
  • C209: 5 (CRS-22, 24, 27, 30, 32)
  • C211: 3 (CRS-26, 29, 33)
• Crew Dragon 2 (flown 2020 to present) - 20 missions by 5 capsules:
  • C206 Endeavour: 6 (Demo-2, Crew-2, Ax-1, Crew-6, 8, 11)
  • C207 Resilience: 4 (Crew-1, Inspiration4, Polaris Dawn, Fram 2)
  • C210 Endurance: 4 (Crew-3, 5, 7, 10)
  • C212 Freedom: 5 (Crew-4, Ax-2, Ax-3, Crew-9, 12)
  • C213 Grace: 1 (Ax-4)
• United States Deorbit Vehicle (USDV) - under development

Spoiler: Fleet utilization by other providers

This section will be more interesting once alternatives to SpaceX's Falcon 9, U.S.-domiciled or otherwise, come into service.

U.S. - Virgin Galactic (suborbital, program on hiatus):

• SpaceShipOne: 17 test (2003-2004)
• VSS Enterprise: >30 test (2010-2014)
• VSS Unity: 25 test (2016-2023) and 7 passenger (2023-2024)

U.S. - Blue Origin New Shepard (suborbital, program on hiatus):

• NS2: 5 (NS-2 to 6)
• NS3: 9 (NS-7 to 13, 17, 23)
• NS4: 17 (NS-14 to 16, 18 to 22, 24 to 26, 28, 30, 32, 34, 36, 38)
• NS5: 6 (NS-27, 29, 31, 33, 35, 37)
• RSS Jules Verne: 7
• RSS First Step: 16 (..., NS-30, 32, 34, 36, 38)
• RSS H. G. Wells: 12 (..., NS-29, 35)
• RSS Kármán Line: 4 (NS-27, 31, 33, 37)

U.S. - Blue Origin New Glenn:

• GS1-SN001 So You're Telling Me There's a Chance: 1
• GS1-SN002 Never Tell Me the Odds: 2
• GS1-SN003 No, It's Necessary: 0

U.S. - USAF/USSF Boeing X-37 aka OTV (spaceplane):

• Vehicle 1: 4 (2010, 2012, 2020, 2025)
• Vehicle 2: 4 (2011, 2015, 2017, 2023)

China - Reusable Experimental Spacecraft (spaceplane):

• One or more vehicles: 4 (2020, 2022, 2023, 2026)

Spoiler: Launch counts by other providers, during increment

Total: +7=50
Orbital providers who have launched, by nationality (49):

• China: +3=25 - CASC: +1=16, CAS Space: +1=2, Chinarocket: +1=3, Deep Blue: 0, ExPace: 1, Galactic Energy: 2, i-Space: 0, LandSpace: 0, Orienspace: 0, SAST: 0, Space Pioneer: 1
• Europe [1]: 1 - Arianespace: 1
• Germany: 0 - HyImpulse: 0, Isar: 0 (pending)
• India: 1 - Agnikul: 0, ISRO: 1
• Japan: 1 - JAXA/MHI: 0, Space One: 1
• Korea, South: 0 - Innospace: 0, KARI: 0
• Russia: +1=6 - Roscosmos: 3, VKS: +1=3
• U.S.: +3=15 - ABL: 0, Blue Origin: +1=2, Firefly: 1, NASA: 1, Northrop: 0, Relativity: 0, Rocket Lab: +2=7, ULA: 2, USDOD : 0

Suborbital providers (1):

• Canada: 0 - Concordia University: 0
• UAE: 1 - TII: 1
• U.S.: 0 - Virgin Galactic: 0

Infrequent or aspiring providers, by nationality (0):

• Australia: 0 - Gilmour: 0
• Canada: * - MLS, NordSpace, Reaction Dynamics
• China: * - LinkSpace, OneSpace, Space Epoch, Space Pioneer
• Denmark: 0 - Copenhagen Suborbitals
• Germany: * - RFA
• France: 0 - Latitude, Maiaspace
• Iran: 0 - IRGC: 0, ISA: 0
• Italy: 0 - Sidereus
• Japan: * - Interstellar Technologies
• Korea, North: 0 - NATA: 0
• Korea, South: * - Perigee
• New Zealand: 0 - Dawn Aerospace
• Netherlands: 0 - T-Minus
• Poland: 0 - LIA: 0, SpaceForest: 0
• Spain: 0 - Pangea, PLD Space
• Sweden: 0 - Pythom
• U.K.: 0 - Astron, Orbex
• U.S.: * - Radian, Stoke Space, Vaya

Spoiler: Abbreviations, methodology, endnotes

Abbrevations
n/a=not available, n/c=not calculated, w/w=week over week, BEO=beyond Earth orbit, F9=Falcon 9 Block 5, FH=Falcon Heavy, (c)=Falcon Heavy center core, (s)=Falcon Heavy side core, ASDS=autonomous spaceport drone ship, CCSFS=Cape Canaveral Space Force Station, GTO=geosynchronous transfer orbit, KSC=Kennedy Space Center, LEO=low Earth orbit, LZ=landing zone, MEO=medium Earth orbit, NLT=no later than, RTLS=return to launch site, SSO=sun-synchronous orbit, VSFB=Vandenberg Space Force Base, YTD=year to date.

For entities and programs: CASC=China Aerospace Science and Technology Corporation, CRS=NASA Commercial Resupply, IRGC=Islamic Guard, ISRO=Indian Space Research Organization, ISA=Iranian Space Agency, JAXA=Japan Aerospace Exploration Agency, LIA=Łukasiewicz Institute of Aviation, MHI=Mitsubishi Heavy Industries, NATA=National Aerospace Technology Administration f.k.a. National Aerospace Development Administration (North Korea), RFA=Rocket Factory Augsburg, SAST=Shanghai Academy of Spaceflight Technology, ULA=United Launch Alliance, U.S.=United States, VKS=Russian Aerospace Forces.

Methodology

• Counts include orbital launches; suborbital launches by entities with aspirations to be orbital, including tourist flights; complete success, qualified success, failure (i.e., launched but terminated); tests and production. They exclude ground tests, pad aborts and unintended launches (i.e., Space Pioneer 6/30).
• Date format: month/day, i.e., U.S. style.
• Annualized rates assume 366 days for year 2024.
• Metrics and formatting are subject to change.
• Counts and calculations are subject to correction.
• Figures are rounded.
• SpaceX launch mass assumes 730 kg for Starlink v2 Mini (per Wikipedia) and Starlink v2 Mini w/ D2C (they're reportedly heavier, but by an unknown amount); 575 kg for v2 Mini Optimized; 5,000 kg for Transporter-10; >12,000 kg for Crew-8; USSF-124 is unknown; Starship IFTs are excluded.
• Launch mass fraction assumes 419,725 kg for ISS.
• Unless otherwise stated, launch rate, fleet utilization and similar calculations include only the F9 fleet, not the (much less common) FH.
• Delay stats are subject to correction if initially-announced schedule (within a sequence of scrubs) is corrected. Aggregates will be marked "n/a" if I have incomplete data.
• Launch probability uses the Poisson method, per Wickwick 7/5.
• Nationality of provider vs. launch site: Rocket Lab is a U.S.-domiciled company but launches from New Zealand. Orbex is British but plans to build engines in Denmark. RFA is German but will launch from the U.K.
• Sources: EverydayAstronaut.com (previous and upcoming), Ken's Launch Schedule, NASA.gov (ISS CRS mission overview PDF), NASASpaceFlight.com, NextSpaceflight.com, Space.com, Space Explored, SpaceX.com, SpaceX.com X feed, SpaceflightNow.com, Wikipedia.

Endnotes

• Note 1. "Nationality" counts "European Union" separately from individual member states, to account for Arianespace launches under the aegis of ESA.
• Note 2. Customer allocation: Some early launches of Starshield were believed to be mixed with Starlink, so they would count towards both "internal" and "U.S. government excl. ISS." Later launches are believed to be categorized openly as NROL.
• Note 3. Starting with Starlink Group 7-9 in January 2024, most launches have manifested a mix of sats with D2C capability. The number of each type is announced by SpaceX, but the mass of the latter is not known.
• Note 4. Sea-level mass is not necessarily comparable to delta-V, given differences attributable to orbital inclinations. In other words, "we've lofted three ISS-masses this year" is not the same as "we could've put three duplicates in the ISS's orbit."
• Note 5. A FH counts for one launch but three landings.
• Note 6. Counting the fleet is tricky because cores may be used as singletons or FH side cores, e.g., B1064 and B1065, which flew five times as F9 and a final time as expended FH.
• Note 7. Turnaround time of F9 cores: Trimmed mean of most recent 20 flights, i.e., start with >=20 flights, pick most recent 20, drop the outliers at each extreme, and average the remaining 18. Unlike other YTD stats, this count spans year-boundary. Uses launcher identity per NextSpaceflight.com.
• Note 8. Most (>65% during 2024) launches are F9 carrying Starlink and hence fungible for spectating purposes, so a long delay attributable to any particular mission (e.g., 10-2 was delayed 12 days while B1073 was swapped for B1078, during which 9-1 and SES-24 launched) doesn't reduce the chance of seeing a launch in general.
• Note 9. Starlink launches always land downrange on a drone ship, so RTLS is a possibility only on the 30% that are paying-customer launches.
• Note 10. A FH dual landing counts as one RTLS viewing opportunity.
• Note 11. Maximum mass, to the launch orbit, with recovery. Crew and Cargo Dragon missions are excluded, because they "are" satellites but do not "deliver" satellites. Potential mass accounts for NRO and rideshare missions that have unspecified payloads to known orbits, i.e., F9 carries 16,000 kg to LEO and 6,500 kg to SSO.
• Note 12. Launch goal during 2024 for SpaceX is per comments of VP Jon Edwards at Everyday Astronaut Astro Awards 1/14/2024; includes four launches delayed from 2023. Goal for 2025 is per comments by Elon Musk on X on 12/17/2024 (>180) and Gwynne Shotwell at CSIS (175-180).

 

[Fatesrider]

But where money does that come from? … SpaceX expects more than 90 percent of that market, or $26.5 trillion, to stem from the AI sector. The vast majority of that, $22.7 trillion, could come from AI for businesses.

No, it couldn't. Full Stop.

Disappointing to read that Musk is now dooming SpaceX to Tesla's fate, but when you have a psychopath in charge of anything, and they suffer from delusions of genius, this is what eventually happens.

His house of cards will never return what he invested. And it will all come crashing down. MAYBE someone can buy out SpaceX on the cheap and keep it going. But I doubt it.

And to explain, I figure Tesla's numbers are fully made up at this point, and all the manipulation of stock value (with SpaceX refinancing $20 Billion in debt at the last minute), this move will never get off the ground.

 

[Statistical]

Also, Starship is a steaming turd of a design, and I have rocket engineering kind in my circle of contacts to inform this opinion. Musk basically uses the name, bigness factor, and the crane capture gimmick to capture the public's attention, but in terms of practicality and achieving real milestones in space exploration, it's more gimmick than practical solution...

I dislike Musk but this is a dubious claim. If we are talking about terrestrial starship, not HLS or his mars colony bullshit, but the 2 stage fully reusable Earth to orbit launch vehicle I believe you will be proven very wrong.

Terrestrial Starship will be as much of a game changer as Falcon 9 was. Speaking of F9 with BO reusing a booster and other companies getting close SpaceX Falcon 9 moat is just about gone. This buys them another decade at least. I would be surprised if BO or anyone else has a starship competitor before 2040.

Of course any gains from Starship are just going to go into the AI hype money furnace.

 

[EllPeaTea]

Blue Origin may have been doing some fairing recovery work during the NG-3 launch.

The GO America finally arrived at Port Canaveral overnight after being clearly seen operating in the New Glenn landing zone.

Something mysterious is tarped on the deck - it's definitely not an intact fairing half, but it's something. Wait and see what the future holds for Blue!

 

[jlredford]

How much of a win are rotating detonation engines? X% more thrust per unit weight or some other metric?

 

[Ianal]

The Reuters article on SpaceX was interesting but I'm not really seeing how it all adds up. Granted, I'm not a finance guy, so there's probably a bunch of stuff that can be taken as read which I'm blissfully unaware of. But even so:

They're positing a TAM of $28 trillion, of which $26.5 trillion will come from the AI sector, so presumably the rest of their business activities - space launch, Starlink etc, are making up the remaining $1.5 trillion.

However, as per the Reuters article: "Those losses eclipsed the $4.4 billion in operating profit generated by Starlink, SpaceX’s satellite internet business and its largest revenue engine, which brought in $11.4 billion of its $18.7 billion total revenue last year."

Ignoring all the AI figures (which I personally think are utter horseshit), there's a big difference between $1.5 trillion and $18.7 billion, especially when SpaceX have already sewn up a large chunk of the space launch market. Unless I'm missing something really obvious, to me that implies that SpaceX think there's room to expand their non-AI services by a factor of roughly 5,000.

(Edit. Brainfart moment - as pointed out down thread, this should be about 100 not 5000. Still seems optimistic to me but makes my next comment a bit embarrassing!)

Did somebody lose a couple of orders of magnitude down the back of the couch cushions or something?

Was also slightly surprised to see the numbers for SpaceX / Starlink revenues. I mean, $18.7 billion isn't chump change by any means, but it's definitely less than I imagined, especially for Starlink.

 

[Statistical]

How much of a win are rotating detonation engines? X% more thrust per unit weight or some other metric?

For rockets (which I assume is what people here care about) you could increase chamber pressure by something in the ballpark of 5% to 20% for a given manifold pressure. So everything else the same (which is likely not true) you would get a modest boost in specific impulse and thrust. Nothing earth shattering but if the tech proves out we may see a lot of engines using it.

It isn't a new rocket engine cycle it is a "new" way to design the combustion chamber. If it works and makes sense from a mass/reliability standpoint it could in theory work with any rocket engine cycle. You could have anything from RDE pressure fed engines all the way up to a RDE boosted Raptor engine which gets 10% higher thrust and 20s higher specific impulse.

On edit: clarified engine cycle is rocket engine cycle.

 

[Statistical]

Was also slightly surprised to see the numbers for SpaceX / Starlink revenues. I mean, $18.7 billion isn't chump change by any means, but it's definitely less than I imagined, especially for Starlink.

Starlink has about 10M customers and $10B in revenue. You can only charge so much per customer especially going forward with competitors coming online over the next decade. Customer count will likely increase might even 10x but revenue per customer will decline. This is similar to the problem Tesla had you know back when Tesla made cares and didn't go off into musk wank territory with AI and robots.

So yeah I don't see $1.5T from Starlink. AFAIK the entire global internet access market is about $600B/yr annually. So likely that $1.5T is being padded with dubious datacenters in the sky nonsense. "If magically datacenters in the sky work and can compete with terrestrial ones and we get 10% of the global datacenter business ...."

Personally I am just praying the fanboys bid the stock up enough it doesn't tank before I can sell (6 months post-IPO).

 

[Jack56]

Worth noting that the 600th successful F9 booster landing was not the 600th booster recovery - two fell over after landing: one due to a fire that weakened a landing leg, the other in heavy seas.

 

[Jack56]

Vulcan issue proves costly to Northrop

I'll bet not as costly as to ULA.

 

[bryanlarsen]

Some context to quantify why that sounds bonkers: Total world GDP in 2025 was around $120 trillion, of which the services sector was $100 trillion. (The conventional three-sector breakdown of economic activities is primary: agriculture; secondary/industry: mining, manufacture, utilities, construction; tertiary/services: trade, transportation, ..., government.) I surmise SpaceX's premise is "AI-software will replace the output attributable to humans in all information, communication, financial, insurance, real estate, professional and administrative roles," or "AI-driven robots will replace all physical activities" or "AI will increase world GDP, even if they replace only a fraction of human labor" -- which is unlikely any time soon, but is consistent with Musk's recent futurist statements (Business Insider Nov 2025, Forbes.com Nov 2025, Fortune Dec 2025 (paywalled), AOL Apr 2026) and insistence that Optimus will soon be the major product of Tesla.

(Source: I work in international macroeconomics.)

Edit: Added supporting links, a third possible premise, phrasing.

Those aren't the only ridiculous premises.

Let's say the last premise is true, and AI doubles the world GDP. But there are still two more ridiculous premises that also must come true to justify that SpaceX valuation.

1. That an AI company captures the bulk of that GDP increase. There are rules of economics that mean that typically suppliers do not capture most of the benefit of a transaction. "Law of One Price", "consumer surplus", "price is the marginal cost", et cetera. AI is not a commodity so they're not iron clad rules, but it does appear that AI will be competitive. We've got at least 3 competitive frontier models at the moment and the Chinese are close behind.

2. That the AI company that captures this profit is SpaceX, not OpenAI, Anthropic, Google, Meta, Alibaba (qwen), DeepSeek, Kimi or Baidu (ERNIE) or somebody else. Heck it's also possible that NVidia or TSMC or somebody at a different level of the stack are the ones doing the capture.

 

[Wickwick]

You could increase chamber pressure by something in the ballpark of 20% compared to the inlet pressure. So modest boost in specific impulse. Nothing earth shattering but if the tech proves out we may see a lot of engines using it.

Constant-volume expansion heat addition leads to a more efficient thermodynamic cycle so the thermal efficiency of your engine can (in theory) go up a bit. However, the big win is not needing to haul around the compressor. So you save engine weight more than anything.

Again, this is all in theory. I have yet to see an RDE of any flavor that's flight-weight hardware.

This is the modern evolution of the pulse detonation engine (PDE) propulsion ideas from the 90's and early 2000's. Notably, there are no (non-research) planes flying with detonation tubes of which I'm aware despite the theoretical weight and efficiency gains.

Edit: fixed a brain fart.

 

[Wickwick]

The Reuters article on SpaceX was interesting but I'm not really seeing how it all adds up. Granted, I'm not a finance guy, so there's probably a bunch of stuff that can be taken as read which I'm blissfully unaware of. But even so:

They're positing a TAM of $28 trillion, of which $26.5 trillion will come from the AI sector, so presumably the rest of their business activities - space launch, Starlink etc, are making up the remaining $1.5 trillion.

However, as per the Reuters article: "Those losses eclipsed the $4.4 billion in operating profit generated by Starlink, SpaceX’s satellite internet business and its largest revenue engine, which brought in $11.4 billion of its $18.7 billion total revenue last year."

Ignoring all the AI figures (which I personally think are utter horseshit), there's a big difference between $1.5 trillion and $18.7 billion, especially when SpaceX have already sewn up a large chunk of the space launch market. Unless I'm missing something really obvious, to me that implies that SpaceX think there's room to expand their non-AI services by a factor of roughly 5,000. Did somebody lose a couple of orders of magnitude down the back of the couch cushions or something?

Was also slightly surprised to see the numbers for SpaceX / Starlink revenues. I mean, $18.7 billion isn't chump change by any means, but it's definitely less than I imagined, especially for Starlink.

The total addressable market for Starlink is everyone on earth not in China or Russia or other countries where they're legally prohibited from operating. One might also throw some population density limits and possibly local revenue limits into that number.

 

[YetAnotherBoris]

edit: Never mind, I'm dumb...

 

[Jack56]

RDEs like Astrobotic's are interesting but not for superior performance - The Wikipedia figure of 25% is true only for comparisons with low pressure open cycle engines. The advantage in performance over high pressure closed cycle engines is 5% or less. E.g. a methalox RDE would gain perhaps 10-15s Isp over a SpaceX Raptor.

The real advantage of RDEs is they have no need of high pressure turbomachinery and preburners, with big savings in weight, complexity and cost. Provided they can be made robust enough to withstand the rotating detonation fronts.

 

[YetAnotherBoris]

RDEs like Astrobotic's are interesting but not for superior performance - The Wikipedia figure of 25% is true only for comparisons with low pressure open cycle engines. The advantage in performance over high pressure closed cycle engines is 5% or less. E.g. a methalox RDE would gain perhaps 10-15s Isp over a SpaceX Raptor.

The real advantage of RDEs is they have no need of high pressure turbomachinery and preburners, with big savings in weight, complexity and cost. Provided they can be made robust enough to withstand the rotating detonation fronts.

I'm having trouble with the "no turbo machinery" part. For a given thrust, you still need to somehow get sufficient propellant mass into the combustion (err, I mean detonation) chamber, at a sufficient steady rate. How do you do that, without some kind of a pump?

As for pressure - how does one prevent the detonation byproducts from propagating back up the propellant feed lines, without pressurizing those feed lines so as to exceed the detonation back-pressure?

 

[Cyrano4747]

I can't help but just think that Musk is just smearing "AI" all over SpaceX in an attempt to boost the valuation at the IPO. He is mainly hype and this is his style. And it seems like any company mentions "AI" and the dumb market automatically boosts their value.

It's exactly this. There was a failing shoe company a few days ago that re-branded as an "AI" company and saw their stock spike 300%+ in a single day's trading. We're just at that point in the bubble when the stupid money gets REALLY stupid and REALLY exuberant and the people trying to cash in don't really need to even pretend to have a goal in mind, just a buzz word.

 

[RadarLuv]

“SpaceX expects more than 90 percent of that market, or $26.5 trillion, to stem from the AI sector. The vast majority of that, $22.7 trillion, could come from AI for businesses.”

To quote Elon himself:

“Frankly, I think we are more likely to discover unicorns dancing in the flame duct.”

 

[Argent Claim]

Interestingly, Reuters has also reported two days ago that SpaceX admitted in its filing that its orbital data center concept may not actually be all it's hyped up to be. To quote:

"Our initiatives to develop orbital AI compute and in-orbit, lunar, and interplanetary industrialization are in early stages, involve significant technical complexity and unproven technologies, and may not achieve commercial viability"

 

[EllPeaTea]

RDEs like Astrobotic's are interesting but not for superior performance - The Wikipedia figure of 25% is true only for comparisons with low pressure open cycle engines. The advantage in performance over high pressure closed cycle engines is 5% or less. E.g. a methalox RDE would gain perhaps 10-15s Isp over a SpaceX Raptor.

The real advantage of RDEs is they have no need of high pressure turbomachinery and preburners, with big savings in weight, complexity and cost. Provided they can be made robust enough to withstand the rotating detonation fronts.

I assume they still need pumps to feed the chamber with propellant? So is it just a case that the pumps can be smaller/lighter? How would they be powered?

 

[Statistical]

I assume they still need pumps to feed the chamber with propellant? So is it just a case that the pumps can be smaller/lighter? How would they be powered?

One option would be just pressure fed engine which has no pumps at all. Like the AJ10-190 used as the Orion Main Engine. Using RDE would allow you to either increase the chamber pressure (higher performance) or use lower pressure tanks (reduced mass). The other option would be you have an engine cycle that involves turbopumps and the RDE just acts as a booster to performance. Jack was pointing out the highest benefit would be the first scenario.

One caveat to my example of AJ1-190 is nobody has a demonstrated a liquid-liquid RDE. At least one has demonstrated liquid-gas engines but this one might just be gas-gas. So there would be challenges with getting the propellant to the engine in the right state with pressure fed designs. It might be possible to gasify one or both of the propellants using a heat exchange on the combustion chamber or for larger engines a low pressure preburner. A good illustration of how simple concepts become complex in rockets.

On edit: fixed incorrect liquid-gas claim

 

[Statistical]

I'm having trouble with the "no turbo machinery" part. For a given thrust, you still need to somehow get sufficient propellant mass into the combustion (err, I mean detonation) chamber, at a sufficient steady rate. How do you do that, without some kind of a pump?

Pressure fed rocket engines exist.

As for pressure - how does one prevent the detonation byproducts from propagating back up the propellant feed lines, without pressurizing those feed lines so as to exceed the detonation back-pressure?

The lines will have to be pressurized to some level. That is unavoidable. The "magic" of rotating part of rotating detonation engine is the pressure wave is not symmetrical. It allows boosting the chamber pressure to closer to the limit of the manifold pressure without the risk of catastrophic back pressuring the lines.

In case it is not clear in all rocket engines the chamber pressure is lower than the inlet/injector/manifold pressure. That gap is in the ballpark of 10% to 25% with higher pressure engines being on the lower end. That gap is wasted performance. By some means you increased the pressure of the propellants and paid for it with increased energy or mass. Then you didn't use all the potential of that increased pressure. RDE allow you to reduce but not eliminate that gap. Not sure this example but prior demonstrations have shown chamber pressures >97% of the manifold pressure. Closing that gap is where some of the increased performance is coming from in rocket engines. The rest is coming from more uniform combustion.

 

[2fifty6]

But where money does that come from?

A question for the ages.

 

[quamquam quid loquor]

I dislike Musk but this is a dubious claim. If we are talking about terrestrial starship, not hls, not mars colony bullshit but the 2 stage fully reusable Earth to Orbit launch vehicle you couldn't be more wrong.

It will be as much of a game changer as Falcon 9 was. I would point out with BO reusing a booster and other companies getting close SpaceX Falcon 9 moat is just about gone. This buys them another decade at least.

Of course it won't matter. Any gains from Starship are just going to go into the AI hype bullshit engine.

My understanding is Starship is primarily going to be used to improve Starlink/Starshield.

At some point in the future they'll launch AI datacenters in space, but they haven't even launched a prototype with Falcon, so it will be a few years before we know if that's economically feasible even at Starship internal launch costs.

 

[Foobarivy]

Some context to quantify why that sounds bonkers: Total world GDP in 2025 was around $120 trillion, of which the services sector was $100 trillion. (The conventional three-sector breakdown of economic activities is primary: agriculture; secondary/industry: mining, manufacture, utilities, construction; tertiary/services: trade, transportation, ..., government.)

Further to the point, the US GDP for 2026 is estimated to be $32 trillion. So SpaceX is saying that the addressable AI market is the size of 82% of the entire US economy. They're saying that their entire addressable market is 89% of the size of the US economy. Yeah, that's bigger than the entire services-producing and agricultural sectors.

Or you could say that it's bigger than the entire EU GDP, that's about $22 trillion. (Shakes head.)

 

[YetAnotherBoris]

A question for the ages.

Well, you see, once Starlink achieves universal direct-to-smartphone service, we will all gladly subscribe and pay $1000/month. Presto!

 

[tRexx]

If you want to overpay for a stock offering, where the actual financials in no way support the offering price, your chance is coming up.

 

[quamquam quid loquor]

Further to the point, the US GDP for 2026 is estimated to be $32 trillion. So SpaceX is saying that the addressable AI market is the size of 82% of the entire US economy. They're saying that their entire addressable market is 89% of the size of the US economy. Yeah, that's bigger than the entire services-producing and agricultural sectors.

Or you could say that it's bigger than the entire EU GDP, that's about $22 trillion. (Shakes head.)

The sneaky out that SpaceX has is that there's no timeline. I can claim the same TAM for toilet paper, but in 1,000 years. They really should be forced to claim TAMs for the filing year.

 

[quamquam quid loquor]

If you want to overpay for a stock offering, your chance is coming up.

Unfortunately for finance theory, this will almost certainly go up.

The amount of money equity trading shops will make on this churn is colossal. Look for Morgan Stanley, JPM, BofA, Citi, Wells, and all the other advisory shops stuff client accounts with the "future". The amount of equity derivatives being sold as people massively short and get squeezed will be unreal. Traders will make billions in the first week.

What this means is you shouldn't touch it, just like you shouldn't touch Tesla. Don't short it, don't buy it.

I will be buying it, but I know going in it's a gamble and disconnected from valuation.