Aerospace, defense & space

Aerospace, defense & space

Industry ID: 08 | Slug: aerospace-defense-space | Researched: 2026-09-15 | Analyst: agent

Reading note on the sector's characteristic error. Announced budget, appropriated funds, obligated contract, and delivered hardware are four different numbers in this sector, and they routinely differ by multiples and by years. Every quantitative claim below is labelled with which of the four it is. Where a widely circulated figure conflates them — Golden Dome's "$151bn contract", the counter-UAS "$29bn quarter", Korea's "$37bn exports" — the conflation is named explicitly rather than laundered.

1. Definition and boundaries

In scope. Commercial aerospace and airframers (Boeing, Airbus, Embraer and their supply chains); aero-propulsion and the engine aftermarket; defence primes; the venture-backed defence-tech challenger cohort; munitions, missiles, interceptors and the energetics/solid-rocket-motor industrial base; launch vehicles and launch services; satellite manufacturing, constellations and ground segment; satellite broadband, satellite connectivity and direct-to-device; military autonomy, drones and counter-drone systems; military space architecture.

Explicitly out of scope, and who owns it.

Boundary disputes worth naming.

1. Satellite connectivity is assigned here, not to telecom. This is a live and contestable call. Starlink's economics in 2026 are mobile-carrier economics — ARPU, churn, spectrum, carrier wholesale deals — not space economics. The assignment is defensible because the capital intensity, the launch dependency and the regulatory regime remain space-sector, but a reader should know that the direct-to-device trend (T-08-12) would sit equally well in telecom.
2. SpaceX is now three businesses and only one is this sector's. Its Q2 2026 disclosure shows Space at 12% of revenue, Connectivity at 55% and AI at 33%. An aerospace analyst covering SpaceX is covering a minority of the company. The AI segment is tracked here only where it bears on space assets (orbital compute) or on the group's capital allocation.
3. Shipbuilding and naval systems straddle this sector and industrials. Naval platforms are treated as out of scope; naval effectors and autonomous surface vessels (Saronic) are in scope because they are procured through the same munitions and autonomy channels.
4. Advanced air mobility (eVTOL) is in scope as an aerospace certification and supply-chain phenomenon, not as an urban-transport phenomenon.

2. Subcategories

1. Large commercial airframers. Boeing and Airbus, a duopoly with a combined ~16,100-aircraft backlog. Competition is on delivery slots and rate, not on order intake.
2. Aero-propulsion and engine aftermarket. CFM (GE/Safran), Pratt & Whitney, Rolls- Royce, MTU. Where commercial aerospace margin actually sits; increasingly the binding constraint on fleet growth rather than a follower of it.
3. Defence primes. Lockheed Martin, RTX, Northrop, General Dynamics, L3Harris, BAE, Thales, Leonardo. Distinguished by cost-plus/fixed-price mix, which now explains more of their margin variance than their end-market exposure.
4. Defence-tech challengers. Anduril, Shield AI, Saronic, Helsing, Castelion. Venture- funded, software-first, and as of June 2026 holding at least one serial-production aircraft designation.
5. Munitions and the energetics industrial base. 155mm, interceptors, solid rocket motors. Capacity-constrained rather than demand-constrained; the slowest-moving node.
6. Launch. SpaceX, Blue Origin, ULA, Arianespace, Rocket Lab. Distinguished in 2026 by the fact that diversification has gone backwards, not forwards.
7. Satellite manufacturing and constellations. Large GEO (contracting), proliferated LEO (expanding), and government architectures (consolidating onto commercial).
8. Satellite broadband and connectivity. Starlink, Amazon Leo, Eutelsat OneWeb, Viasat, SES. A volume business with falling ARPU.
9. Direct-to-device satellite. Starlink D2C, AST SpaceMobile, Amazon's proposed 5,105-satellite network. Carrier-channel rather than consumer-channel.
10. Military autonomy and drones. From FPV attritables at consumer-electronics cost to Collaborative Combat Aircraft. Unit economics span five orders of magnitude.
11. Counter-UAS. The demand-side mirror of (10), and the clearest case of defence-tech product-market fit.
12. Military space services. Starshield, MILNET, the Space Data Network — a government-owned, contractor-operated model displacing government-built systems.

3. Market structure

Concentration. The sector is a set of adjacent oligopolies with one emerging winner-take-most segment.

• Commercial airframes: a strict duopoly. Airbus led Boeing 1,157 to 498 on gross orders through August 2026 — a two-to-one order split that is unusual even for this duopoly and reflects Boeing's certification and rate constraints rather than product inferiority.
• Launch: trending toward winner-take-most, and in 2026 it went backwards on diversification. SpaceX flew 78 times in H1 2026. Blue Origin's New Glenn destroyed its own pad in May.
• Defence primes: a regulated oligopoly with five US and roughly six European players, where the customer sets prices through negotiated rates and the government owns the demand curve.
• Munitions and energetics: an effective duopoly at the solid-rocket-motor node (L3Harris/Aerojet and Northrop), which is why a single subcontract — L3Harris's $4.7bn award in September 2026 — is systemically important.
• Satellite broadband: one dominant operator (12.0m subscribers) and a well-capitalised challenger with 396 satellites on orbit.

Where margin actually sits. Not where revenue sits.

The pattern is consistent and counter-intuitive: building the hardware loses money; servicing it and operating it makes money. Launch is the loss leader for connectivity. New-aircraft manufacture is the loss leader for the aftermarket. Fixed-price defence development destroys margin while production munitions and sustainment create it.

Barriers to entry. Extreme and rising: certification (no Western eVTOL type certificate exists after a decade and billions spent); qualification cycles (30-month solid-rocket-motor lead times); security clearance and facility accreditation; capital intensity (SpaceX spent $28.5bn of capex in six months); and, newly, regulatory content rules — the EU's SAFE instrument caps non-EU/EEA/EFTA/Ukraine content at 35%, which is a barrier defined by nationality rather than capability.

Who has pricing power. Engine OEMs over airlines, through parts and repair-approval control — which is precisely what IATA is contesting. Solid-rocket-motor suppliers over missile primes. SpaceX over everyone in launch, and now over the US government in military communications. Nobody has pricing power over the US government in fixed-price development, which is why that contract type keeps producing charges.

Market size — with modeller named, treated as estimate. A widely syndicated figure puts the global space economy at $429bn, reported by SatNews on 2026-05-14 and attributed to the commercial-satellite-dominated annual industry survey. Competing secondary figures of $470bn and $626bn circulate on aggregator sites with no stated methodology. These are estimate, not fact, they do not reconcile with each other, and none was obtainable in primary form within this research budget. They should not be used as a denominator. The defensible market-size anchors in this dossier are company-reported: Lockheed's $230bn backlog, Boeing's $715bn backlog, Rheinmetall's €80.5bn backlog, and SpaceX's $47.5bn backlog — all fact, all dated, all from filings.

4. Who matters

Leading companies. SpaceX (https://www.spacex.com) · Boeing (https://www.boeing.com) · Airbus (https://www.airbus.com) · Lockheed Martin (https://www.lockheedmartin.com) · RTX (https://www.rtx.com) · Northrop Grumman (https://www.northropgrumman.com) · General Dynamics (https://www.gd.com) · L3Harris (https://www.l3harris.com) · Rheinmetall (https://www.rheinmetall.com) · BAE Systems (https://www.baesystems.com) · Thales (https://www.thalesgroup.com) · Leonardo (https://www.leonardo.com) · GE Aerospace (https://www.geaerospace.com) · Safran (https://www.safran-group.com) · General Atomics ASI (https://www.ga-asi.com) · Blue Origin (https://www.blueorigin.com) · Amazon Leo (https://www.aboutamazon.com/news/amazon-leo) · SES (https://www.ses.com) · AST SpaceMobile (https://ast-science.com) · Hanwha Aerospace (https://www.hanwhaaerospace.com) · Korea Aerospace Industries (https://www.koreaaero.com)

Notable startups. Anduril (https://www.anduril.com) · Shield AI (https://shield.ai) · Saronic (https://www.saronic.com) · Helsing (https://helsing.ai) · Quantum Systems (https://quantum-systems.com) · Starcloud (https://www.starcloud.com) · Castelion · Mach Industries · plus the named Ukrainian cohort in the EU joint-venture programme (Tencore, Deviro, Athlon Avia, UForce, Vyriy Industry).

Active investors. Andreessen Horowitz American Dynamism (https://a16z.com/american-dynamism/) · Founders Fund · General Catalyst · 8VC · Benchmark and EQT (into orbital compute) · Sequoia. Sector venture total was $14.6bn across 107 rounds in H1 2026, against $9.6bn for all of 2025 and $1.6bn in 2020.

Platforms and standards bodies. 3GPP (non-terrestrial-network releases underpinning direct-to-device) · ITU (spectrum coordination and orbital slots) · NATO Standardization Office (STANAGs) · RTCA and EUROCAE (avionics standards) · the USAF Autonomy Government Reference Architecture, which is a deliberate anti-lock-in standard for CCA autonomy.

Regulators. FAA (https://www.faa.gov) — certification, production oversight, launch licensing and mishap investigation · FCC Space Bureau (https://www.fcc.gov/space) — constellation authority and spectrum · EASA · DDTC/BIS for ITAR and EAR export control · European Commission DG DEFIS (https://defence-industry-space.ec.europa.eu) · CAAC.

Government agencies as buyers. US DoD · Missile Defense Agency (https://www.mda.mil) · Space Development Agency (https://www.sda.mil) · US Space Force · NRO · NASA · ESA (https://www.esa.int) · Korea's DAPA · Japan's ATLA · NATO NSPA.

Research institutions. DoD Office of Inspector General (https://www.dodig.mil) — the most reliable public check on funded-versus-delivered · GAO (https://www.gao.gov) · CRS (https://crsreports.congress.gov) · Kiel Institute (https://www.kielinstitut.de) · FPRI (https://www.fpri.org) · SIPRI (https://www.sipri.org) · RAND · IISS · CSIS.

Trade organisations. IATA (https://www.iata.org) · Aerospace Industries Association (https://www.aia-aerospace.org) · ASD Europe (https://www.asd-europe.org) · Satellite Industry Association (https://sia.org) · AUVSI (https://www.auvsi.org).

Consumer and civil-society groups. Secure World Foundation (https://swfound.org) on space sustainability · Union of Concerned Scientists satellite database · Article 36 and the Campaign to Stop Killer Robots on autonomous-weapons norms · the American Astronomical Society on constellation light pollution and radio interference.

5. Products, business models, technologies, customers

Major products. Narrowbody and widebody airliners; turbofan engines and their overhaul; fighter aircraft and Collaborative Combat Aircraft; interceptors and guided munitions; artillery ammunition and propellant; launch services; broadband constellations and user terminals; counter-UAS effectors and sensors; military satellite communications.

How money is actually made today, and how that is changing.

• Aftermarket over new-build. The installed base, not the production line, is the profit engine. IATA's June 2026 study quantifies why this is intensifying: LEAP shop visits rise from 600–800/year in 2025 to 5,000+ by 2040. Whoever controls parts and repair approvals captures that.
• Recurring connectivity revenue over hardware sales. SpaceX's Connectivity segment produced $1.656bn of operating income in a single quarter on 12.0m subscribers. This is a subscription business that happens to require rockets.
• Ceiling-based enterprise agreements over programme-by-programme contracts. Palantir ($10bn ceiling, 2025), Anduril ($20bn ceiling, consolidating 120–130 orders), MDA SHIELD ($151bn ceiling, 2,100+ awardees, zero obligated at base). The economics shift from winning a programme to winning task orders, and bid-and-proposal cost moves onto the supplier with no revenue guarantee.
• Capacity-ahead-of-contract as a strategy. Lockheed's CEO explicitly attributed Q2 2026 performance to "expanding munitions capacity before contracts materialized" — a bet that only works when demand is politically underwritten.
• Consumables over platforms. Ukraine's stated drone production of 4m+ (2025) rising to a 7m+ target (2026) makes the relevant unit a consumable at consumer-electronics cost, not a platform at aerospace cost.

Technologies that matter. Solid-rocket-motor and energetics manufacturing (the true bottleneck); phased-array antennas and optical inter-satellite links; reusable launch; autonomy stacks and the government reference architectures that keep them vendor-swappable; 3GPP NTN waveforms for direct-to-device; additive manufacturing for rate-limited castings and forgings; high-power microwave and gun-based counter-UAS effectors.

Major customer segments and what they buy on.

6. Geography

Production. Airframes concentrate in Washington State, South Carolina, Toulouse, Hamburg and Mobile. Engines in Ohio, Connecticut, Paris and Derby. Munitions and energetics in a narrow set of US plants — Camden, Arkansas (L3Harris solid rocket motors, two new facilities due 2027); Mesquite, Texas and the Iowa Army Ammunition Plant for 155mm — plus Germany, Spain and a new Rheinmetall plant in Lithuania. Launch is overwhelmingly Florida and Texas. Satellite manufacturing is splitting between US LEO volume production and a contracting European GEO base.

Capital. Venture capital is overwhelmingly US — the $14.6bn H1 2026 defence-tech total is US-dominated, consistent with the macro brief's finding that the Americas took 66% of Q2 2026 global VC and Europe $25.6bn is the structurally weak leg. Sovereign capital is the opposite: the EU's €150bn SAFE facility and national European budgets are the largest incremental capital pools in the sector, and they are explicitly designed to be spent inside Europe.

Demand. Defence demand is concentrated in NATO Europe (nearly 20% growth in core expenditure in 2025 versus 2024), the US, the Gulf and the Indo-Pacific. Commercial aviation demand is global with Asian growth leading. Satellite broadband demand is most intense where terrestrial infrastructure is weakest.

Regulation. The US regulates access to technology (ITAR/EAR) and to its market; the EU now regulates access to its money (SAFE content rules); China regulates components (drone subsystems). These are three different levers and they are being pulled in different directions.

Non-US market covered with a regional source — South Korea. Seoul Economic Daily (Korean-origin, 2026-09-02) documents the breadth of the Korean export push: KAI in final-stage KF-21 talks with Indonesia after 140+ FA-50 exports to six countries; LIG's Cheongung-II sold to the UAE, Saudi Arabia and Iraq with customers requesting early delivery and follow-on orders; Hanwha's K9 going to Spain via Indra Sistemas — a localisation structure that is readable as a direct response to European content rules; HD Hyundai negotiating a Peruvian submarine design contract; Hyundai Rotem's K2 approved for Peruvian evaluation. Korea's competitive advantage is spare qualified capacity and delivery date, which is exactly what Western lines lack. Japan's relaxation of its arms export restrictions adds a second Asian competitor in the same window.

Note the source-quality trap here: secondary aggregator sites circulate a "$37bn 2026 Korean defence exports" figure with no named methodology. The only figure traceable to official reporting is $15.4bn for 2025. The $37bn figure is not used in this dossier.

7. Historical trend patterns

Ten to twenty-five years of cycles in this specific sector, and what the false positives looked like.

The post-Cold-War consolidation and its unwinding (1993–present). The 1993 "Last Supper" drove US primes from dozens to five. The 2020s are the first genuine reversal: Anduril holding a production aircraft designation in 2026 is the first time in thirty years a new entrant has reached serial production of a combat aircraft. The lesson from the consolidation era is that the barrier was never purely technical — it was programme-of-record access, which is exactly what enterprise agreements and OTAs now route around.

The commercial-space hype wave of 2007–2015, and what actually survived. Dozens of smallsat-launch and Earth-observation ventures were funded on a "launch costs will collapse" thesis. Most failed. The survivors — SpaceX and Rocket Lab — won on manufacturing execution, not on business-model novelty. False positive to remember: around 2014–2016 the consensus was that a competitive multi-provider heavy-lift market would exist by the early 2020s. In September 2026, after the New Glenn pad explosion, it does not.

The 2020–2022 space SPAC wave. Dozens of pre-revenue space companies went public via SPAC on 2025–2030 revenue projections. Most missed by an order of magnitude and several delisted or were acquired at a fraction of their listing value. This is the single most relevant precedent for 2026's defence-tech and orbital-compute valuations, and it resolved in roughly three years.

Airframer rate promises (2018–present). Boeing's 737 rate has been promised at 57/month, then 52, then 47, then 42, repeatedly. Airbus has cut delivery guidance mid-year more than once. The pattern is now so consistent that the base case for any announced rate should be a miss, not a hit. As of August 2026, Airbus at 475 YTD against 870 is on track to continue the pattern.

Munitions replenishment (2022–present). Every Western government announced large ammunition capacity expansions after February 2022. Four years on, the US 155mm line is at roughly 36% of its stated target. The lesson is that the physical constraint — qualification cycles, energetics, machine tools, skilled labour — is measured in years regardless of how fast money moves.

Advanced air mobility (2016–present). Promised entry into service in 2020, then 2023, then 2024, then 2025, then 2026. No Western type certificate exists as of August 2026. This is now a decade-long false positive, and the failure mode migrated from technology to certification, which is slower to fix.

Satellite broadband's earlier round (1990s). Iridium and Globalstar both went bankrupt after building constellations against overestimated demand and underestimated terrestrial competition. Starlink's 12m subscribers vindicate the architecture — but its ARPU falling from $85 to $66 is the same demand-elasticity signal that killed the first generation, arriving at a very different scale.

What this history implies. In this sector, capability claims resolve slowly and capacity claims resolve even more slowly. The trends that have persisted are the ones backed by physical throughput (SpaceX's launch cadence, Rheinmetall's backlog conversion); the ones that have failed are the ones backed by projected demand curves (SPAC-era space, AAM) or by announced capacity (155mm).

8. What is changing now (as of 2026-09-15)

Grounded in the macro brief, the five changes that matter:

1. The sector's largest private company became public, and the disclosure reframed it. SpaceX's IPO on 2026-06-12 raised the largest sum in history; its first reported quarter (2026-08-04) revealed that launch is 12% of revenue and the only loss-making core segment. The stock closed at $125.33 on report day — below the IPO price and 44% off its June peak, with over $1tn of market value erased from that peak. The macro brief's warning that the FOMC flagged "high AI-firm equity valuations and increased leveraged financing of infrastructure buildout" as a financial-stability risk has a concrete instance here: SpaceX issued $25bn of senior notes at a 5.855% weighted-average rate in June and spent $15.8bn of a $18.4bn quarterly capex bill on AI.

2. European appropriations are converting to orders — unevenly. Rheinmetall's €80.5bn backlog and book-to-bill above 3 are hard evidence of conversion. NATO reports nearly 20% core-expenditure growth in 2025. But the Kiel Institute's critique stands: conversion is concentrated in a few well-positioned suppliers and product lines, and the F126 frigate cancellation shows growth is not monotonic.

3. The gap between appropriated and delivered is now officially documented. The DoD IG's July 2026 finding — 36,000 155mm rounds/month against a 100,000 target, after $4.5bn spent and one plant producing zero conforming shells over three and a half years — is the most important single document in this sector this year. FPRI's parallel finding on interceptors (132:1 consumption-to-production during the Iran conflict; 24–30 month lead times; April 2026 funding delivering mid-2028) shows the same structure.

4. The Middle East conflict the FOMC flagged has a direct sector consequence. It drained interceptor stockpiles at rates that make the replenishment question acute, and it is the proximate driver of the PAC-3 MSE ramp from ~600 to a 2,000/year target — a target that requires two new Camden, Arkansas facilities that do not open until 2027.

5. Launch diversification reversed. New Glenn's static-fire explosion on 2026-05-28 destroyed Launch Complex 36 with repairs expected to take over a year, after an April 2026 second-stage anomaly. Every business case that assumed multi-provider heavy-lift by 2027 needs re-basing — including Amazon Leo's, which has deployed 396 satellites without a single New Glenn flight.

Appropriations overhang. As of this research date no FY2027 appropriations bills had been enacted and Senate appropriators remained divided on the defence topline. Gen. Guetlein warned that without continued funding after 30 September, Golden Dome "will stop with everything we've already built and delivered." Reporting indicates a continuing resolution was moving, but the enacted status as of 2026-09-15 could not be confirmed to a primary source and is recorded as a data gap. The September 16 FOMC decision falls one day after this research date and is unresolved.

9. The five lists

Five most important current trends

1. Munitions and interceptor output far below stated targets despite full funding (T-08-02)
2. SpaceX's first public disclosure reframing what the sector's largest company is (T-08-04)
3. European rearmament converting into firm multi-year backlog (T-08-01)
4. Record prime backlogs alongside thin or negative programme margins (T-08-03)
5. Announced aircraft production rates outrunning actual deliveries (T-08-07)

Five fastest-growing signals

1. Defence-tech challengers crossing into serial-production programmes (T-08-09)
2. Starlink subscriber growth with ARPU compression (T-08-05) — overlaps list 1 via T-08-04
3. Drone mass production as a sovereign industrial capability (T-08-14)
4. Counter-UAS moving into funded procurement lines (T-08-06)
5. SpaceX's Starshield displacing government-built military space architecture (T-08-11)

Five trends most likely to affect businesses

1. Engine MRO capacity as the binding constraint on fleet growth (T-08-08)
2. European-content rules turning EU defence money into industrial policy (T-08-13)
3. Ceiling-value enterprise agreements replacing programme contracting (T-08-10)
4. Heavy-lift launch diversification reversing (T-08-18)
5. Record backlogs with thin margins (T-08-03) — overlaps list 1

Five trends most likely to affect consumers

1. Starlink at 12m subscribers with ARPU falling to $66 (T-08-05) — overlaps list 2
2. Direct-to-device satellite becoming a carrier product (T-08-12)
3. Aircraft delivery and engine constraints feeding through to airline capacity and fares (T-08-07, T-08-08)
4. eVTOL air taxis continuing not to arrive (T-08-17)
5. Drone proliferation and counter-drone authority over civil airspace (T-08-06, T-08-14)

Where these overlap, explicitly. T-08-04 and T-08-05 are two faces of the same company and appear across three lists. T-08-07 and T-08-08 are a single commercial- aerospace constraint viewed from the airframe and the engine side. T-08-06 and T-08-14 are the defence and offence sides of one arms race.

10. Overhyped / overlooked / cooling / reversing

Most overhyped

Golden Dome as an immediate contract bonanza. The evidence that hype outruns substance is arithmetic. The $151bn figure in circulation is an IDIQ ceiling shared by more than 2,100 awardees, with no funding obligated at the base contract level — MDA said so explicitly. The programme office's actual FY2026 money was $22.5bn, it did not arrive until March 2026, and it was 90% obligated by August. That is a roughly sevenfold gap between the circulating number and the money in hand, against an FY2027 appropriation that was not enacted as of this research date. The programme is real, funded and testing; the number attached to it in public discourse is not.

Orbital data centres as near-term AI infrastructure. Starcloud is valued at $1.1bn having flown one 130-pound satellite carrying a single Nvidia H100. Its own CEO puts favourable economics three to five years out and mainstream adoption "about a decade out". Microsoft's president said the company is "keeping our feet on the ground". The thesis is nonetheless being used to justify SpaceX AI capex of $15.8bn in one quarter against a $1.257bn segment operating loss — and the market has already begun repricing it, with SPCX below its IPO price two months after listing.

Honourable mention: "defence tech is replacing the primes." $14.6bn of H1 2026 venture funding is real. So is Lockheed's $230bn backlog and Boeing's $715bn. The challengers have won a genuine and important foothold (T-08-09); they have not won the revenue base, and the funding-to-programme ratio is the thing to watch.

Most overlooked

The solid-rocket-motor and energetics chokepoint. Almost all attention goes to interceptor programmes; almost none to the fact that a 30-month motor lead time and a two-supplier market set the ceiling on every one of them. L3Harris's $4.7bn September 2026 subcontract, with two Camden facilities not operational until 2027, is arguably the most consequential award of the year and got a fraction of the coverage of the Golden Dome ceiling number. Attention has missed it because it is a subcontract, not a programme, and because it is boring chemistry rather than exciting autonomy.

Engine MRO as a capacity constraint on the entire aviation system. IATA is making this argument loudly and it is being read as a trade dispute rather than as a structural constraint. GTF groundings peaked at 648 aircraft — 28% of the fleet. Shop visits grow 6–8x by 2040. This determines airline capacity and fares more than airframer delivery rates do, and it is where aerospace profit is migrating.

Starlink's ARPU compression. The 12m-subscriber headline travelled everywhere; the $85 → $66 ARPU decline travelled almost nowhere. It is the single most important economic datum in satellite broadband because it sets the price every competing constellation must clear, and it was invisible before June 2026 because the company was private.

SAFE's 35% content rule as a trade weapon. Treated as procurement minutiae; it is industrial policy that may determine whether European budget growth becomes US prime revenue. Korea has already found the workaround (localise via Indra); US primes mostly have not.

Trends that appear to be cooling — with the indicator that turned

• GEO satellite expansion. Indicator: SES cancelling IS-41 and IS-44 outright for failing its internal rate-of-return threshold, joining Eutelsat. Cancellation, not deferral, is the turn. Within SES, Networks grew over 100% while Media/Video fell 11%.
• eVTOL / advanced air mobility. Indicator: EHang — the only holder of a type certificate — withdrawing its 2026 revenue guidance. When the certified player cannot forecast revenue, the uncertified players' timelines carry no information.
• Heavy-lift launch diversification. Indicator: Launch Complex 36 destroyed with a repair estimate exceeding a year, following an April second-stage anomaly. Three New Glenn flights in total, none carrying Amazon Leo satellites.

Trends that may reverse — and the mechanism

• European rearmament conversion could stall, not through political reversal but through fiscal reversal: SAFE is a loan instrument, and in the macro brief's sticky-inflation, hawkish-lean environment, debt-financed defence expansion in high-debt member states is the first thing to be trimmed. The mechanism would show up as slipping contract signatures against approved plans, not as budget announcements being withdrawn.
• SpaceX's military-space consolidation could reverse legislatively. Congress already reinstated $500m for SDA Tranche 3 in FY2026 over vendor-dependence concerns. A single appropriations cycle could restore a multi-vendor architecture.
• The defence-tech funding cycle could reverse on a single failed programme. The cohort's valuations assume programme capture. A high-profile CCA cost overrun or schedule failure under a fixed-price structure — the exact mechanism that produced Boeing's KC-46 and VC-25B charges — would reprice the category quickly.
• Counter-UAS effectiveness could invert the drone trend. If directed-energy and gun-based effectors achieve favourable cost-per-kill at scale, the economics that make mass attritable drones dominant flip. The indicator would be effector cost-per-engagement falling below drone unit cost in a documented operational context.

11. Risks and major uncertainties

Sector-specific risks.

1. Appropriations discontinuity. No FY2027 bills enacted as of this research date; a named programme manager publicly warning of a 30 September stop. US defence revenue recognition is hostage to a legislative calendar.
2. Fixed-price development risk transfer. It has destroyed margin at Boeing (VC-25B $280m charge in Q2 2026 alone) and is now being applied to first-time manufacturers with thinner balance sheets.
3. Single-node industrial dependencies. Solid rocket motors, PAC-3 seekers (a single Boeing facility in Huntsville), energetics precursors, and large castings and forgings. Any one of these caps output regardless of funding.
4. Vendor concentration in military space. Over 480 MILNET satellites on one supplier's terminals, with Tranche 3 defunded.
5. Valuation risk in the challenger cohort. $14.6bn deployed in six months into companies whose programme revenue is largely undisclosed, in a rising-rate environment the macro brief explicitly flags.
6. Demand reversal. A Ukraine settlement or a Middle East de-escalation would leave newly built munitions capacity searching for demand, with the capacity arriving (2027–2028) after the demand peak.
7. Concentration risk in commercial aerospace. Two airframers, three engine makers. A quality escape at any one of them is a system-wide event.
8. Orbital debris and spectrum conflict as constellation counts rise past 10,000 active satellites for a single operator.

Genuine unknowns — and the distinction.

Things we don't know but could find out: actual FY2027 appropriations outcome; Anduril's true post-money valuation; CCA contract values and quantities (not disclosed); direct-to-device paying subscribers (not separately reported by anyone); actual SAFE disbursements as opposed to approvals; Ukraine's audited drone output; classified MILNET cost and performance; Korea's official 2026 export authorisations.

Things nobody can know: whether Starship reaches a cadence that makes orbital compute economic; whether counter-UAS or drones win the cost-exchange race; whether European political commitment to 3.5%/1.5% survives two election cycles and a recession; whether the defence-tech cohort's software advantage persists once primes are forced to match it; whether a Ukraine settlement occurs and on what timeline.

12. Scenarios to 2030

Base case — grinding conversion. European budgets keep converting at roughly the current rate; US munitions output climbs but stays below stated targets into 2028; Boeing and Airbus each settle at ~10–15% below announced rates; Starlink passes 25m subscribers at declining ARPU; Anduril and General Atomics deliver a few hundred CCAs; launch stays SpaceX-dominated with Vulcan and New Glenn as marginal second sources. Falsifiable early indicator: Rheinmetall book-to-bill staying above 1.5 through FY2027 while US 155mm output passes 60,000/month by end-2027.

Upside — the industrial base actually scales. The Camden facilities open on time in 2027, PAC-3 MSE output passes 1,200/year by 2029, 155mm reaches 100,000/month, SAFE disbursements convert at above 70% into signed contracts, and Airbus and Boeing hit 900 and 700 deliveries respectively by 2029. Indicator: the first DoD IG or GAO report that finds a munitions programme meeting its stated rate — something that has not happened in this cycle.

Downside — appropriations and demand both wobble. FY2027 lands as a year-long CR, Golden Dome slows, a Ukraine settlement removes the European political forcing function, and newly built capacity opens into falling orders. Defence-tech valuations reset hard as the cohort's revenue is revealed to be a fraction of its funding. Indicator: a full-year continuing resolution enacted, plus the first down round or flat round at a top-five defence-tech company.

Disruption — mass attritables displace exquisite platforms. Drone and cruise-missile mass production at consumer-electronics cost makes multi-million-dollar platforms economically indefensible for a widening set of missions; procurement restructures around consumables. Indicator: a Western military cancelling or truncating a major manned platform programme explicitly in favour of attritable mass — as distinct from adding attritables alongside it.

Regulatory — content rules and export control fragment the market. SAFE's 35% cap is enforced strictly, the US responds with its own reciprocity conditions, and the transatlantic defence market splits into two supply chains with Korea and Türkiye arbitraging between them. Indicator: a major European competition where a US prime is excluded on content grounds rather than on capability or price.

Failure — a systemic quality or safety event. A fatal commercial aviation accident traced to rate-pressure quality escape, or a launch failure that destroys a crewed or strategically critical payload, resets certification and oversight across the sector. Indicator: an FAA or EASA production-rate cap reimposed on a major programme, or an NTSB/FAA finding attributing a hull loss to manufacturing conformity.

13. Data gaps and limitations

Could not verify (highest priority).

1. FY2027 appropriations status as of 2026-09-15. A law-firm appropriations blog indicated a continuing resolution funding the government to 11 December had passed the House 370–48 and the Senate 90–6 and been signed, but the page mixed a July publication date with September updates, and no primary source (congress.gov, whitehouse.gov) could be retrieved to confirm enactment. congress.gov blocks automated retrieval via robots.txt, which is a structural access problem for this sector and a Phase 2 licensing/ingest decision. Given Golden Dome's stated 30 September cliff, this is the single most consequential unresolved fact in the dossier.
2. Anduril's post-money valuation. Crunchbase News reports the May 2026 Series H at $30.5bn; multiple secondary trackers report $61bn. Both cannot be post-money on a $5bn round. No primary filing was located. Recorded as unverified.
3. CCA contract values and quantities. The June 2026 production awards to General Atomics and Anduril were announced without disclosed values or unit counts. Unit cost — the entire premise of the CCA concept — is therefore not publicly checkable.
4. Actual SAFE disbursements. The Council publishes approvals for 18 member states but not amounts disbursed or contracts signed. This is precisely the appropriation-to-contract conversion datum the sector most needs and it is not published.
5. Direct-to-device commercial traction. No operator — SpaceX, AST, or any carrier — discloses D2D subscribers or revenue separately. The category's adoption is unmeasurable from public sources.

Conflicting figures recorded rather than resolved.

Missing for non-US markets. No primary Chinese-language source on Guowang/Qianfan constellation progress or Chinese military aerospace production. No Japanese-language primary source on ATLA export policy — the Japan export-liberalisation claim rests on a Korean secondary source and should be verified against ATLA before use. No Indian or Turkish primary sources despite both being materially significant exporters. Russian production figures are entirely absent and are only inferable from Ukrainian claims, which are an interested party.

Paywalled or access-restricted. Aviation Week (AWIN subscription) — the Ukraine drone figures came from an excerpt only. SpaceNews and congress.gov restrict automated retrieval. CNBC and Stars-and-Stripes-adjacent outlets returned 403s on direct fetch. IATA's underlying economics datasets are licensed separately from the free press releases. Janes, Teal Group and Forecast International's full datasets are enterprise- priced and are where the reliable production-forecast data actually lives — Phase 2 should expect this to be the sector's main licensing cost.

Methodological limitation. Several figures central to this dossier are self-reported by parties with an interest: Ukraine's drone production (a government at war), Golden Dome's intercept test result (the programme office), CCA schedule performance (the service), and every company's own backlog definition. Rheinmetall's "nomination backlog" in particular includes framework volumes that are not firm orders and is not comparable like-for-like with Lockheed's backlog. These are labelled in the trend records but the reader should not treat them as equivalent to audited data.

14. Ranking scorecard

15. Sources

1. SpaceX Reports Second Quarter 2026 Results — SpaceX Investor Relations — https://s21.q4cdn.com/184289198/files/doc_financials/2026/q2/SpaceX-Reports-Second-Quarter-2026-Results.pdf — 2026-08-04 — A
2. SpaceX's revenue rises as its once-soaring stock price drifts back to Earth — NPR — https://www.npr.org/2026/08/04/nx-s1-5918536/spacex-first-earnings-report-since-ipo — 2026-08-04 — B
3. SpaceX's Record-Breaking IPO (SPCX) — Crunchbase News — https://news.crunchbase.com/public/spacex-record-breaking-ipo-spcx/ — 2026-06-12 — B
4. Defence Investment Update: Record Spending in Europe and Canada — NATO — https://www.nato.int/en/news-and-events/articles/news/2026/07/07/defence-investment-update-record-spending-in-europe-and-canada — 2026-07-07 — A
5. Rheinmetall's H1: Strong sales growth and record profitability (Half-Yearly Financial Report H1 2026) — Rheinmetall AG — https://www.rheinmetall.com/Rheinmetall%20Group/Presse/News/Documents/2026/08/2026-08-06-Rheinmetall-News-Half-Yearly-Financial-Report-H1.pdf — 2026-08-06 — A
6. Europe is spending more on defence than ever before: Time to spend smart — Kiel Institute for the World Economy (Rodrigo Carril) — https://www.kielinstitut.de/publications/news/europe-is-spending-more-on-defence-than-ever-before-time-to-spend-smart/ — 2026-02-13 — A
7. What is Security Action for Europe (SAFE)? — Council of the European Union — https://www.consilium.europa.eu/en/policies/safe/ — 2026-04-10 — A
8. Bang for Army's billions of bucks not measuring up in 155 mm shell production, IG says — Stars and Stripes, reporting the DoD Inspector General — https://www.stripes.com/branches/army/2026-07-14/army-artillery-shell-shortfall-22262070.html — 2026-07-14 — B (underlying IG report is Tier A but was not retrieved directly)
9. Scaling Patriot Production: The Industrial Base Crisis Explained — Foreign Policy Research Institute (Amoah, Bazilian, Matisek) — https://www.fpri.org/article/2026/05/scaling-patriot-production-the-industrial-base-crisis-explained/ — 2026-05-18 — B
10. Lockheed Martin to Bring PAC-3 MSE Interceptor's Production to 2,000 per Year — The Aviationist — https://theaviationist.com/2026/01/06/lockheed-martin-pac-3-mse-production-2000-year/ — 2026-01-06 — B
11. L3Harris Wins $4.7 Billion PAC-3 MSE Contract to Raise US Patriot Interceptor Production Capacity — Army Recognition — https://www.armyrecognition.com/news/army-news/2026/l3harris-wins-4-7-billion-pac-3-mse-contract-to-raise-us-patriot-interceptor-production-capacity — 2026-09-09 — C
12. Lockheed Martin Lifts 2026 Guidance on Record $230B Backlog — GovCon Wire, reporting Lockheed Martin Q2 2026 results — https://www.govconwire.com/articles/lockheed-martin-q2-2026-earnings-guidance — 2026-07-24 — B
13. Boeing Reports Second Quarter Results — The Boeing Company Investor Relations — https://boeing.mediaroom.com/2026-07-28-Boeing-Reports-Second-Quarter-Results — 2026-07-28 — A
14. Airbus And Boeing Report August 2026 Commercial Aircraft Orders and Deliveries — Forecast International (Flight Plan) — https://flightplan.forecastinternational.com/2026/09/09/airbus-and-boeing-report-august-2026-commercial-aircraft-orders-and-deliveries/ — 2026-09-09 — B
15. Urgent Action Needed to Ease Engine MRO Bottlenecks — IATA — https://www.iata.org/en/pressroom/2026-releases/06-24-urgent-action-needed-to-ease-engine-mro-bottlenecks/ — 2026-06-24 — A
16. Golden Dome funding is 90% obligated, Guetlein says — Aerospace America (AIAA) — https://aerospaceamerica.aiaa.org/golden-dome-funding-is-90-obligated-guetlein-says/ — 2026-08-20 — B
17. MDA Expands SHIELD Pool With 1,086 Additional Awards for Golden Dome — GovCon Wire — https://www.govconwire.com/articles/mda-1086-second-tranche-shield-awards — 2025-12-19 — B
18. Golden Dome Project Begins With a Massive $151 Billion Contract — The Motley Fool — https://www.fool.com/investing/2026/01/03/golden-dome-project-begins-with-a-massive-151-bill/ — 2026-01-03 — C (cited as an example of the ceiling-vs-obligation conflation, not as evidence)
19. Sector Snapshot: Defense Startup Funding Hits An All-Time Record As VCs Begin To Eye Exits — Crunchbase News — https://news.crunchbase.com/defense-tech/startup-venture-funding-all-time-record-ai-anduril/ — 2026-06-02 — B
20. US Air Force awards first CCA production contracts to General Atomics, Anduril — Defense News — https://www.defensenews.com/industry/techwatch/2026/06/18/us-air-force-awards-first-cca-production-contracts-to-general-atomics-anduril/ — 2026-06-18 — B
21. Anduril's Pentagon contract turning point — Fortune — https://www.fortune.com/2026/03/22/anduril-pentagon-contract-turning-point — 2026-03-22 — B
22. Anduril Raises $5B Series H at $61B Valuation — Tectonic Defense — https://www.tectonicdefense.com/anduril-raises-5b-series-h-at-61b-valuation/ — 2026-05-29 — C (cited only to document the valuation contradiction)
23. Space Force plans to invest billions in sprawling Space Data Network in FY27 — DefenseScoop — https://defensescoop.com/2026/04/28/space-force-space-data-network/ — 2026-04-28 — B
24. Networks Business Propels Revenue Growth for SES, Operator Cancels 2 GEO Sats — Via Satellite — https://www.satellitetoday.com/finance/2026/05/12/networks-business-propels-revenue-growth-for-ses-operator-cancels-2-geo-sats/ — 2026-05-12 — B
25. SES joins Eutelsat in canceling GEO expansion satellites — SpaceNews — https://spacenews.com/ses-joins-eutelsat-in-canceling-geo-expansion-satellites/ — 2026-05-13 — B (headline and framing only; full text not retrievable, robots.txt restricted)
26. The State of Satcom 2026 — Payload — https://payloadspace.com/the-state-of-satcom-2026/ — 2026-04-06 — B
27. AST SpaceMobile pledges 2026 service launch — Telecoms.com — https://www.telecoms.com/satellite/ast-spacemobile-pledges-2026-service-launch — 2026-01-26 — B
28. Amazon seeks federal approval to launch 5,105 satellites for direct-to-device network — CNBC — https://www.cnbc.com/2026/07/27/amazon-satellite-internet-network.html — 2026-07-27 — B (headline and lede only; page returned 403 on retrieval)
29. Amazon Leo mission updates: 375+ satellites now in orbit after successful Atlas V launch — Amazon — https://www.aboutamazon.com/news/innovation-at-amazon/project-kuiper-satellite-rocket-launch-progress-updates — 2026-07-02 — A
30. 2026 New Glenn rocket explosion — Wikipedia — https://en.wikipedia.org/wiki/2026_New_Glenn_rocket_explosion — 2026-08-05 — C (used for a materially important event; underlying Blue Origin and FAA statements were not retrieved directly, and confidence on T-08-18 is capped accordingly)
31. eVTOL Certification Roundup: August 2026 — The Flight Brief — https://www.theflightbrief.com/articles/evtol-certification-roundup-august-2026 — 2026-08-26 — C
32. Orbital AI: Seattle-area startup Starcloud hits $1.1B valuation to build space-based data centers — GeekWire — https://www.geekwire.com/2026/orbital-ai-seattle-area-startup-starcloud-hits-1-1b-valuation-to-build-space-based-data-centers/ — 2026-03-30 — B
33. Korean Defense Exports Widen to Missiles, Jets, Submarines — Seoul Economic Daily (Korean-origin regional source) — https://en.sedaily.com/finance/2026/09/02/korean-defense-exports-widen-to-missiles-jets-submarines — 2026-09-02 — B
34. Korea's Defense Exports Hit $15.4 Billion in 2025, Set to Grow Further — Seoul Economic Daily — https://en.sedaily.com/politics/2026/03/25/koreas-defense-exports-hit-154b-in-2025 — 2026-03-25 — B
35. Ukraine Eyes Drone Production Topping 7 Million Units — Aviation Week (Robert Wall) — https://aviationweek.com/defense/supply-chain/ukraine-eyes-drone-production-topping-7-million-units — 2026-04-21 — B (excerpt only; full article paywalled behind AWIN)
36. EU, Ukraine sign drone deal to fast-track unmanned systems production — Breaking Defense — https://breakingdefense.com/2026/07/eu-ukraine-sign-drone-deal-to-fast-track-unmanned-systems-production/ — 2026-07-23 — B (publication date inferred from context; article did not state it explicitly — flagged)
37. Here's how the Army plans to spend nearly $1 billion in procuring small counter drone tech — Breaking Defense — https://breakingdefense.com/2026/05/heres-how-the-army-plans-to-spend-nearly-1-billion-in-procuring-small-counter-drone-tech/ — 2026-05-14 — B (headline and search metadata only; direct fetch failed with a redirect loop)
38. Department of Defense Counter Unmanned Aircraft Systems: Background and Issues for Congress (R48477) — Congressional Research Service — https://www.congress.gov/crs-product/R48477 — 2026-03-01 — A (existence and scope confirmed; full text not retrievable, robots.txt restricted)
39. Congress Returns While the FY2027 Funding Clock Ticks — Akin Gump, The Approps Drop — https://www.akingump.com/en/insights/blogs/the-approps-drop/congress-returns-while-the-fy2027-funding-clock-ticks — 2026-09-09 — B (page mixes a 2026-07-13 publication date with September updates; CR enactment status treated as unconfirmed)
40. Global Space Economy Reaches 429 Billion Dollars as Commercial Satellite Sector Dominates — SatNews — https://satnews.com/2026/05/14/global-space-economy-reaches-429-billion-dollars-as-commercial-satellite-sector-dominates/ — 2026-05-14 — C (market-size estimate; modeller not confirmed, figure recorded as estimate and not used as a denominator)

More research in this sector · Explore trend records