Defense Technology

Overview / thesis

The defense-tech investment landscape — drones, counter-UAS, defense electronics, directed energy, space robotics, and the critical minerals that feed the supply chain. This is where geopolitical risk becomes investment thesis.

The thesis in one paragraph

Defense spending is structurally increasing, not cyclically. The Citrini "War From Home" thesis argues that remote warfare — drones, autonomous systems, cyber — is reshaping military doctrine permanently: you don't need million-dollar missiles when $500 FPV drones work. Two rearmament supercycles — European and Japanese — underpin the entire theme directionally, layered on top of a US budget that crossed $1 trillion in FY2026. But stock selection within the theme matters enormously. The investable universe spans diversified primes re-rating on order growth, niche Tier 2/3 suppliers with sole-source moats and multi-year backlogs trading at sub-$250M enterprise values, and speculative pure-plays priced entirely on a future that hasn't happened. The spread between the cheapest and most expensive names runs from ~21x forward P/E to effectively infinite, and the quality dispersion is just as wide.

Why it matters: the structural-vs-cyclical question

The single most important framing for portfolio construction is whether this is structural or cyclical. The bull case — Citrini, Mawer's Joshua Samuel, and the small-cap deep dive all converge here — is that the shift is durable:

• Demand is multi-decade and policy-locked, not budget-cycle-dependent. NATO's Hague Summit endorsed a 3.5% GDP defense spending target (up from 2%). European defense budgets are projected to reach €800B by 2030 — €300B more than 2025. Japan's FY2026 defense budget hit ¥9.04 trillion ($58B), a 9.4% increase and the 14th consecutive record, already exceeding 2% of GDP two years ahead of schedule, with a five-year Defense Buildup Program allocating ¥43 trillion through FY2027.
• Supply, not demand, is the binding constraint. Decades of post-Cold War underinvestment (European spending fell from ~2.75% of GDP in the Cold War era) created production and capability gaps that cannot be solved quickly. Multi-year manufacturing build-out timelines give qualified, incumbent suppliers pricing power and visible revenue. Munitions stockpiles depleted by Ukraine and Israel reinforce this — the US Army targets 100,000 rounds/month of 155mm artillery; Patriot GEM-T backlog stretches ~5 years.
• The drone economics are real. Ukraine and the Middle East demonstrated mass employment of cheap drones and the unsustainability of $100K+ interceptor missiles against $500 attack drones. This is the demand pull behind both counter-UAS and directed energy.

The bear framing — the open question Pink flags — is that defense budgets remain politically volatile, and any ceasefire narrative or "spending fatigue" can trigger sharp corrections (Leonardo saw a 5-10% snap correction in August 2025). The resolution most sources land on: the theme is real and still in early innings, but it is a stock-picker's sector, not a buy-the-basket sector.

The "so what": where the risk/reward actually favors you

The cross-cutting investment conclusion across the comparison files is that the macro tailwind is necessary but not sufficient. The same defense budget that lifts everyone is already priced into the obvious names. The edge is in dispersion:

• Quality primes at reasonable prices are the lowest-risk way to play it. LDO (Leonardo) recurs as the highest-conviction name — cheapest major EU defense stock at 21-24x forward P/E with a sub-1.0 PEG, €44B+ order backlog (2.5 years of revenue coverage), book-to-bill at 1.3x, and free operating cash flow that quadrupled from €209M (FY2022) to €826M (FY2024). Its discount to Rheinmetall (39x forward P/E) and Saab (46x) is the cushion no other name offers.
• Tier 2/3 suppliers with designed-in moats offer embedded value the market overlooks — too small for institutions, too specialized to disrupt. The recurring screen: sole-source/designed-in positions (3-5 year MIL-SPEC qualification cycles create genuine moats), 2x+ revenue in backlog, zero/low debt, production-stage (not development-stage) programs, and margin-expansion runway. ESP (Espey) and CVU (CPI Aerostructures) repeatedly screen as the best backlog-to-valuation disconnects — CVU's 0.13x EV/Backlog ($509M backlog vs. ~$70M EV) is the starkest in the universe.
• Speculative pure-plays carry the most upside and the most risk. ONDS (Ondas) is the recurring cautionary tale — a $4B+ market cap on ~$48M revenue (~90x trailing sales) with 355% share dilution in 14 months. LASR (nLIGHT) trades at 13.7x EV/Revenue and ~203x forward P/E on the directed-energy production-contract optionality.

Sizing the opportunity (TAM / market builds)

The opportunity is best understood as a stack of sub-markets rather than one number, since "defense tech" spans budgets, weapons, and component sub-segments. The discrete figures recurring across the sources:

Market / segment	Size & growth	Source context
US national defense spending (FY2026)	>$1 trillion total; DoD request $892.6B + $150B reconciliation ($20.4B for munitions/supply chain)	Small-cap deep dive
US RDT&E (FY2026)	$179B, +27% YoY — one of the largest R&D allocations in defense history	Small-cap deep dive
US Navy shipbuilding (FY2026)	$47.3B, up from $28B in 2024	Small-cap deep dive
Global arms sales (primes)	~$632B; 5 primes (Lockheed, RTX, Northrop, GD, Boeing) dominate post-1993 "Last Supper" consolidation	Small-cap deep dive
European defense budgets (2030E)	€800B by 2030, +€300B vs. 2025; each 1% GDP increase ≈ €6B additional business for Leonardo	Four-stocks-ranked
Japan Defense Buildup Program	¥43 trillion through FY2027; FY2026 budget ¥9.04T ($58B)	Four-stocks-ranked
Directed energy weapons	~$12.4B (2025) → ~$28B (2034), 9.5% CAGR; high-energy laser sub-segment growing faster	Fiber-laser primer
Directed energy weapons (alt sizing)	$5-10B+ by 2030	FEIM-vs-LASR
Industrial fiber laser	~$6.9-7.7B (2024) → ~$17.5B (2033), 11.1% CAGR — but commoditizing for Western players	Fiber-laser primer
Precision timing (space/defense)	~$2-3B TAM, small incumbent share	FEIM-vs-LASR
US mil-spec frequency/RF	$500M-$800M	MPTI showdown
US military power electronics	$3B-$5B	MPTI showdown
Advanced imaging sonar + diver tech	$300M-$500M	MPTI showdown
Global aerostructures	$15B+ (SAM smaller for any single supplier)	MPTI showdown

A key dynamic across the component sub-segments: the defensible markets are defense, medical, ultrafast, and specialty applications where ITAR/security-clearance barriers and deep technical moats keep Chinese commoditization out. In standard industrial laser cutting/welding, Chinese manufacturers (Raycus, MaxPhotonics) have permanently impaired Western economics — the same lesson applies broadly: pick segments China can't easily enter.

The demand drivers, distilled

1. Counter-drone / counter-UAS as the killer app. Cheap drone swarms make missile-based defense economically unsustainable. Lasers flip the cost-exchange ratio — pennies-to-dollars per shot vs. tens of thousands per interceptor, with infinite magazine depth and speed-of-light engagement. The Pentagon stated (March 2026) it wants directed energy fielded at scale within 36 months; the Army's E-HEL is slated to be the first directed-energy program of record.
2. European rearmament supercycle — NATO 3.5% target, replacing Soviet-era equipment with NATO-standard systems, air/missile defense, ammunition replenishment, extended logistics.
3. Japanese / allied-Asia rearmament — friend-shoring of the defense industrial base, GCAP next-gen fighter, missile systems.
4. Munitions and supply-chain replenishment — depleted stockpiles plus reshoring/domestic-sourcing mandates (e.g., NDAA Section 834 mandating elimination of US defense reliance on Chinese/Russian optical glass by Jan 1, 2030).
5. Autonomous systems across domains — air (FPV/loitering munitions, tactical ISR), ground (UGVs), and subsea (AUVs, synthetic-aperture sonar), with the AI/autonomy layer as the value-capture question.
6. Defense-critical minerals — antimony, rare earths, tungsten. UAMY's DLA contracts ($597M+) prove the US treats mineral supply as national security; ablative composites and rare earths recur in The Meridian Report's coverage.

Sub-themes the sources develop (industry-level)

• Directed energy / high-energy lasers — the architecture competition (Lockheed Martin's spectral beam combining vs. nLIGHT's coherent beam combining, playing out in HELSI Phase 2 toward megawatt-class). Counter-drone pulls lasers from prototype to production. Israel's Iron Beam (Dec 2025) is the first operational HEL defense system; HELIOS and DE M-SHORAD are fielded. Timeline risk is the perennial caveat — directed energy has been "almost ready" for decades. See the fiber-laser/directed-energy detail on individual ticker pages (LASR, IPGP, COHR, FEIM).
• Drone / tactical ISR — pure-plays (Red Cat / RCAT) vs. subsystem suppliers (NextVision) vs. diversified primes (BAE, L3Harris, RTX). Blue UAS certification and FAA Type-Certification (Ondas Optimus) are regulatory moats; the open question is whether margin lives in the drone or the autonomy/AI layer.
• Defense electronics & components — frequency control/precision timing (MPTI, FEIM), RF, mission processing (MRCY), IR optics/photonics (LPTH). Mission-critical, embedded, ITAR-protected, with multi-decade program lifecycles.
• Subsea / naval — synthetic-aperture sonar, subsea batteries (PNG Kraken), submarine power electronics (ESP) tied to the Columbia-class program (DoD's #1 acquisition priority, $130B, 12 boats through 2042).
• Space as defense — MDA space robotics at the defense/commercial boundary; precision-timing payloads; proliferated-satellite contracts.

The key debates and open questions

1. Structural vs. cyclical? The dominant tension. Citrini and the rearmament-cycle bulls say structural; the skeptic says budgets are politically volatile and vulnerable to ceasefire narratives. Most sources resolve this as "structural but stock-pick it."
2. Drone-warfare commoditization — if FPV drones cost $500, where is the margin? Is the value in the airframe or the autonomy/AI/sensor layer on top?
3. Small-cap execution risk — the universe is full of great technology on terrible balance sheets. Separating compounders (sole-source moats, real backlog conversion) from capital incinerators (dilution-funded, development-stage) is the core skill.
4. Speculative valuation — names like ONDS and LASR price in a future that hasn't materialized; every quarter becomes a pass/fail test at those multiples.
5. Defense ESG tension — some allocators won't touch the sector. Mispricing opportunity or genuine portfolio constraint?
6. Architecture / technology bets — the SBC-vs-CBC directed-energy competition will determine which laser architecture the DoD standardizes for production; an analogous "which standard wins" risk runs through autonomous systems.

What's still missing (coverage gaps)

• No Palantir coverage (largest defense-AI / software-defined-defense company; named as a disruptive threat across the analytics/C2 layer).
• No large-cap prime deep dives (Lockheed, RTX, Northrop, L3Harris, GD, Boeing) beyond the prime-vs-supplier structural framing.
• No standalone counter-UAS market sizing (critical to the drone-warfare thesis).
• No consolidated defense budget tracker (US / allied / global trends in one place).
• No cyber-defense coverage.
• Ukraine / Taiwan scenario investment implications not mapped.
• PKE (sole supplier of PAC-3 ablative composites — Mike/Multibagger Monitor thesis) flagged but not yet written up.

How it works

Defense technology is not one business. It is a stack of physically distinct sub-sectors — directed-energy lasers, precision timing, RF and frequency control, infrared optics, autonomous drones and counter-UAS, subsea sonar, ruggedized power electronics, aerostructures, and the critical-mineral feedstocks underneath all of them. What ties them together economically is not the product but the qualification regime: once a component is designed into a weapons platform that survives shock, EMI, thermal, and nuclear-environment testing, it is embedded for the life of the platform. The moats here are physics plus paperwork — energy efficiency, beam quality, parts-per-billion frequency stability, non-germanium glass chemistry — multiplied by ITAR registration, MIL-SPEC qualification, and 3-5 year design-in cycles that nobody short-cuts. This section reproduces the technical mechanism, unit economics, and cost structure of each, drawn across the vault's directed-energy primer, the small-cap supplier deep dives, the drone/ISR work, and the peer comparisons.

Why the defense supply chain has the structure it does

The industry consolidated from 51 primes to 5 after the 1993 "Last Supper" — Lockheed Martin, RTX (Raytheon), Northrop Grumman, General Dynamics, and Boeing now control the bulk of roughly $632 billion in global arms sales. But primes cannot make everything. A modern weapons system runs to millions of parts; a Boeing 787 alone needs 45+ major companies and 4+ million parts. The Tier 1/2/3 structure exists because of three forces. First, complexity management — no single entity can efficiently control that supplier web. Second, specialization economics — the "investment-discouragement effect" outweighs the "markup-avoidance effect," so independent suppliers invest more in cost reduction than captive subsidiaries would. Third, and most important for investors, qualification barriers create moats: MIL-PRF-3098, MIL-S-901D Grade A shock, MIL-STD-461 EMI, MIL-STD-740 noise, AS9100, ITAR registration, and CMMC compliance impose 3-5 year qualification cycles. Once a part is designed into a platform, switching costs are prohibitive for the platform's lifetime — typically 20-40 years. The binding constraint in the current cycle is supply, not demand: the defense industrial base is capacity-constrained, which hands qualified incumbents pricing power and multi-year visibility.

The economic signature this produces, visible across the small-cap names, is the backlog-to-enterprise-value disconnect — $500M+ backlogs trading at sub-$100M enterprise values, because the market discounts micro-cap illiquidity and conversion risk:

Company	EV	Backlog	EV/Backlog	Gross Margin	Debt	Moat
CVU	~$70M	$509M	0.13x	20-22%	$16M	Single-source missile wings / EW pods
MPTI	~$212M	$59M	3.6x	43-46%	$0	40+ sole-source frequency programs
ESP	~$75M	$141M	0.53x	35%	$0	Submarine power sole-source
CODA	~$82M	N/A	—	68-70%	$0	Only real-time 3D sonar
MOBX	~$27M	N/A	—	50-60%	High	Growing but fragile

The screening logic that falls out: sole-source or designed-in positions, 2x+ revenue in backlog, zero/low debt, production-stage (not development-stage) programs, margin-expansion runway from fixed-cost leverage, and multiple prime relationships to dilute customer concentration. Company-level detail lives on the ticker pages (MPTI, ESP, CVU, CODA, MOBX, ONDS, EVLV, LPTH, MRCY, PNG, FEIM, LASR, IPGP, COHR, LDO, 7013); the mechanisms below are industry-wide.

Directed-energy lasers — the physics, from first principles

Every laser runs on Einstein's 1917 mechanism: stimulated emission. Pump energy from semiconductor laser diodes is injected into a gain medium — for high-power systems, a length of optical fiber doped with rare-earth ions, almost always ytterbium. Pump photons excite the ytterbium ions to a higher energy state; when enough ions are excited you reach population inversion (more atoms in the upper level than the lower), the prerequisite for lasing. A passing photon of the right energy triggers an excited ion to drop and emit a second identical photon — same wavelength, phase, direction. Two become four, four become eight: exponential amplification. Fiber Bragg gratings (FBGs) at each end act as mirrors forming the optical cavity — one ~99% reflective, the other 10-50% reflective to let the output beam escape. The result is a coherent, monochromatic beam at ~1060-1080 nm for ytterbium fiber.

Ytterbium dominates for concrete physical reasons: a simple two-level energy structure with no excited-state absorption (so nearly every pump photon contributes to lasing); a small quantum defect (pump at ~915-976 nm, lasing at ~1060-1080 nm — the gap is waste heat, and a small gap means less heat, so more power before thermal damage); a broad absorption band (900-1000 nm, tolerant of diode wavelength variation); and an output wavelength absorbed well by metals, which is why fiber lasers dominate metal cutting. The fiber itself is fused silica, whose thin-strand geometry gives an extremely high surface-area-to-volume ratio — heat dissipation is inherently efficient compared to a bulky crystal or gas tube.

The architecture that unlocked high power is double-clad cladding pumping. The fiber has three layers: a single-mode doped core (10-30 μm) where lasing happens, an inner cladding (125-400 μm, sometimes 600+), and a lower-index outer cladding. Pump light is launched into the inner cladding, not the core, so you can couple in cheap high-power multimode diodes that need no beam quality; as the pump bounces through the inner cladding it repeatedly crosses the doped core and over several meters achieves >90% absorption. The laser light stays confined to the single-mode core and keeps excellent beam quality regardless of how sloppy the pump is. This decoupling of pump quality from output quality is what made kilowatt-class fiber lasers possible. Pump sources are GaAs (gallium arsenide) diodes at 915 or 976 nm: individual emitters at 5-15 W, diode bars at 40-200+ W, modules at hundreds of watts, fused by pump combiners to inject thousands of watts. Pump diodes are the single most expensive component aside from the complete assembly. FBGs also act as pump reflectors, bouncing unabsorbed pump back for a second pass — shown to lift slope efficiency from 17% to 23% in erbium systems — and because they are written directly into the fiber there are no free-space optics to align or contaminate, a major reliability edge.

Scaling lasers to weapon power — beam combining

A single fiber laser tops out around 5-10 kW before nonlinear effects — stimulated Raman scattering (SRS), stimulated Brillouin scattering (SBS) — and transverse mode instability (TMI) wreck beam quality. To reach 50 kW, 100 kW, 300 kW, or ultimately 1 MW, you combine many fiber lasers. Two approaches compete:

• Spectral Beam Combining (SBC) — each laser runs at a slightly different wavelength; a diffraction grating overlaps them into one beam. Completely passive, no active control, robust, graceful degradation if one laser fails. This is Lockheed Martin's approach (300 kW HELSI demonstrator). Downside: the output carries multiple wavelengths, so it is not truly single-frequency.
• Coherent Beam Combining (CBC) — all lasers seeded from a common source and amplified in parallel, with active phase control locking them so they interfere constructively. Brightness scales as N² (N = number of beams), preserving single-frequency output and theoretically higher brightness. This is nLIGHT/Nutronics' approach, targeting megawatt-class. Downside: continuous feedback sensing and complex phase locking.

The SBC-vs-CBC contest is the Lockheed-vs-nLIGHT HELSI competition, and its outcome will steer which architecture the DoD standardizes for production weapons.

Laser unit economics — the cost-per-shot revolution

The entire defense-laser thesis is one number: cost per shot. An interceptor missile costs $10,000-$100,000+. A laser shot costs pennies to a few dollars in electricity. Against mass drone swarms — dozens of $500-$5,000 one-way attack drones — the cost-exchange ratio of missile interceptors is unsustainable; lasers invert it. The Army expects each ~$4.17M E-HEL unit to engage swarms of up to 30 small drones, rockets, or mortars at up to 4 km within 1 minute — roughly $30 of electricity to defeat $15,000-$150,000 of drones. Lasers also offer an effectively infinite magazine (no ammunition to run out during a sustained swarm), speed-of-light engagement (no flight time or lead calculation), graduated response (dazzle → disable sensors → destroy), and no ammunition logistics chain. The physics knobs that determine whether a system fits on a truck:

Wall-plug efficiency (WPE) — optical out / electrical in — directly sizes the power supply. At 30% WPE a 300 kW laser needs 1 MW of input; at 50% it needs 600 kW. That difference decides truck-mounted vs. dedicated generator trailer.

Laser type	Typical WPE
Fiber (Yb-doped)	30-50%
Direct diode	40-60%
CO2	10-15%
Nd:YAG (diode-pumped)	10-20%
Nd:YAG (lamp-pumped)	2-5%

Beam quality — the Beam Parameter Product (BPP, mm·mrad) and M² factor (M²=1.0 is diffraction-limited; real single-mode fiber achieves 1.0-1.3, multimode 5-20+) — sets how much power lands on a small spot at long range. For a 300 kW laser to kill a missile at 1+ km you need excellent beam quality, which is why beam combining must preserve brightness, not merely add power. Power density at focus runs ~10⁶-10⁷ W/cm² for cutting, 10⁵-10⁶ for welding, 10³-10⁵ for surface treatment. Wavelength selection: Yb fiber at 1030-1100 nm (the workhorse), Er at ~1550 nm ("eye-safe," telecom/LIDAR), Tm at ~1900-2000 nm (medical, growing), Ho at ~2050-2100 nm. Pulse modes span continuous-wave (CW, for cutting/welding/directed energy, 1 kW to 100+ kW), quasi-CW, nanosecond pulsed, and ultrafast picosecond/femtosecond ("cold ablation" with GW-TW peak power for damage-free micromachining).

The power-class ladder for defense:

Power	Application	Status
2-5 kW	Counter-UAS (small drones), dazzling	Fielded (IPG Crossbow, Boeing CLWS)
10-20 kW	Counter-UAS (Group 1-2), sensors	Fielded/testing
50 kW	Counter-UAS all groups, rockets, mortars	Early deployment (DE M-SHORAD / Guardian on Stryker)
100 kW	Cruise missiles, larger targets	In development
300 kW	Anti-ship cruise missiles, hardened targets	Prototype (HELSI Phase 1)
500 kW	Extended range, faster kill	Development (HELSI Phase 2, Lockheed)
1 MW+	Ballistic / hypersonic threats	Development (HELSI Phase 2, nLIGHT/Nutronics)

The persistent risk on the economics: directed energy has been "almost ready" for decades, the timeline keeps slipping, and manufacturing reliable, maintainable laser weapons in rugged field conditions remains unsolved. Countermeasures (reflective coatings, smoke/chaff, maneuvering) and atmospheric turbulence at multi-kilometer range can blunt effectiveness — which is why adaptive optics for atmospheric correction is the gating technology for making megawatt lasers practically deployable.

The laser value chain and where margin sits

Five tiers, each with different competitive dynamics — and the key structural insight is that this industry rewards vertical integration more than almost any other, because component performance is interdependent (pump wavelength → fiber absorption → thermal load → beam quality), specialty components have few suppliers, each tier adds margin, and the DoD increasingly demands domestic supply chains.

1. Semiconductor pump diodes — GaAs chips at 793-976 nm. nLIGHT runs its own GaAs fab (Vancouver, WA), modules to 700 W; Coherent shipped 793 nm modules at 50 W CW with 100,000-hour MTBF; IPG and TRUMPF make their own. Top five suppliers hold ~38% of the market; lead times stretch 26-32 weeks; GaAs wafer production is geographically concentrated.
2. Specialty optical fiber — Yb-doped double-clad silica, e.g. 20/400 μm core/cladding at low NA (0.06), anti-SRS / anti-TMI designs. Coherent/Nufern is the largest independent maker — a chokepoint, since anyone not vertically integrated buys gain fiber from a competitor.
3. Laser source integrators — assemble diodes, fiber, FBGs, combiners into the "laser engine." IPG (most complete vertical integration, ~$1.0B revenue, historically 50%+ gross margin), TRUMPF (~17% share, #2), nLIGHT, plus Chinese players Raycus (~30% domestic share) and MaxPhotonics.
4. OEM machine builders — integrate engines into cutting/welding machines (TRUMPF, Amada, Bystronic, Han's Laser). Top ten hold >60% share.
5. End users — industrial (welding 37% / cutting 21% of an IPG-type mix), defense (directed energy), medical (urology/lithotripsy ~$758M, 8% CAGR — thulium fiber lasers have 4x the water absorption of holmium and cut 30-50% faster), semiconductor (laser annealing ~$816M, critical for sub-5nm), telecom EDFAs (~$898M, 7.2% CAGR), and additive manufacturing.

The defining industrial dynamic: Chinese commoditization. Raycus and MaxPhotonics capture ~70% of China's 1-6 kW segment; Chinese cutting machines sell at $10K-$30K against $150K-$500K Western equivalents. The pattern repeats up the power curve (1→3→6→10→now 12-20 kW), high-power CW average selling prices fall ~7%/year, and IPG's revenue slid from a $1.4B (2018) peak to ~$1.0B (2025). The lesson for investors: the industrial fiber-laser market is structurally unattractive for Western players competing on the source alone — the defensible segments are defense (ITAR + clearances + decade-long programs + beam-combining know-how), medical, ultrafast, and specialty, where Chinese firms do not yet compete. nLIGHT's response — exiting cutting/welding entirely to go all-in on defense — is the cleanest expression of this thesis (detail on LASR); IPG fights a two-front war defending industrial while building defense from near-zero via Crossbow (detail on IPGP).

Barriers to entry in defense are high and durable: multi-hundred-million-dollar GaAs fabs and epitaxial-growth expertise; specialty-fiber doping/index/draw know-how accumulated over decades; beam combining at 100+ kW (especially CBC with adaptive optics); and systems integration spanning thermal management, beam directors, atmospheric propagation, and target tracking — none of which are component-level competencies. In industrial, those barriers are eroding (China has shown the core tech can be replicated, patents are hard to enforce globally, and component supply chains are opening up).

Precision timing and frequency control — the metronome inside every weapon

The opposite of a laser in size and visibility, identical in mission-criticality. Crystal oscillators exploit the piezoelectric effect in quartz: stress the crystal and it generates a precise electrical oscillation. An OCXO (oven-controlled crystal oscillator) holds the crystal in a temperature-controlled chamber to reach stability measured in parts-per-billion; TCXOs and DOCXOs are variants. The product family (oscillators, RF filters, crystal resonators, integrated microwave assemblies) operates from 1 MHz to 90 GHz. Higher up the stack sit rubidium atomic clocks and hydrogen masers — the most precise references — plus secure GPS receivers. These are the smallest, cheapest component in a satellite or weapon, and the one that makes everything else work:

• Radar — precise timing enables target detection and Doppler processing.
• GPS-denied navigation — when satellites are jammed, a precision local oscillator holds position accuracy.
• Electronic warfare — frequency-agile jammers need stable reference oscillators.
• Secure comms — encrypted radio links require synchronized timing across the network.
• Satellites and missile guidance — without precise timing, satellites drift and guidance fails.

The unit economics are the moat. The timing component is a tiny fraction of total system cost ("nobody cancels a $500M satellite to save $200K on oscillators"), which delivers pricing power and budget resilience, while decades of flight heritage (FEIM: 120+ space programs over 63 years; MPTI: 40+ sole-source programs running through 2030-2035+) create switching costs no competitor can replicate. The forward-looking growth vectors are miniaturization for new platforms — e.g. miniaturized atomic clocks for drone-fleet timing and airborne radars (~$20M annual opportunity class) — and missile-defense ("Golden Dome") timing demand. The lone structural threat: MEMS oscillators (SiTime) gradually displacing quartz in some applications. Company specifics live on FEIM and MPTI.

Infrared optics — the chemistry of seeing heat

IR imaging systems (thermal sensors, EO/IR seekers, shipboard surveillance, counter-UAS optics, Navy SPEIR programs) need lenses that transmit long-wave infrared, which ordinary glass blocks. The traditional material is germanium, but germanium is a Chinese-controlled critical mineral, and that is precisely the chokepoint the technology and the legislation revolve around. The differentiated alternative is chalcogenide glass (the BlackDiamond family) — a non-germanium, moldable IR glass that is one of very few NDAA-compliant germanium replacements. The economic catalyst is therefore legislative rather than merely cyclical: Section 834 of the FY2026 NDAA legislatively mandates eliminating U.S. defense reliance on optical glass from China and Russia by January 1, 2030, and China's own germanium export restrictions tighten the squeeze. Vertical integration matters here too — owning the glass chemistry up through camera-module assembly captures margin and locks in supply. The competitive set at larger scale is Coherent/II-VI and Teledyne FLIR. Company detail on LPTH.

Defense electronics processing and RF/EW

The mission-computing and signal-processing layer — RF/EW front ends through secure microelectronics — is governed by open-architecture standards (SOSA) that let modules from different vendors interoperate, plus ITAR and security-clearance barriers and switching costs across hundreds of fielded programs (Mercury Systems spans 300+ programs in 35+ countries). The economic profile is mid-margin (target 40%+ gross, depressed to ~28% through a trough), heavy on R&D, and protected by program incumbency. The disruption risk is insourcing: large primes (Lockheed, RTX, Northrop, L3Harris) can pull subsystem work back in-house. See MRCY.

Drones and ISR — where the value migrates to autonomy

The tactical ISR drone stack splits into platform makers (Red Cat, Skydio, AeroVironment), subsystem specialists (NextVision's micro stabilized E/O gimbals), and prime integrators (BAE, L3Harris, RTX) layering C4ISR — Command, Control, Communications, Computers, Intelligence, Surveillance, Reconnaissance. The technical differentiators are autonomy level (manual remote → full autonomous), sensor package (EO/IR, LIDAR, synthetic aperture radar, multi-spectral), endurance (30+ minutes for small tactical, hours for larger), and integration into broader command systems. A reference small platform: the Teal 2 quadcopter with a FLIR Hadron 640R EO/IR sensor at 30+ minutes endurance.

Two regulatory mechanics drive the economics. Blue UAS certification is the path to U.S. government procurement, created by DoD initiatives to replace Chinese-made small UAS with domestic alternatives — a policy-manufactured demand pool. FAA Type Certification (held by Ondas' Optimus, the first FAA Type-Certified autonomous drone) and BVLOS (beyond-visual-line-of-sight) waivers form a regulatory moat for autonomous operation. The open strategic question — and the one the vault repeatedly flags — is commoditization: if an FPV drone costs $500, the margin is not in the airframe but in the autonomy/AI layer and the counter-UAS stack on top. Counter-UAS (C-UAS) is the pull-through application for the entire laser story; the drone threat is what makes infinite-magazine, cheap-per-shot lasers economically necessary. Company specifics: ONDS; competitive landscape from the drone/ISR analysis.

Subsea robotics and sonar — seeing underwater in zero visibility

Two distinct technical moats. First, synthetic aperture sonar (SAS) and real-time 3D volumetric imaging sonar (the Echoscope family) — the latter generates complete 3D models from 16,000+ soundings per acoustic transmission, refreshed 12-20 times per second for video-like underwater data where competitors offer only 2D or slow 3D. Real-time volumetric perception is the candidate "standard sensor" for the emerging UUV/AUV (unmanned/autonomous underwater vehicle) market, projected ~$11B by 2030 — the underwater analog to the aerial-drone revolution. The Echoscope PIPE NANO ("a shade bigger than a smartphone") miniaturizes this for small underwater drones. The Diver Augmented Vision Display (DAVD) pipes real-time 3D sonar into a diver's helmet head-up display under an exclusive Navy license. These technologies command 66-70% gross margins — the pricing power of a genuine monopoly. Second, subsea batteries (SeaPower) certified by the U.S. Navy. Applications: mine countermeasures, port security, submarine operations, UUV guidance, salvage. See CODA and PNG.

Ruggedized power electronics — the moat is physical survival

MIL-SPEC power supplies, transformers, and converters spanning 12V to 45,000V and 20W to 3.8MW for severe-environment use. The moat is literally physical: submarine power electronics must survive extreme shock (MIL-S-901D Grade A), operate silently (MIL-STD-740 noise specs), withstand EMI (MIL-STD-461), and function in nuclear environments. Qualification testing spans years; once designed in, the part cannot be un-designed from a nuclear submarine already under construction. The flagship demand driver is the Columbia-class submarine program — the DoD's #1 acquisition priority, a ~$130B program building 12 submarines through 2042, the largest subs ever built with first-of-kind electric-drive propulsion. The economics: net-cash balance sheets, mid-30s% gross margins expanding on fixed-cost leverage, and backlog at ~3x annual revenue providing multi-decade visibility. Single-facility concentration is the offsetting operational risk. See ESP.

Aerostructures and contract manufacturing — the weakest moat

Structural assemblies — pod structures, wing assemblies, control surfaces, engine inlets, complex metallic assemblies — built to print on fixed-price and cost-plus contracts. This is the model with the thinnest moat in the sector: build-to-print providers do not own proprietary technology, so margins are structurally lower (15-22% gross vs. 35-66% for IP-owning peers), and fixed-price contracts carry tail risk (a legacy A-10 program wind-down produced ~$8.1M in estimate-at-completion charges, depressing gross margin to ~13% before normalizing to ~22% once clean). Protection comes from program lock-in, certifications, and single-source status on specific programs (e.g. NGJ-MB electronic-warfare pods, single-source missile wing assemblies). The value driver is backlog conversion at normalized margin, not technology. See CVU.

Munitions, missiles, and ablative materials — the consumables layer

The hardware-intensive replenishment layer is where current budget growth concentrates, and the economics are pure consumables-at-capacity. Ukraine and Israel have depleted stockpiles; the Army targets 100,000 rounds/month of 155mm artillery, and the Patriot GEM-T backlog stretches ~5 years. The supply-chain insight is sole-source feedstock dependency — e.g. ablative composites for the PAC-3 interceptor, where a single supplier (PKE) holds the position, and antimony as a defense mineral (2-3% of antimony goes to ammunition/ordnance, with flame retardants — 43-49% of demand — also defense-relevant for fireproofing vehicles, ships, aircraft). The U.S. treats this as national security: UAMY's antimony franchise sits on $597M+ in DLA strategic-stockpile contracts precisely because it is the only Western smelter. Reference systems span MLRS/HIMARS (M270, M142), ATACMS, thermobaric (TOS-1A), and precision-guided munitions. Defense minerals detail lives under critical-minerals and UAMY.

The macro demand mechanics — structural, not cyclical

The unit economics above only matter because the spending backdrop is reclassifying from cyclical to structural. The vault's framing (Citrini's "War From Home") is that remote warfare — drones, autonomous systems, cyber — permanently reshapes doctrine: you do not need million-dollar missiles when $500 FPV drones work. The hard numbers:

• U.S.: FY2026 DoD request $892.6B plus a $150B reconciliation package (including $20.4B for munitions/supply chain), total national defense spending exceeding $1 trillion. RDT&E allocation of $179B, a 27% YoY increase. Navy shipbuilding jumped to $47.3B (from $28B in 2024). Pentagon directive (March 2026): field directed-energy weapons at scale within 36 months; E-HEL slated as the first directed-energy program of record.
• Europe: NATO's Hague Summit endorsed a 3.5% GDP defense target (up from 2%); European budgets projected to reach €800B by 2030, ~€300B above 2025. Decades of post-Cold-War underinvestment (Cold War spending ran ~2.75% of GDP) created production constraints that cannot be solved quickly — multi-year supply chains hand contractors long demand visibility and pricing power. Each 1% of GDP increase in EU defense spending translates to roughly €6B of additional business for a major prime like Leonardo.
• Japan: FY2026 defense budget ¥9.04 trillion ($58B), +9.4%, the 14th consecutive record; spending already exceeded 2% of GDP two years ahead of schedule, with the five-year Defense Buildup Program allocating ¥43 trillion through FY2027. The aero-engine aftermarket compounds in parallel (~20% CAGR on the maturing V2500 fleet and growing GEnx installed base).

Two recurring structural tensions sit underneath the economics. The commoditization question — if drones and some laser power-classes commoditize, margin migrates to autonomy/AI, beam combining, and ITAR-protected niches. And the timeline-slippage question — directed energy's history of being perpetually "around the corner" means production-contract transitions (the step from $171M development contracts to $500M-1B+/year production) are the real value inflection, and they keep moving right.

Subsectors

The defense-tech universe in this vault is not one market. It is a dozen distinct sub-areas, each with its own physics, its own buyer behavior, its own moat structure, and its own investment angle. They share one macro tailwind — structurally rising defense spending, counter-drone urgency, space proliferation, and friend-shoring of the industrial base — but the way you make or lose money differs completely across them. What follows enumerates each sub-area: what it is, the technology, who plays, and where the money is.

Drones / ISR / Counter-UAS / Autonomous Systems

The headline subsector and the one driving the whole "war from home" thesis: cheap unmanned systems are rewriting military doctrine. You don't spend a million-dollar missile when a $500 FPV drone does the job. This breaks into three layers.

Tactical ISR platforms are small, rugged drones built for intelligence, surveillance, and reconnaissance. The differentiation runs along autonomy level (from manual remote operation to advanced onboard autonomy), sensor package (EO/IR, LIDAR, synthetic aperture radar, multi-spectral), and endurance (30+ minutes for small tactical quadcopters up to hours for larger systems). The single most important commercial gate is Blue UAS certification — the path to U.S. government procurement, created explicitly to replace Chinese-made small UAS (DJI) in U.S. operations. Players: Red Cat Holdings (RCAT, pure-play tactical manufacturer; Teal 2 quadcopter with FLIR Hadron 640R EO/IR sensor), NextVision (NXSN, a subsystem supplier specializing in micro stabilized E/O systems rather than full platforms), plus the diversified primes who fold UAS into broader portfolios — BAE Systems (modular scalable UAS architectures), L3Harris (C4ISR sensors and comms), and Raytheon/RTX.

Autonomous multi-domain systems combine air drones, ground robots (UGVs), and counter-UAS into one platform. ONDS (Ondas Holdings) is the vault's marquee name here — Optimus (the first FAA Type-Certified autonomous drone, a genuine regulatory moat), Raider (counter-UAS), Roboteam (tactical UGVs), and FullMAX private wireless. The model is acquisition-driven (Roboteam, Sentrycs, Apeiro, five acquisitions in 2025) and funded by a ~$1.5B cash war chest built through staggering dilution — shares went from ~93M to ~424M in 14 months (a 4.6x increase) with another 121.6M warrants outstanding. Guidance of $170-180M FY2026 revenue against a ~$48M FY2025 base is almost entirely M&A-dependent. The Ondas Networks segment, once a core thesis pillar, generates effectively zero ($82K quarterly).

Counter-UAS (C-UAS) is the defensive flip side and the killer app pulling lasers, RF jammers, and interceptors into production (see Directed Energy below). The drone threat validated by Ukraine and the Middle East makes counter-drone the most urgent procurement line in the sector.

Investment angle: the value migrates up the stack. If FPV drones cost $500, the margin is not in the airframe — it is in the autonomy/AI layer, the certification moat (Blue UAS, FAA Type Cert), and the counter-drone systems. Pure-play airframe makers face commoditization; the defensible positions are software, sensors, and regulatory gatekeeping. Disruptive private threats: Anduril, Shield AI, Skydio (none FAA Type-Certified, unlike ONDS).

Directed Energy / High-Power Fiber Lasers

The fastest-growing and most technically demanding subsector. Directed energy weapons (DEW) replace interceptor missiles with light. The core value proposition is cost-per-shot economics: an interceptor missile costs $10,000-$100,000+; a laser shot costs pennies to a few dollars in electricity. Against drone swarms — dozens of $500-$5,000 one-way attack drones — the missile cost-exchange ratio is unsustainable. Lasers flip it. They also offer infinite magazine depth, speed-of-light engagement, and graduated response (dazzle, disable, or destroy). Full technical primer at fiber-laser-directed-energy-primer.

The technology is the high-power ytterbium-doped fiber laser (1030-1100 nm), running on stimulated emission, cladding-pumped via cheap multimode GaAs semiconductor diodes, with fiber Bragg gratings as cavity mirrors. A single fiber tops out near 5-10 kW before nonlinear effects (stimulated Raman/Brillouin scattering) and transverse mode instability degrade the beam, so reaching 50 kW / 100 kW / 300 kW / 1 MW requires beam combining. Two architectures compete: Spectral Beam Combining (SBC) — passive, robust, graceful degradation, used by Lockheed Martin for its 300 kW HELSI demonstrator — versus Coherent Beam Combining (CBC) — active phase control, brightness scaling as N², used by nLIGHT/Nutronics, targeting megawatt class. The SBC-vs-CBC contest is the HELSI competition playing out, and which the DoD standardizes on will shape production laser weapons for a generation.

Market: directed energy weapons ~$12.4B (2025) → ~$28B (2034), 9.5% CAGR; the high-energy laser sub-segment grows faster. Power-class-to-application map:

Power Level	Application	Status
2-5 kW	Counter-UAS (small drones), dazzling	Fielded (IPG Crossbow, Boeing CLWS)
10-20 kW	Counter-UAS (Group 1-2), sensors	Fielded/testing (LOCUST, palletized)
50 kW	Counter-UAS all groups, rockets, mortars	DE M-SHORAD / Guardian on Stryker
100 kW	Cruise missiles, larger targets	In development
300 kW	Anti-ship cruise missiles, hardened targets	HELSI Phase 1 completed
500 kW	Extended range, faster kill	HELSI Phase 2, Lockheed Martin
1 MW+	Ballistic, hypersonic threats	HELSI Phase 2, nLIGHT/Nutronics

Key U.S. programs: HELSI (High Energy Laser Scaling Initiative — Phase 1 300 kW class via Lockheed/SBC, nLIGHT/CBC, General Atomics/distributed gain; Phase 2 scaling to 500 kW-1 MW, nLIGHT holds a $171M contract expanded from $86M, completion late 2026); DE M-SHORAD "Guardian" (50 kW on Stryker A1, live-fired at Fort Sill June 2025, four systems deployed to CENTCOM); IFPC-HEL (Lockheed prototype, demonstrator only); E-HEL (the Army's planned first directed-energy program of record, RFI Oct 2025, competitive selection Q2 FY2026, ~$4.17M per laser, each engaging swarms of up to 30 drones at up to 4 km within 1 minute); JLWS for cruise-missile defense feeding into Golden Dome; Navy HELIOS (60 kW on USS Preble, reportedly downed Iranian drones in 2026) and ODIN dazzler; DARPA MELT (Northrop). International proof points: Israel's Iron Beam (Rafael/Elbit, delivered to IDF Dec 2025, the world's first operational high-energy laser air defense, up to 10 km).

Players and the supply-chain split: laser-source companies supply the "engine" to defense primes who integrate the weapon system. LASR (nLIGHT) is the vault's pure-play — vertically integrated from its own GaAs wafer fab (Vancouver, WA) through pump diodes, specialty fiber, fiber lasers, beam combining (Nutronics, acquired 2019), to complete laser weapon systems (70 kW LWS, 30 kW, 10 kW HEL). It has bet the company by exiting industrial cutting/welding entirely in Q4 2025 to go all-in on defense; A&D revenue was $175.3M (67% of total), up 60%. IPGP (IPG Photonics) is the industrial fiber-laser giant fighting a two-front war — defending a shrinking industrial franchise against Chinese commoditization while building defense from near-zero via the Crossbow counter-UAS laser (3 kW MINI, 100+ drones neutralized, a ~$10M Lockheed order). Coherent (COHR) owns Nufern, the largest independent specialty-fiber maker and a supply-chain chokepoint. Lumentum (LITE) and TRUMPF (private, ~17% share, disk lasers) round out sources. Primes integrating: Lockheed Martin (SBC, Sanctum C-UAS), RTX/Raytheon (HELWS), Northrop (Phantom 10 kW <200 lbs, MELT), Boeing (CLWS 5 kW, GA partnership). Head-to-head FEIM vs LASR analysis at feim-vs-lasr-showdown.

Investment angle: industrial fiber laser is a commodity race to the bottom (IPG fell from ~$1.4B revenue in 2018 to ~$1.0B in 2025 under Raycus/MaxPhotonics pricing). The defensible market is defense — ITAR, security clearances, decade-long program timelines, and beam-combining know-how that Chinese makers cannot cross. nLIGHT is pure optionality on production contracts (potentially $500M-1B+/year by 2028-2030) but has never posted a profitable year in 25 years and trades at ~14x sales, 203x forward P/E. Timeline risk is the perennial caveat — directed energy has been "just around the corner" for decades; the difference now is operational urgency, fielded demonstrators, and a 36-month Pentagon "field at scale" directive (March 2026).

Defense Electronics, Processing & Power

The subsystems and "mission-critical electronics" layer inside platforms. Three flavors appear in the vault.

Mission processing and sensor electronics: MRCY (Mercury Systems) is the mid-cap leader — mission-critical processing and sensor electronics across 300+ defense programs in 35+ countries, the SOSA open-architecture standard-bearer, products spanning RF/EW through mission computing and secure microelectronics. ITAR barriers and 300+ program switching costs are the moat. The story is a margin-recovery turnaround (28% gross vs. 40%+ historical target) on a record ~$1.5B backlog. The diversified defense-electronics primes also sit here: Leonardo DRS, Indra Sistemas (IDR, Spain's radar/C3I champion), and Leonardo S.p.A.'s Defense Electronics & Security division.

Ruggedized power systems: ESP (Espey Manufacturing) builds MIL-SPEC power supplies, transformers, and converters for severe-environment military use (12V to 45,000V, 20W to 3.8MW). The moat is being designed into the platform — sole-source distribution transformers and power-distribution panels on Columbia-class submarines ($29.5M + $19.8M contracts), Virginia-class, Arleigh Burke DDG-51 (AMDR radar), E-2D Hawkeye, F-18. Once your panels are designed into a nuclear submarine under construction, you cannot be un-designed. A new $7.4M Navy-funded Magnetics Center of Excellence adds capacity. This is the closest thing to guaranteed revenue in defense — Columbia is the Navy's #1 acquisition priority, a $130B program building 12 submarines through 2042.

Defense connectors / EMI filtering / RF components: the deepest tier. MOBX (Mobix Labs) supplies EMI filtered connectors and Mil-Spec circular connectors across real programs (Navy vessels, M-1 Abrams via Honeywell, Tomahawk, Javelin guidance, Patriot via Spacecraft Components, F-35, military drones), holds BAE Gold Tier Supplier status — but is financially precarious (deeply unprofitable, ~$3.1M cash against $8.5M annual operating loss, 69% share dilution in a year, Nasdaq delisting risk). The defense exposure is real; the equity structure is not investable for most.

Investment angle: qualification barriers (MIL-PRF, MIL-S-901D, MIL-STD-461, AS9100, ITAR, CMMC) create 3-5 year qualification cycles; once designed in, components are embedded for the platform's life. Sole-source positions plus multi-year backlogs trading at low EV give embedded value, but customer concentration and program cancellation are the recurring risks.

Precision Timing / Atomic Clocks / Frequency Control

The "metronome" subsector — the smallest, cheapest component in a satellite or weapon, and the one that makes everything else work. Without precise timing, encrypted comms can't synchronize, satellites drift, and missile guidance fails.

Two adjacent technology families. Frequency control / RF components: crystal oscillators exploiting the piezoelectric effect in quartz — OCXOs (oven-controlled), TCXOs, DOCXOs — plus RF filters and microwave assemblies, achieving stability in parts-per-billion, operating 1 MHz to 90 GHz. MPTI (M-tron Industries, founded 1965) is the vault's pure-play: 40+ long-term defense/space programs, many sole-source, ROIC ~35%, the best profitability profile among small-cap defense electronics (46% gross, 19% op margin), debt-free. Mission-critical for radar (Doppler processing), GPS-denied navigation, electronic warfare (frequency-agile jamming), and secure comms.

Atomic clocks / precision timing for space: FEIM (Frequency Electronics, founded 1961) makes crystal oscillators, OCXOs, rubidium atomic clocks, hydrogen masers, and secure GPS receivers — on 120+ space programs over 63 years. Revenue mix: satellite payloads 59%, non-space DoD 38%, commercial 3%, ~94% U.S. The new TURbO miniaturized atomic clock targets a ~$20M annual market from FY2027 for drone-fleet timing and airborne radars. Catalysts: proliferated-satellite contracts (two ~$45M awards March 2026), Golden Dome / missile-defense timing. The competitive moat is heritage — 63 years of flight data creating switching costs no competitor can replicate; the product is too cheap relative to system cost to ever rip out (nobody cancels a $500M satellite to save $200K on oscillators).

Investment angle: the most durable, lowest-beta corner of defense-tech (FEIM beta 0.31). Mission-critical, deeply embedded, budget-resilient. TAM is modest (~$2-3B for precision timing) so growth is incremental, but the franchise is essentially impossible to displace. Watch threats from MEMS oscillators (SiTime) gradually displacing quartz in some applications.

Sonar / Underwater / Subsea Robotics

Undersea is its own warfare domain, riding the Navy shipbuilding budget surge ($47.3B FY2026 from $28B in 2024) and the coming underwater-drone (UUV/AUV) revolution — the aquatic analog of the aerial drone story.

Two vault names. CODA (Coda Octopus) makes the world's only real-time 3D/4D/5D/6D volumetric imaging sonar — the Echoscope series, 16,000+ soundings per acoustic transmission refreshed 12-20 times/second for video-like underwater perception in zero visibility. The DAVD (Diver Augmented Vision Display) puts that sonar into military divers' helmet head-up displays under an exclusive Navy license from NSWC Panama City (~12 Navy Commands including Special Forces; DAVD revenue inflecting $1.2M → $3.5-4.0M). The Echoscope PIPE NANO GEN (smartphone-sized) targets the ~$11B UUV/AUV market for small underwater drones. Gross margins 66-70% reflect a genuine technology monopoly; competitors offer only 2D or slower 3D. Defense apps: mine countermeasures, port security, submarine ops, UUV guidance, salvage; customers include the Navy, European navies, Raytheon (Phalanx CIWS), Northrop (mine hunting).

PNG (Kraken Robotics, Canadian, TSXV) provides synthetic aperture sonar (SAS), subsea SeaPower batteries (certified by the US Navy), and underwater robotics; ~70% defense revenue, highest gross margin in its peer group (59%), an Anduril partnership, riding NATO undersea-warfare spend. US investors carry currency and liquidity risk on the TSXV listing.

Investment angle: technology-monopoly pricing power (CODA) and patented batteries/sonar (PNG) make this a defensible niche, but both trade at steep multiples already pricing years of execution. The greenfield UUV market is the long-runway optionality.

Aerostructures

Structural manufacturing for military aircraft and electronic-warfare pods — wings, control surfaces, engine inlets, pod structures, complex metallic assemblies. The weakest-moat subsector because the dominant model is build-to-print contract manufacturing: program lock-in and certifications protect near-term revenue, but the supplier doesn't own proprietary technology and margins are structurally lower (15-22% gross).

Vault name: CVU (CPI Aerostructures). The crown jewel is the Next Generation Jammer Mid-Band (NGJ-MB) program — pod structures and Air Management Systems for the Raytheon EW system on EA-18G Growler aircraft ($80M+ funded since 2016, 300+ pods over a decade, IOC Dec 2024, Lot 5 Nov 2025). Also single-source on Raytheon missile wing assemblies, NGJ Low-Band for L3Harris, plus legacy T-38, MH-60 Seahawk, F-16. The striking feature is a $508.96M total backlog ($100.05M funded) against a ~$70M enterprise value — 0.13x EV/Backlog, the most extreme valuation disconnect in the small-cap defense set. The catch: an A-10 program wind-down created $8.1M in EAC charges, and fixed-price cost overruns plus an SEC restatement history (resolved) leave a governance scar. The bet is normalized 20%+ gross margins on backlog conversion.

The large diversified primes also carry aerostructures as a drag — Leonardo S.p.A.'s Aerostructures division (tied to Boeing 787 rates) and IHI's airframe work are dilutive relative to their defense-electronics and aero-engine segments.

Investment angle: optically cheap on backlog, but you're buying execution on fixed-price manufacturing, not a technology moat. Highest conversion risk in the group.

Photonics for Defense (Infrared Optics & Imaging)

Infrared optics, camera assemblies, and the specialty glass that feeds defense imaging and targeting — distinct from the high-power directed-energy lasers above (this is sensing/imaging, not weapons).

Vault name: LPTH (LightPath Technologies). Vertically integrated from proprietary BlackDiamond chalcogenide glass through IR camera assemblies, transformed by the 2025 G5 Infrared and 2026 AMI acquisitions from a components company into an IR imaging platform. The legislated catalyst is the single strongest small-cap driver in the vault: Section 834 of the FY2026 NDAA mandates eliminating U.S. defense reliance on optical glass from China and Russia by January 1, 2030 — and BlackDiamond is one of very few NDAA-compliant germanium replacements, reinforced by China's own germanium export restrictions. This is law, not a "might happen." Defense pulls: shipboard surveillance, counter-UAS, Navy SPEIR. Revenue accelerated to a $65M+ run rate (120% Q2 YoY), gross margin expanded 27% → 37%, adjusted EBITDA turned positive. Larger competitors: Coherent/II-VI, Teledyne FLIR. Segment mix: Assemblies & Modules 44%, IR Components 31%, Visible Components 21%.

Investment angle: a legislated demand driver with a genuine non-Germanium materials moat. The risk is valuation (~9-10x NTM EV/Sales vs. 2-4x for profitable optics peers), customer concentration, and a $200M shelf-registration dilution overhang.

Nuclear-Triad Primes / Nuclear Defense Adjacency

Nuclear sits at the energy-defense boundary. The triad (land-based ICBMs, submarine-launched, air-launched) and naval nuclear propulsion are the strategic-deterrence backbone, set by the 2022 Nuclear Posture Review. The vault flags Northrop Grumman (NOC) as the triad prime (analyzed alongside uranium miners in NOC-uranium-companies-northrop) — nuclear submarines, carriers, and propulsion. The submarine angle connects directly to the Sonar/Underwater and Power subsectors above (Columbia/Virginia-class drive ESP, CODA, and the $47.3B shipbuilding budget). This is large-cap and thinly covered in the vault — a flagged gap rather than a developed thesis.

Investment angle: the most politically durable defense spend (bipartisan missile-defense and deterrence support, Golden Dome), but concentrated in mega-cap primes the vault hasn't deep-dived.

Security Screening / AI Weapons Detection

Defense-adjacent rather than core military — physical-security screening for civilian venues, schools, and stadiums, but procured under the same threat environment.

Vault name: EVLV (Evolv Technology). AI-powered weapons-detection systems, ~70% share in AI weapons screening for venues, sold as a subscription/SaaS model (~85% of ARR) with hardware deployment. Highest gross margin in its peer set (53.7%) reflecting the recurring-revenue model, fastest organic growth among profitable-path names (41% YoY, $117M ARR). The knocks: deeply unprofitable (-42% op margin), negative FCF, and a financial-restatement history that is a real credibility risk.

Investment angle: a niche differentiator with a genuine SaaS recurring-revenue model and the clearest near-term path to profitability among the speculative growth names, but the end market (venues) is narrow and the restatement history is a governance flag.

Adjacent Sub-areas Found in Sources

Defense services / mission support: not a technology play at all but the low-margin labor-and-logistics backbone — VVX (V2X) provides logistics, aerospace maintenance, training, and IT mission support across 47 countries (~16,000 employees, $11-12.5B backlog, 8.6% gross margin). It is the cheapest name across every defense comparison (0.66x EV/Revenue) precisely because cost-plus government services carry no technology differentiation. A stable compounder, not a growth story.

Munitions, artillery & precision strike: the replenishment subsector driven by Ukraine/Israel stockpile depletion. The Army targets 100,000 rounds/month of 155mm artillery; the Patriot GEM-T backlog stretches ~5 years. Systems referenced: MLRS/M270, HIMARS (M142), ATACMS, TOS-1A (Russian thermobaric), plus PGMs and cluster munitions (glossary in utr-defense-primers). The European angle (Mawer/Maser, maser-military-landscape): decades of underinvestment (Cold War spend ~2.75% of GDP, collapsed post-1991) created capacity-constrained supply chains where limited supplier bases hand pricing power and multi-year visibility to incumbents — Germany's rearmament package leading, NATO's Hague Summit 3.5% GDP target, European budgets projected to €800B by 2030.

Electronic warfare (EW): cross-cuts ISR, defense electronics, and aerostructures — NGJ (Next Generation Jammer) programs reaching full-rate production, frequency-agile jamming requiring stable reference oscillators (the MPTI/FEIM tie-in), and EW pods (the CVU/NGJ-MB tie-in). Prioritized for Great Power competition.

Critical minerals for defense: the supply-chain floor. UAMY's antimony (the only Western smelter; $597M+ DLA strategic-stockpile contracts) feeds ammunition, ordnance, and flame retardants (fireproofing for military vehicles, ships, aircraft). Rare earths, tungsten, ablative composites (PKE as sole supplier of PAC-3 ablative composites per the Multibagger Monitor thesis) round it out. Coverage at critical-minerals and the Meridian Report source. The US government treating mineral supply as national security (DLA contracts) is the proof point.

Marine autonomous / wave-energy tech: a tangential, pre-commercial corner — Ocean Power Technologies (OPTT) makes PowerBuoy wave-energy converters, WAM-V autonomous surface vehicles, and the Merrows AI maritime-monitoring platform, with emerging defense relationships (USCG purchase order). Minimal revenue ($3.75M), extreme distress (Altman Z ~-11.67); included for completeness, not as a thesis.

Quantum & space for defense: flagged in the vault hub but thinly developed — quantum computing for cryptography, sensing, and GPS-denied navigation, and space robotics (MDA Space — satellite servicing at the defense/commercial boundary). Both are early-stage themes rather than developed subsectors here.

Coverage Gaps Flagged in Sources

The vault explicitly notes missing coverage that would otherwise be subsectors: Palantir (the largest defense-AI / software-defined-defense company, competing on the analytics/C2 layer across every subsector), the large-cap primes (Lockheed, RTX, Northrop, L3Harris) beyond passing reference, formal counter-UAS market sizing, a defense-budget tracker, and cyber defense (entirely uncovered).

Value chain

The defense-tech value chain runs from raw materials and components, up through subsystems and integrated weapon systems, to the prime contractors who sell platforms to governments. The single most important structural fact: the constraint is not demand, it is supply. The defense industrial base is capacity-constrained after decades of consolidation and underinvestment, which hands pricing power and multi-year visibility to whoever holds a qualified, designed-in position. Where the margin actually pools is counterintuitive — not at the giant primes, but at the small, obscure Tier 2/3 component suppliers with sole-source positions, and at the specialty-component chokepoints that no one can second-source quickly.

The prime / tier structure — and why primes don't vertically integrate

The U.S. defense industry consolidated from 51 to 5 major primes after the 1993 "Last Supper": Lockheed Martin, RTX (Raytheon), Northrop Grumman, General Dynamics, and Boeing now control the lion's share of the $632 billion in global arms sales. But these primes cannot manufacture everything themselves, and deliberately don't. The supply chain organizes into a Tier 1/2/3 structure: Tier 1 suppliers provide major subsystems directly to primes (Collins Aerospace, L3Harris); Tier 2 companies manufacture subassemblies and components for Tier 1s; Tier 3 suppliers produce individual components, raw materials, and specialized parts — often thousands of SKUs in small batches.

Three reasons primes don't insource everything, and these are exactly what create the investable moats downstream:

1. Complexity management — a modern weapons system contains millions of parts across thousands of suppliers. A Boeing 787 requires 45+ major companies and 4+ million parts. No single entity can efficiently control that complexity.
2. Specialization economics — academic research shows the "investment-discouragement effect" outweighs the "markup-avoidance effect": independent suppliers invest more in cost reduction and innovation than captive subsidiaries would.
3. Qualification barriers create moats — MIL-SPEC qualifications (MIL-PRF-3098, MIL-S-901D Grade A shock, MIL-STD-461 EMI, MIL-STD-740 noise), AS9100 certification, ITAR registration, CMMC compliance, and program-specific certifications create 3-5 year qualification cycles. Once a component is designed into a platform, it is virtually embedded for the platform's lifetime. You cannot un-design a part out of a nuclear submarine already under construction.

The investment implication is the spine of the whole sector: small-cap suppliers with sole-source positions on critical programs can exhibit surprisingly durable competitive advantages and pricing power — if they can execute. The flip side is that the primes (Lockheed, RTX, Northrop, L3Harris) can, and sometimes do, insource subsystem work that suppliers like MRCY and LPTH currently provide, which is the structural threat hanging over the merchant-supplier tiers.

Where the margin pools by tier

Margin in defense is inverted relative to size. Government services — the bottom of the value stack — run on cost-plus economics and earn single-digit gross margins. Differentiated components and proprietary subsystems earn 35-70% gross margins. The pattern across the tracked universe is stark:

Tier / business type	Representative names	Gross margin	Operating economics
Government services (logistics, training, MRO)	VVX	~8.3-8.6%	~4.3% op margin; cost-plus, commodity-like
Build-to-print contract manufacturing (aerostructures)	CVU	~18-22% normalized	~5-9% op margin; weakest moat, replaceable
Diversified defense primes / conglomerates	LDO, IDR, 7013	~12-18%	~9-10% EBITA; scale + program lock-in
Precision components, sole-source	MPTI, ESP	35-46%	17-22% op margin; designed-in, ITAR moat
Proprietary-technology subsystems	CODA, PNG	59-70%	technology monopoly pricing
Directed-energy / merchant laser (pre-scale)	LASR	~30%	negative op margin; investing ahead of production

The lesson from the services end: VVX trades at 0.66x EV/Revenue precisely because the cost-plus government services model structurally caps margin expansion — "no technology differentiation, low margins, slow growth," even on an $11-12.5B backlog. At the other end, CODA's 69-70% gross margins "demonstrate pricing power from technological differentiation" as the world's only real-time 3D/4D volumetric imaging sonar maker. Pricing power tracks proprietary IP and sole-source status, not scale.

Sole-source pricing power — the clearest evidence

The cleanest demonstration of how designed-in positions convert to pricing power is Espey Manufacturing (ESP), which holds exclusive sole-source positions for distribution transformers on U.S. Navy submarine programs (Columbia-class, Virginia-class). Qualification barriers are formidable — MIL-T-15108C, MIL-S-901D Grade A shock testing, MIL-STD-740 noise — and a competitor would need 3-5 years minimum to qualify, with switching unacceptable on a fixed production schedule running through 2042. The result: gross margins expanded from 12.1% (FY2021) to 28.9% (FY2025), a 1,680 basis point improvement, which the analysis reads directly as "substantial pricing authority." MPTI tells the same story from the RF/frequency-control angle — 40+ sole-source programs of record, 46% gross margins, ITAR-compliant domestic manufacturing in a reshoring era. Sole-source + qualification barrier + multi-decade program = pricing power.

Backlog-to-valuation as the supplier scorecard

Because supplier revenue is locked into multi-year program lifecycles, backlog coverage (and EV/Backlog) is the operative value-chain metric — book-to-bill above 1.0x is the signal the moat is still compounding. The micro-cap supplier comparison:

Company	EV	Backlog	EV/Backlog	Gross margin	Debt	Moat
CVU	~$70M	$509M	0.13x	20-22%	$16M	Single-source missile wings / NGJ-MB pods
MPTI	~$212M	$59M	3.6x	43-46%	$0	40+ sole-source programs
ESP	~$75M	$141M	0.53x	35%	$0	Submarine power sole-source
CODA	~$82M	N/A	—	68-70%	$0	Only real-time 3D sonar
MOBX	~$27M	N/A	—	50-60%	High	Growing but fragile

CVU's 0.13x EV/Backlog ($508.96M total backlog, $100.05M funded, against ~$70M EV) is the most striking disconnect in the small-cap space — the market is pricing in execution risk and micro-cap illiquidity, not the embedded value of converting that backlog at normalized 20% margins (~$102M of gross profit potential over contract lives). The crown-jewel program economics: CVU's Next Generation Jammer Mid-Band (NGJ-MB) pods for Raytheon's EA-18G Growler — $80M+ funded since 2016, 300+ pods over the next decade, "the most valuable program in our 40-year history."

The screening criteria that actually matter for picking winners in the supplier tier: (1) sole-source or designed-in positions; (2) backlog visibility of 2x+ revenue; (3) zero/low debt for covenant flexibility and M&A optionality; (4) production-stage vs. riskier development-stage programs; (5) margin-expansion runway from learning curves and fixed-cost leverage; (6) multiple prime relationships — customer concentration is the primary risk factor. (See MPTI, ESP, CODA, CVU ticker pages for company specifics.)

The fiber laser / directed energy value chain — five distinct tiers

The directed-energy supply chain is the cleanest worked example of a tiered value chain in this sector, with sharply different competitive dynamics at each stage:

Tier 1 — Semiconductor laser diodes (pump sources). GaAs-based diode chips emitting at 793-976 nm provide the pump energy; the single most expensive component in a fiber laser. The top five suppliers (Coherent/COHR, Lumentum/LITE, TRUMPF, ams-OSRAM, IPG) hold only ~38% of the market — moderate fragmentation. The choke point: lead times for specialized pump diodes stretch to 26-32 weeks, GaAs wafer production is geographically concentrated, and 2025 US tariffs added complexity for diode importers. Building a diode capability requires multi-hundred-million-dollar GaAs fabs and deep epitaxial growth expertise (LASR runs its own GaAs wafer fab in Vancouver, WA; completed a $22M epitaxy/MOCVD upgrade in 2025).

Tier 2 — Specialty optical fiber. Ytterbium-doped, double-clad silica gain fiber. This is the most underappreciated chokepoint: Coherent/Nufern is the largest independent specialty-fiber manufacturer, so any laser maker not vertically integrated (i.e., not IPGP or nLIGHT) is likely buying its gain fiber from Coherent — a competitor. Manufacturing it requires decades of accumulated know-how in doping and refractive-index profiles.

Tier 3 — Fiber laser source manufacturers ("the laser engine"). Assembles diodes, fiber, FBGs, combiners, and electronics. IPG Photonics (IPGP) is the dominant, most vertically integrated player (~$1.0B revenue, 50%+ gross margins at peak); TRUMPF is #2 at ~17% share. Chinese players Raycus (~30% domestic share) and MaxPhotonics together hold ~70% of China's mid-power (1-6 kW) segment.

Tier 4 — OEM machine builders. Integrate laser sources into cutting/welding machines. Top ten OEMs hold >60% global share (TRUMPF, Bystronic, Amada at the high end; Han's Laser, HSG, Bodor in China).

Tier 5 — End users. Industrial (automotive, aerospace, shipbuilding), defense (directed-energy weapons, LIDAR), medical (urology/lithotripsy), semiconductor (laser annealing, via drilling), telecom (EDFAs), additive manufacturing.

The margin-capture logic of vertical integration. This industry rewards vertical integration more than most because component performance is interdependent (pump wavelength → fiber absorption → thermal load → beam quality), specialty components have limited suppliers, each tier adds margin, and the DOD increasingly demands domestic control of critical components. IPG's historically high margins came from capturing margin at every tier. The most integrated players — IPG (most complete), nLIGHT (semiconductor through complete weapon system), TRUMPF (diodes through machine tools), Han's Laser (sources through machines) — have structural cost and performance advantages over pure component suppliers.

Choke points and bottleneck tiers

The sector's bottlenecks cluster in specialty materials and qualified capacity, not in assembly:

• Specialty optical fiber (Nufern/Coherent) — the upstream chokepoint of the laser chain; if you don't make your own, you buy from a rival.
• GaAs pump-diode fabs — geographically concentrated, 26-32 week lead times, tariff-exposed.
• Infrared optical glass / germanium — China restricts germanium exports; Section 834 of the FY2026 NDAA legislatively mandates eliminating U.S. defense reliance on optical glass from China and Russia by January 1, 2030. LPTH's proprietary BlackDiamond chalcogenide glass is one of very few NDAA-compliant germanium replacements — a legislated demand catalyst, not a maybe. The G5 camera business booked $80M+ in orders in its first year.
• PAC-3 ablative composites — PKE is cited as the sole supplier of ablatives for the PAC-3 interceptor (Mike/Multibagger Monitor defense thesis), a single-source position deep in the missile-defense supply chain.
• Defense minerals — antimony — UAMY as the only Western antimony smelter, underpinning $597M+ DLA strategic-stockpile contracts; 2-3% of antimony goes to ammunition/military with flame retardants (43-49%) also defense-relevant. Antimony, rare earths, and tungsten are the raw-material floor of the chain.
• Munitions and interceptor capacity — the Army targets 100,000 rounds/month of 155mm; the Patriot GEM-T backlog stretches ~5 years. Demand exists; qualified production capacity is the binding constraint.
• MIL-SPEC qualified capacity broadly — the 3-5 year qualification cycle is itself a chokepoint, since new entrants can't simply add supply.

The European angle reinforces the same dynamic (Mawer/Joshua Samuel): decades of post-Cold-War underinvestment (spending fell from ~2.75% of GDP) created manufacturing bottlenecks that "cannot be quickly solved," giving the limited supplier base pricing power and margin expansion, with multi-year contract visibility — exactly why European defense contractors are argued to have higher growth potential than mature U.S. names.

Merchant supplier vs. defense prime — the integration boundary in directed energy

A defining structural tension in directed energy: laser source companies (nLIGHT, IPG) supply the laser "engine" to defense primes (Lockheed Martin, RTX/Raytheon, Northrop Grumman, Boeing) who integrate it into weapon systems (beam directors, thermal management, target tracking, atmospheric propagation — none of which are component-level competencies). The contested question is whether the source makers move up the value chain into prime territory:

• LASR/nLIGHT is both a merchant supplier (sold the 50 kW module for DE M-SHORAD) and increasingly a full weapon-system maker (70 kW, 30 kW, 10 kW HEL products it markets directly to DOD and allied governments). The analysis flags the inherent channel conflict — "analogous to how Intel both sells chips to PC makers and makes its own PCs." Whether nLIGHT can win prime-contractor positions against Lockheed and Raytheon is the key strategic question for the segment.
• IPG plays the traditional supplier role — providing Crossbow laser modules to Lockheed Martin for integration into Lockheed's Sanctum C-UAS architecture (a ~$10M order in Feb 2026), not pursuing prime positions.

The architecture competition itself is a value-chain fork: Lockheed's spectral beam combining (SBC, passive, proven at 300 kW) vs. nLIGHT's coherent beam combining (CBC, higher theoretical brightness, targeting 1 MW under the $171M HELSI Phase 2 contract). The HELSI Phase 2 outcome will influence which architecture DOD standardizes for production laser weapons — and therefore which suppliers capture the production margin pool.

Cost-per-shot — the economics driving the directed-energy chain

The demand pull at the end of the laser value chain is a cost-exchange argument. An interceptor missile costs $10,000-$100,000+; a laser shot costs pennies to a few dollars of electricity. Against mass drone swarms ($500-$5,000 one-way attack drones), missile interceptors are economically unsustainable, while lasers offer infinite magazine depth, speed-of-light engagement, and graduated response. The Army expects each E-HEL laser (~$4.17M per unit) to engage swarms of up to 30 small drones, rockets, or mortars at up to 4 km within 1 minute — roughly $30 of electricity to defeat $15,000-$150,000 of drones. This is the killer-app logic that converts directed energy from a perpetual R&D curiosity into a production value chain.

The drone / ISR value chain — platform vs. subsystem

Tactical ISR drones show the same platform-vs-component split. At the platform end sit pure-play manufacturers (Red Cat / RCAT — Teal 2 with FLIR Hadron 640R EO/IR sensor, Blue UAS-certified) and the diversified primes (BAE, L3Harris, RTX) who integrate UAS into broader C4ISR ecosystems. At the subsystem/component end sit specialists like NextVision (NXSN — micro stabilized E/O systems) and the sensor/payload suppliers (EO/IR, LIDAR, SAR). Key value-chain levers: Blue UAS certification is the gating credential for U.S. government procurement (replacing Chinese-made small UAS), and the analysis flags that "limited competition and domestic suppliers drive margins" — the same supply-constrained, qualification-gated dynamic as the rest of the sector. The open commoditization question (raised on the defense-technology hub): if FPV drones cost $500, the margin is not in the airframe — it migrates to the autonomy/AI layer and the stabilized sensor payloads on top.

Contract economics — where pricing power lives in the chain

Two contract structures govern margin behavior. Cost-plus contracts (government services, some development work) cap upside but de-risk the supplier — VVX's 94% prime-contract, cost-plus services model is the archetype, with structurally thin margins. Fixed-price contracts carry execution risk: CVU's A-10 program was a 2019 fixed-price deal whose higher-than-planned manufacturing costs forced $8.1M in EAC adjustments on a retiring platform — the live demonstration that fixed-price exposure can wreck a quarter's margins. Sole-source production contracts on long-lived platforms are where durable pricing power actually sits (ESP's submarine transformers, MPTI's 40+ programs), because the qualification barrier eliminates competitive bidding pressure. The capital cost of operating in these tiers is correspondingly modeled at a 12-16% WACC for micro-cap suppliers (4.5% risk-free + 5.5% ERP + 3-5% size premium + 0-3% concentration/leverage premium), with betas typically 0.8-1.2.

(Source note: the defense-tech-monitor-2026-03-01.md file contained only frontmatter — no value-chain content to extract. All listed source paths existed.)

Players

The investable defense-tech universe in this vault skews small. Deliberately. The thesis is that the most interesting risk/reward sits not in the $100B+ primes (Lockheed, RTX, Northrop, General Dynamics, Boeing — the five that survived the 1993 "Last Supper" and now split most of the ~$632B in global arms sales) but in the Tier 2/3 suppliers with sole-source positions, multi-year backlogs, and designed-in moats. The primes don't vertically integrate everything: complexity management, specialization economics, and 3-5 year MIL-SPEC qualification cycles (MIL-PRF-3098, MIL-S-901D Grade A, MIL-STD-461, AS9100, ITAR, CMMC) all push work down to specialists. Once a component is designed into a platform, it's embedded for the platform's life. That dynamic — capacity-constrained supply, not demand — is what gives these names pricing power and visibility.

Coverage is organized below by sub-sector. Company-specific deep detail lives on the ticker pages; this is positioning plus the comparative view. The five head-to-head analyses worth reading in full: feim-vs-lasr-showdown (FEIM vs LASR), MPTI-vs-CODA-vs-ESP-and-more-showdown (MPTI vs CODA vs ESP vs CVU), four-defense-stocks-ranked (Leonardo vs IHI vs Ondas vs LightPath), defense-tech-comparison (EVLV, VVX, LDO, IDR, 7013, ONDS), and defense-tech-peer-comparison (LPTH, MPTI, MRCY, ONDS, PNG, VVX, OPTT).

Small-cap defense components and electronics

This is the core of the vault's defense-tech franchise — the Tier 2/3 suppliers where a $500M+ backlog can trade at a sub-$100M enterprise value.

• MPTI — M-tron Industries. Precision frequency control: crystal oscillators (OCXOs, TCXOs, DOCXOs), RF filters, and microwave assemblies from 1 MHz to 90 GHz, founded 1965. The metronome inside radar, GPS-denied navigation, EW, and secure comms. Sole-source on 40+ long-term defense/space programs running through 2030-2035+. The best profitability profile among the small-cap peers (46% gross, ~19% op margin, positive FCF, ~35% ROIC, debt-free) — yet trades at the lowest EV/Revenue (3.4x) of the tech names. Premium valuation on a growth-adjusted basis (PEG ~4x with decelerating growth and 2025 margin compression); the showdowns rate it WATCH, not buy. CEO hired specifically to execute the first M&A deal.

• ESP — Espey Mfg. & Electronics. MIL-SPEC ruggedized power supplies, transformers, and converters (12V-45,000V, 20W-3.8MW) for severe-environment military use. Sole-source on distribution transformers/power-distribution panels for Columbia-class submarines (the Navy's #1 acquisition priority, a ~$130B program building 12 boats through 2042), plus Virginia-class, DDG-51 AMDR radar, E-2D and F-18 radar power. The "sleep-well-at-night" play: zero debt 5+ years, $17.8-22M cash, ~27% FCF margin, $134-141M backlog (~3.1-3.2x revenue), gross margins that expanded from 12.1% (FY21) to ~29-35% — clear pricing power. Cheapest in its peer group (~12x forward P/E, PEG ~0.3-0.5x) with the deepest visibility. The chairman bought $17M+ on the open market. Rated the top pick in the MPTI/CODA/ESP/CVU showdown (score 4.21, BUY) and a BUY in the five-stock DCF analysis (PT $62). Single-facility risk in Saratoga Springs is the structural flaw.

• CVU — CPI Aerostructures. Structural assemblies — pod structures, wing assemblies, control surfaces — for military aircraft and EW pods. The crown jewel is NGJ-MB (Next Generation Jammer Mid-Band) pod structures for Raytheon's EA-18G Growler program ("the most valuable program in our 40-year history," 300+ pods, IOC Dec 2024), plus single-source Raytheon missile wing assemblies, NGJ Low-Band for L3Harris, and legacy T-38/MH-60/F-16 work. The valuation anomaly: ~$509M total backlog ($100M funded) against a ~$70M EV — 0.13x EV/Backlog. Weakest moat of the group (build-to-print, no proprietary tech, 18-22% gross margins) plus an SEC restatement scar and a Q2 2025 material-weakness flag. A turnaround bet, not a compounder — the A-10 program termination ($8.1M EAC charge) depressed earnings; Q3 2025 was the first clean quarter at 22.3% gross margin. Rated SPECULATIVE BUY (DCF PT $6.50, 50%+ upside).

• CODA — Coda Octopus Group. The world's only real-time 3D/4D/5D/6D volumetric imaging sonar (Echoscope: 16,000+ soundings per transmission, 12-20 Hz refresh). DAVD (Diver Augmented Vision Display) head-up displays under an exclusive Navy license from NSWC Panama City; the new Echoscope PIPE NANO GEN (smartphone-sized) targets the projected ~$11B UUV/AUV market. A genuine technology monopoly — 66-70% gross margins, 45% insider ownership, $26-29M cash exceeding annual revenue, zero debt. The longest growth runway in the small-cap group (greenfield AUV perception market) but mediocre ROIC (~7-13%) from an undeployed cash pile, and a 38x trailing P/E near highs. Rated WATCH — wait for a post-earnings pullback. Alliance Global Partners PT $11.

• MOBX — Mobix Labs. Real defense exposure (Navy vessels, M-1 Abrams EMI connectors via Honeywell, Tomahawk, Javelin, Patriot via Spacecraft Components, F-35, military drones; BAE Gold Tier Supplier), but a precarious financial structure: ~$3.1M cash against ~$8.5M annual adjusted operating loss, ~$42M TTM net loss, 69% share dilution in a year, penny-stock price (~$0.22) and Nasdaq bid-price compliance concerns. The deep dive's verdict: defense franchise is real, equity is not investable for most; bankruptcy probability >50% within 18 months. AVOID. (No ticker page in vault.)

• IEHC, CHG, AXON — adjacent ticker pages. Covered as standalone canonical pages in the vault (IEH Corp hyperboloid connectors; Charge Enterprises; Axon Enterprise) rather than profiled inside the showdown documents. See the ticker pages directly.

The screening criteria the deep dive distills: sole-source/designed-in positions, 2x+ backlog coverage, zero/low debt, production-stage (not development) programs, margin-expansion runway, and multiple prime relationships (customer concentration is the dominant risk). The summary comparison:

Company	EV	Backlog	EV/Backlog	Gross Margin	Debt	Moat
CVU	~$70M	$509M	0.13x	20-22%	$16M	Single-source missile/pods
MPTI	~$212M	$59M	3.6x	43-46%	$0	40+ sole-source programs
ESP	~$75M	$141M	0.53x	35%	$0	Submarine power sole-source
CODA	~$82M	N/A	—	68-70%	$0	Only real-time 3D sonar
MOBX	~$27M	N/A	—	50-60%	High	Growing but fragile

Five-stock risk/reward ranking from the DCF analysis (prices as of the report): ESP ($47.32, PT $62, BUY, Low risk) > CVU ($4.30, PT $6.50, BUY, Medium) > MPTI ($64.75, PT $68, HOLD, Low-Med) > CODA ($9.85, PT $11, SPECULATIVE BUY, Medium-High) > MOBX ($0.22, AVOID, Extreme).

Directed energy and precision timing

The two opposite poles of the defense-photonics theme — one is the cheap, mission-critical heartbeat; the other is the explosive growth story. See feim-vs-lasr-showdown for the full head-to-head.

• FEIM — Frequency Electronics. Precision timing since 1961 — crystal/oven-controlled oscillators, rubidium atomic clocks, hydrogen masers, secure GPS receivers. The smallest, cheapest component in a satellite or weapons system, and the one that makes everything else work; 120+ space programs over 63 years, ~94% US government revenue. Defensive (beta 0.31), GAAP-profitable (37.9% gross, ~10% op margin, 10% ROIC), $83M record backlog with $45M in fresh satellite awards. Risks: founder Martin Bloch (~93, still Exec Chairman/CSO) key-person exposure, lumpy quarters, a precarious cash position ($86K at one quarter-end). The showdown's pick of the two — BUY at ~$40 (weighted score 3.46), the better risk/reward after a 34% pullback. TURbO miniaturized atomic clock (~$20M/yr from FY27, drone-fleet timing) and Golden Dome are the catalysts.

• LASR — nLIGHT. Vertically integrated high-power fiber laser company that has pivoted hard from commoditizing industrial cutting/welding into defense directed energy — semiconductor pump diodes (own GaAs fab in Vancouver, WA), LIEKKI specialty fiber, fiber lasers, and complete laser weapon systems via coherent beam combining (CBC) and adaptive optics (2019 Nutronics acquisition). Flagship programs: HELSI Phase 2 ($171M, scaling CBC to 1 megawatt-class, completion late 2026) and DE M-SHORAD ($34.5M, 50kW counter-drone on Stryker, four deployed to CENTCOM), plus 70/30/10 kW HEL products. $261.3M revenue (2025, +31.6%), A&D up 60% to 67% of mix; net cash after a $190-201M raise. Never posted a profitable year in 25 years; 11-14% SBC, 69% dilution since IPO, insiders selling. The larger TAM ($5-10B+ directed energy by 2030 vs ~$2-3B precision timing) but at 14x sales / 203x forward P/E it prices in a future that hasn't arrived — the showdown rates it WATCH (score 2.96), wait for a pullback to $45-50. Beta 2.34.

The broader fiber-laser/directed-energy value chain (see fiber-laser-directed-energy-primer for the deep version) ranks the merchant laser suppliers and primes:

• IPG Photonics (IPGP) — the original fiber-laser giant, most vertically integrated (grows crystals, draws fiber, makes diodes), ~$1.0B revenue but in structural decline from ~$1.4B (2018) as Chinese commoditization (Raycus, MaxPhotonics) erodes the industrial moat. Belated defense entry via the Crossbow counter-UAS laser ($10M Lockheed order, tiny). Fighting a two-front war.
• Coherent (COHR) — owns Nufern, the largest independent specialty-fiber maker (a supply-chain chokepoint for anyone not named IPG or nLIGHT); makes pump diodes; center of gravity has shifted to AI/datacom optics.
• Lumentum (LITE) — increasingly a telecom/AI play (NVIDIA's $2B March 2026 investment dwarfs any laser opportunity); minor directed-energy participant.
• TRUMPF (private, ~17% share, #2) and Jenoptik (JEN) round out the Western set; Raycus / MaxPhotonics / Han's Laser / JPT (China) own the commoditizing industrial end but do not yet compete in defense.
• Defense primes in directed energy: Lockheed Martin (spectral beam combining, SBC; 300kW delivered, scaling to 500kW; HELIOS shipboard 60kW on USS Preble; IFPC-HEL; Sanctum C-UAS integrating IPG Crossbow), RTX/Raytheon (HELWS), Northrop Grumman (Phantom 10kW <200lbs; DARPA MELT), Boeing (CLWS 5kW; 300kW HELWS with General Atomics). The strategic question is SBC (Lockheed, passive/robust) vs CBC (nLIGHT, higher theoretical brightness, targeting 1MW) — HELSI Phase 2 decides which architecture the DoD standardizes. International benchmarks: Israel's Iron Beam (Rafael/Elbit, first operational HEL air-defense system, delivered Dec 2025, up to 10km).

Drones, autonomous systems, and counter-UAS

The "War From Home" thesis (src-citrini-defense-primer-wfh) — remote warfare is structural, not cyclical; $500 FPV drones beat million-dollar missiles, so the economics flip toward cheap autonomy and toward lasers/counter-UAS for defense. See drone-isr-analysis for the tactical-ISR platform review and utr-drone-warfare-deep-dive for the technology landscape.

• ONDS — Ondas Holdings. Defense-focused autonomous systems: Optimus (first FAA Type-Certified autonomous drone — a real regulatory moat), Raider counter-UAS, Roboteam tactical UGVs, FullMAX private wireless. The most transformative growth story and the worst risk/reward in the vault's view. Guiding $170-180M revenue for FY2026 off a ~$48-49M base (~3.5x), almost entirely M&A-dependent (five 2025 acquisitions to integrate), against a ~$4B market cap (~90x trailing sales) and deeply negative EBITDA. The disqualifiers: 355% share dilution in 14 months (93M → 424M shares, 121.6M warrants at $28); the Ondas Networks segment, once core, generates ~$82K/quarter; and every analyst covering it served as a placement agent for its capital raises — making the unanimous "Strong Buy" worthless as validation. The huge gap between consensus (~$65M) and management guidance ($170-180M) is the tell. four-defense-stocks-ranked rates it AVOID (rank 4 of 4); a venture-stage business at a mid-cap valuation. $1.5B cash war chest is the only real asset.

• PNG — Kraken Robotics (Canadian, TSXV). Synthetic aperture sonar, SeaPower subsea batteries (US Navy-certified), underwater robotics; ~70% defense revenue. Highest gross margin in the peer-comparison group (59%), 50%+ growth, net cash, strong Anduril partnership. But extreme valuation (~120x EV/EBITDA, ~143x P/E) plus CAD/TSXV currency and liquidity risk for US investors. (No ticker page; covered in defense-tech-peer-comparison.)

• Tactical ISR platforms (drone-isr-analysis): Red Cat Holdings (RCAT) — small rugged Blue-UAS-certified ISR drones (Teal 2 / FLIR Hadron 640R), pure-play, riding DoD's replacement of Chinese-made small UAS; NextVision (NXSN) — micro stabilized E/O subsystem supplier; plus BAE, L3Harris (C4ISR), and RTX as the diversified primes. Private disruptors named across the sources: Anduril (autonomous systems; also a PNG partner), Shield AI and Skydio (military drones, neither FAA Type-Certified), and Palantir (PLTR) at the analytics/C2 layer (flagged as uncovered in the vault).

Security screening and broad defense/aerospace

• EVLV — Evolv Technology. AI-powered weapons detection / security screening for venues, schools, stadiums; SaaS model (~85% of ARR), ~70% share in AI weapons screening for venues. Highest gross margin (53.7%) and fastest growth (41% YoY) of the defense-tech-comparison set, $117M ARR, but deeply unprofitable (-42% op margin, negative FCF) with a restatement-credibility scar. The more-credible of the two speculative names there (vs ONDS) — positive adjusted EBITDA, real recurring revenue.

• VVX — V2X. Defense mission-support services (logistics, aerospace maintenance, training, IT) across 47 countries, ~16,000 employees. The cheapest name across both multi-company comparisons — 0.66x EV/Revenue, ~9x EV/EBITDA, ~12.5x P/E vs sector medians far higher — with an $11-12.5B backlog (~2.5x revenue) and rapid deleveraging (3.7x → 2.2x). The catch is the cost-plus services model: 8-9% gross margin caps upside, and growth is low-single-digit. A "value compounder" rather than a tech play. (No ticker page; in defense-tech-comparison and defense-tech-peer-comparison.)

• MRCY — Mercury Systems. Mid-cap defense-electronics processing (RF/EW through mission computing) across 300+ programs, SOSA open-architecture leader, ITAR barriers, record ~$1.5B backlog. A real margin-recovery turnaround (28% gross vs 40%+ target) but largely priced in — stock +135% in 12 months, ~89x forward P/E, above consensus PT. Risk/reward less compelling at current levels. (No ticker page; in defense-tech-peer-comparison.)

• LPTH — LightPath Technologies. Vertically integrated IR optics, camera assemblies, and proprietary BlackDiamond chalcogenide glass — transformed by the G5 Infrared (2025) and AMI (2026) acquisitions into a defense IR-imaging platform. The strongest small-cap catalyst in the vault: FY2026 NDAA Section 834 legislatively mandates eliminating US defense reliance on Chinese/Russian optical glass by Jan 1 2030, and BlackDiamond is one of very few NDAA-compliant germanium replacements (reinforced by China's own germanium export curbs). Revenue accelerating ($32M → $65M+ run-rate, +120% Q2 YoY), adjusted EBITDA just turned positive, gross margin 27% → 37%, $97.8M backlog. But at ~9-10x NTM EV/Sales (vs 2-4x for profitable optics peers) it demands near-perfect execution, with a $200M shelf overhang. four-defense-stocks-ranked rates it Selective Buy (rank 2) — a measured 2-3% speculative position. (No ticker page; in defense-tech-peer-comparison and the four-stock ranking.)

The large-cap diversified primes appear primarily as benchmarks rather than coverage targets:

• Leonardo (LDO) — the vault's highest-conviction defense name in four-defense-stocks-ranked (rank 1, Overweight). Cheapest major European defense stock (21-24x forward P/E, PEG <1.0) vs Rheinmetall at 39x and Saab at 46x, €44B+ backlog (2.5 years coverage), FOCF quadrupled €209M → €826M, net debt/EBITDA 3.1x → ~2.0x, riding the European rearmament supercycle (NATO Hague 3.5% GDP target; EU defense budgets toward €800B by 2030). Italian government's 30% stake is the structural discount.
• IHI Corporation (7013) — Japanese aero-engine/defense conglomerate (Aero Engine, Space & Defense = 85.5% of operating profit on 34% of revenue), GCAP partner, Japan's #1 jet-engine maker (~70% share). Best business quality in the four-stock ranking but trading 10-30% above analyst targets (PEG 2.95) — rated Neutral/Wait. Japan's FY2026 defense budget hit ¥9.04T (14th consecutive record), already past 2% of GDP two years early.
• Indra Sistemas (IDR) — Spain's national defense champion (radar, C3I), the standout performer (stock tripled in 52 weeks, +675% Q4 defense order growth) — but the re-rating has largely happened (~21x EV/EBITDA).
• OPTT — Ocean Power Technologies — wave-energy/autonomous-marine with tangential defense exposure; the most speculative name in defense-tech-peer-comparison (Altman Z -11.67, going-concern risk).

Defense-mineral and nuclear adjacency

• UAMY — US Antimony. Sits at the defense/critical-minerals intersection: antimony for ammunition, ordnance, and flame retardants; only Western smelter, hence $597M+ DLA strategic-stockpile contracts. The proof that the US government treats mineral supply as national security. See critical-minerals.
• NOC-uranium-companies-northrop — Northrop Grumman plus uranium miners (nuclear subs, carriers, propulsion); nuclear as both energy and defense. See BWXT (naval reactors) and nuclear-smr.
• MDA — MDA Space — space robotics / satellite servicing at the defense/commercial boundary. See MDA-space-thesis.
• PKE (Park Aerospace) — flagged as a sole-supplier of PAC-3 ablative composites (Mike/Multibagger Monitor defense thesis); a defense supply-chain name to track. The Meridian Report covers the rare-earths/ablative-composites/critical-minerals defense supply chain (METC, AREC).

Coverage gaps the hub itself flags: no Palantir, no L3Harris/Raytheon/Lockheed standalone analysis, no counter-UAS market sizing as a discrete page, no defense-budget tracker, no cyber-defense coverage, and Ukraine/Taiwan scenario investment implications unmapped.

Monitor

This is the rolling log for defense-tech: dated catalysts, policy moves, earnings signals, and a standing watch-list. Company-specific detail lives on ticker pages — this section tracks the industry-level developments and the recurring framing that the dated briefings keep returning to. A dedicated dated monitor file (defense-tech-monitor-2026-03-01.md) exists but is a stub — frontmatter only, no body — so all of the dated content below is reconstructed from the comparison, primer, and deep-dive sources, which are themselves timestamped snapshots (most carry March 2026 data dates).

The standing thesis being monitored

Defense spending is structurally rising, not cyclically. Three legs hold it up and each generates its own stream of catalysts to watch:

1. The drone/counter-drone cost-exchange flip. Ukraine and the Middle East proved you cannot spend $100K per interceptor against $500 one-way attack drones. This is the macro that pulls directed energy, counter-UAS, and tactical ISR from prototype into procurement.
2. European and Japanese rearmament supercycle. Decades of underinvestment (European spending fell from ~2.75% of GDP in the Cold War) created production gaps that take years to close — which means multi-year demand visibility and pricing power for qualified suppliers.
3. US budget at record levels. FY2026 total national defense spending exceeds $1 trillion for the first time.

The open questions that recur across the briefings: is this structural or politically volatile (a ceasefire narrative triggered a 5-10% snap correction in EU defense in August 2025); does FPV-drone commoditization leave any margin in the hardware vs. the autonomy layer; and will the directed-energy timeline slip again as it has for decades.

Budget and policy moves (the macro catalyst stack)

United States — FY2026 / FY2027:

• DoD budget request hit $892.6 billion, plus a $150 billion reconciliation package (including $20.4 billion for munitions and supply chain). Total national defense spending exceeds $1 trillion in FY2026.
• RDT&E allocation of $179 billion, a 27% YoY increase — one of the largest R&D allocations in defense history.
• Navy shipbuilding budget jumped to $47.3 billion in FY2026, up from ~$28B in 2024. Columbia-class is the DoD's #1 acquisition priority ($130B program, 12 submarines through 2042).
• Army targets 100,000 rounds/month of 155mm artillery; Patriot GEM-T backlog stretches ~5 years.
• Trump administration's proposed $1.5 trillion FY2027 defense budget; Defense Secretary Hegseth's stated goal of 300,000+ drones.
• Pentagon directive (March 2026): field directed-energy weapons at scale within 36 months. This is the single most cited near-term policy catalyst across the laser/DE sources.

NDAA — the legislated catalyst:

• Section 834 of the FY2026 NDAA legislatively mandates eliminating US defense reliance on optical glass from China and Russia by January 1, 2030. This is law, not a "might-happen" — directly drives the germanium-replacement IR-optics thesis (see LPTH BlackDiamond). Reinforced by China's own germanium export restrictions.

Europe:

• NATO Hague Summit endorsed a 3.5% GDP defense spending target (up from 2%). European defense budgets projected to reach €800B by 2030 — €300B more than 2025. Rule of thumb tracked by LDO: each 1% GDP increase in EU defense spending = ~€6B additional business for Leonardo.

Japan:

• FY2026 defense budget hit ¥9.04 trillion (~$58B), a 9.4% increase and the 14th consecutive record. Total spending has already exceeded 2% of GDP two years ahead of schedule; five-year Defense Buildup Program allocates ¥43 trillion through FY2027. Drives 7013 (IHI) via GCAP and fighter engine components.

Directed-energy / laser program calendar (watch these milestones)

This is the densest stream of dated catalysts and the area where program timing moves stocks. Source: the fiber-laser/directed-energy primer.

• E-HEL (Enduring High Energy Laser) — the Army's planned first directed-energy program of record. RFI issued October 30, 2025; competitive source selection planned for Q2 FY2026. Each ~$4.17M laser expected to engage swarms of up to 30 small drones/rockets/mortars at up to 4 km within 1 minute. First test of whether DE moves to scale production. Watch the down-select.
• HELSI Phase 2 — $171M nLIGHT contract (expanded from $86M) scaling coherent-beam-combining to megawatt class; expected completion late 2026. Lockheed scaling its spectral-beam-combining architecture to 500 kW+. The SBC-vs-CBC architecture competition determines which standard the DoD productionizes. Any successful demo is a LASR catalyst.
• DE M-SHORAD / "Guardian" — 50 kW laser on Stryker; live-fire at Fort Sill (June 2025); four systems deployed to CENTCOM; nLIGHT delivered the 50 kW module Q4 2025. Continued deployments are the proof-of-concept watch-item.
• IFPC-HEL — Lockheed prototype shipping to Dugway Proving Ground, UT for developmental testing March–summer 2026. Will NOT field operationally — feeds the next-gen Joint Laser Warfighting System (JLWS), which folds into the Trump "Golden Dome" missile-defense initiative.
• Navy — HELIOS (60 kW on USS Preble) reportedly downed Iranian drones early 2026; ODIN dazzler used in Operation Epic Fury (February 2026); Songbow is the next-gen Navy DE effort.
• DARPA MELT — $8M Phase 2 to Northrop (April 2025); optional Phase 3 (2027+) assembles a nine-tile panel array.
• International proof points: Israel's Iron Beam (Rafael/Elbit) delivered to the IDF December 2025 — the world's first operational high-energy laser air-defense system (to 10 km). Allied demand pipeline: Australia, UK, Saudi Arabia, Japan. nLIGHT showcased a 70 kW LWS at the World Defense Show (Riyadh, Feb 2026) and the POST Conference (Honolulu, March 2026).
• Merchant-supplier signals: IPG received a ~$10M Crossbow order from Lockheed Martin (February 2026) — tiny vs. its ~$1B revenue but a new vector; IPG lost a UPC patent ruling to TRUMPF (March 2026, <1% of sales, appealing). NVIDIA invested $2B in LITE (March 2026) — confirms Lumentum's pivot away from defense/industrial lasers toward AI optical networking.

Earnings and contract signals logged (the dated tape)

These are the company-level data points the comparison/showdown sources captured around their March 2026 data dates — kept here because they read as the dated earnings/contract tape for the sector. Company detail belongs on ticker pages; this is the chronological signal log.

• FEIM — Q3 FY2026 earnings released March 11, 2026; revenue softness drove a 34% pullback from the January high of $61.47. Offsetting: $45M in fresh satellite contract awards (March 2026) and an $83M record backlog, with management guiding awards of "similar magnitude" later in 2026. Insiders buying (4 buys, 0 sells in the past year). Craig-Hallum initiated Buy (Sep 2025); Freedom Capital downgraded to Hold (Dec 2025).
• LASR (nLIGHT) — record 2025: $261.3M revenue (+31.6%), A&D revenue $175.3M (+60%). Officially exited industrial cutting/welding in Q4 2025 (~$25-30M 2026 headwind, margin-accretive). $201M / $190M+ equity offering February 2026 (~$133.6M cash). Q1 FY2026 guide $70-76M. CEO insider selling ($20M+). Next pass/fail test flagged: Q1 FY2026 earnings, May 7. Baird initiated at $95 (March 2026).
• IDR (Indra) — defense order growth of +675% in Q4 2025; 2026 revenue guided >€7.0B (+28%). Stock tripled in 52 weeks (€19 → €63).
• LDO (Leonardo) — €23.8B new orders (+14.5%), book-to-bill 1.3x H1 2025, FY2025 orders guided €22.25-22.75B; FOCF guided €920-980M (quadrupled from €209M in FY2022). Net debt/EBITDA collapsed 3.1x → ~2.0x heading toward 0.6x. FY2025 results due early March 2026. PEG <1.0 — flagged the cheapest major EU defense name.
• ONDS (Ondas) — preliminary 2025 revenue ~$48-49M; 2026 guidance $170-180M (raised from $110M), vs. consensus ~$65M — a 2.7x discrepancy that is itself a credibility watch-item. Shares exploded 93M → 424M in 14 months (355% dilution) plus 121.6M warrants at $28. $1.5B cash war chest. Every covering analyst served as a placement agent — the unanimous Strong Buy is non-independent.
• MRCY (Mercury) — record $1.5B backlog; stock +135% in 12 months, trading above the $80 consensus PT; forward P/E ~89.5x.
• LPTH (LightPath) — Q2 FY2026 revenue +120% YoY; adjusted EBITDA turned positive Q1/Q2 FY2026 ($0.4M, $0.6M); gross margin 27% → 37%; $97.8M backlog; G5 camera business booked $80M+ orders in year one; $200M shelf registration is the dilution overhang.
• Micro-cap earnings dates that were imminent at the March 2026 data date: CODA Q1 FY2026 earnings March 17, 2026 (stock extended, 38x trailing P/E near highs); MPTI Q4 FY2025 earnings March 26, 2026 (margin-recovery data is the swing factor — Q4 2025 results were "pending March 26").
• ESP (Espey) — Q1 FY26 backlog $141.1M (+49% YoY, ~3.2x revenue); declared a $0.75/share special dividend September 2025; chairman bought $17M+ on the open market — the standout insider signal in the micro-cap group. Columbia-class contracts: $29.5M (Dec 2024) + $19.8M (April 2025), deliveries through 2030.
• CVU (CPI Aerostructures) — $508.96M total backlog ($100.05M funded) vs. ~$70M EV (0.13x EV/backlog); NGJ-MB Lot 5 awarded November 2025 ($21.1M initial, $42.3M NTE); single-source Raytheon missile-wing award (October 2025) sent the stock +59-96%; A-10 termination created $8.1M EAC charges, but Q3 2025 (first clean quarter) showed 22.3% gross margin.
• MPTI (M-tron) contract wins: $20M air-defense program (Dec 2025, production through 2030+), $5.5M naval weapons (June 2025, through 2035+), multi-domain data-link program (Sept 2025).
• MOBX (Mobix) — FY2025 prelim revenue $9.7-9.9M (+54%) but ~$3.1M cash against ~$8.5M annual operating loss; 69% share dilution in a year; Nasdaq bid-price compliance concern (180-day extension). Spacecraft Components acquisition (Nov 2024).

Standing watch-items checklist

Carry these forward — they are the recurring metrics and decision points the sources keep flagging:

• Directed energy: E-HEL source-selection outcome (Q2 FY2026); HELSI Phase 2 megawatt demo (late 2026); whether the 36-month "at scale" Pentagon timeline holds or slips again. SBC vs. CBC architecture decision.
• Counter-UAS: does the killer-app economics (≈$30 of electricity to defeat $15K-$150K of drones) convert prototypes into production contracts. Watch Crossbow, HELWS, Iron Beam adoption.
• Book-to-bill as the single best small-cap health signal — flagged explicitly for MPTI (stay above 1.0x) and ESP.
• ONDS guidance gap: consensus $65M vs. management $170-180M for 2026 — the cleanest credibility test in the sector; watch whether the five-acquisition integration delivers.
• Margin normalization: CVU post-A-10 (target 20%+ gross), MRCY turnaround (28% gross vs. 40%+ target), LPTH path to GAAP profitability.
• Backlog-to-EV disconnects: CVU at 0.13x and ESP at ~0.53x are the embedded-value watch-items.
• EU re-rating exhaustion: IDR and LDO have already tripled — watch for ceasefire-narrative snap corrections (precedent: August 2025).
• Japan timing: 7013 (IHI) trades above virtually every analyst PT (¥4,200 vs. consensus ¥3,200-3,800); PEG 2.95. Wait-for-pullback name; PW1100G inspection liability (~350 aircraft grounded through FY2026) and ¥1.9B op-profit-per-¥1 FX sensitivity are the drags to monitor.
• Chinese commoditization of industrial fiber lasers — the structural headwind for Western laser source companies; watch Raycus/MaxPhotonics moving up the power curve (now 12-20 kW). Defensible segments remain defense, medical, ultrafast.
• Supply-chain / critical-minerals feed: germanium and antimony export restrictions; UAMY DLA contracts as the proof that the US treats mineral supply as national security. The Meridian Report (★★ paid, Steve) covers rare earths / ablative composites / defense supply-chain bottlenecks; Multibagger Monitor (★★ paid, Mike/@BlackSchol) covers the PAC-3 ablative-composites supply chain (PKE as sole supplier).

Coverage gaps to close

Flagged as missing in the topic hub and worth queueing for future monitoring: Palantir (largest defense-AI name), large-cap primes (L3Harris, Raytheon, Lockheed, Northrop), a quantified counter-UAS market sizing, a standing defense-budget tracker (US/allied/global), cyber defense, and explicit Ukraine/Taiwan scenario investment mapping. The dated monitor file itself (defense-tech-monitor-2026-03-01.md) is an empty stub and should either be populated on the next monitor run or folded into this section.

Sources

The defense-tech research on this page draws on a mix of named newsletter/analyst sources with dedicated pages under _sources/, attributed third-party reports synced into the vault, and a long tail of company filings, IR pages, market-data terminals, and trade press cited inline. Below is the consolidated source inventory.

Source authors and publications (with vault pages)

• source-citrini — Citrini Research. The "War From Home" thesis is the conceptual spine of the whole sector: remote warfare (FPV drones, loitering munitions, autonomous targeting, cyber) is a structural shift in military doctrine, not a cyclical budget bump. Cited in the topic hub as the framing for whether defense exposure is structural vs. cyclical. The per-piece note appears in the wiki as src-citrini-defense-primer-wfh.
• source-meridian-report — The Meridian Report (themeridianreport.com), author "Steve," ★★ paid. Covers defense and critical minerals — rare earths, ablative composites, defense supply-chain bottlenecks. Cross-references UAMY, METC, AREC.
• source-mike-multibagger — Mike / Multibagger Monitor (Substack @huntingmultibaggers; X @Mike10947310 / @BlackSchol), ★★ paid. Microcap, event-driven, VC-style public-markets theses. Defense angle: PKE as sole supplier of PAC-3 ablative composites.

Source authors and publications (no vault page yet)

• Mawer — Joshua Samuel, CFA and Rob Campbell, CFA. "Investment Opportunities in a Changing Military Landscape" (Mawer podcast EP187, April 23, 2025) — the basis for the European-rearmament / historical-underinvestment thesis (European spending ~2.75% of GDP in the Cold War, post-Cold-War collapse, Germany's package, supply-chain capacity constraints creating pricing power). Source file: Maser Capital - Investment Opportunities in a Changing Military Landscape.docx. Note: the consolidated wiki file is named maser-military-landscape.md ("Maser" appears to be a typo for "Mawer").
• Defense Stocka. "Autonomy and Tactical ISR Drone Analysis" — comparative platform review (Red Cat / RCAT, NextVision / NXSN, BAE Systems, L3Harris, RTX). Source file: DR - Autonomy and Tactical ISR Drone Analysis.docx.
• Under The Radar (UTR). "Defense Primers: Glossary and Weapons Systems Reference" (MLRS/HIMARS/ATACMS, IFVs, USEUCOM, C4ISR). Source file: UTR - Defenses Primers.docx. The topic hub also references a broader UTR defense series — utr-drone-warfare-deep-dive, utr-primer-ballistic-missile-defense, utr-defense-sector-risks, utr-quantum-for-defense, utr-quantum-update-q-ctrl.

Key external references cited across the sources

Market-data and aggregator terminals (recurring across the peer-comparison files): StockAnalysis.com, Yahoo Finance, Investing.com, MacroTrends, Smartkarma, Seeking Alpha, StockTitan, Simply Wall St, TipRanks, Alpha Spread, MarketBeat, PR Newswire / GlobeNewswire, The Corner (thecorner.eu).

Sell-side / analyst coverage cited: Sidoti, HC Wainwright, Morgan Stanley, Craig-Hallum (FEIM Buy initiation, Sep 2025), Freedom Capital (FEIM downgrade to Hold, Dec 2025), Baird (LASR initiation at $95, Mar 2026).

Company filings and IR (SEC/SEDAR + investor-relations pages): Evolv Technology IR, V2X IR, Leonardo IR, Indra IR, IHI Corporation IR, Ondas Holdings IR, IPG Photonics IR, nLIGHT press releases.

Fiber-laser / directed-energy primer external references: RP Photonics (fiber lasers, double-clad fibers, spectral beam combining), Wikipedia (fiber laser), Tech Briefs (SBC vs. CBC), IPG Photonics "Fiber Lasers 101," nLIGHT (Directed Energy, Semiconductor Lasers, HELSI $171M, 2025 record revenue, 70kW LWS, Nutronics acquisition), Coherent Corp (Nufern specialty fibers), Lockheed Martin (300kW HELSI delivery, scaling to 500kW), US Army (300kW HELSI program, E-HEL RFI, 2026 counter-drone laser competition), Military Times and National Defense Magazine (Pentagon "laser weapons at scale within 36 months"), Army Recognition (DE M-SHORAD first operational use; Iron Beam delivery), Navy Times (HELIOS drone intercept), RTX/Raytheon, Northrop Grumman (DARPA MELT, Phantom laser), Boeing CLWS, General Atomics (Boeing 300kW HELWS), Raycus, Laser Focus World, DefenseScoop, Defense News, plus market-sizing houses Grand View Research, Fortune Business Insights, Precedence Research, Mordor Intelligence, IMARC, and two CRS reports (DOD Directed Energy Weapons R46925; Navy Shipboard Lasers R44175).

Small-cap deep-dive external references (inline web citations): Umbrex ("How the Defense Contractor Industry Works"), Brennan Inc (aerospace supply-chain tiers), TD Economics (US defense-spending impacts — $892.6B DoD request, $150B reconciliation, >$1T total FY2026), Stars and Stripes (Navy FY2026 shipbuilding budget $47.3B), Global X ETFs (defense technology / $179B RDT&E, +27% YoY), plus per-company aggregator citations (Seeking Alpha, StockAnalysis, Stock Titan, TipRanks, GitHub market-research lists).

Consolidated source files folded into this sector page

• defense-tech-comparison.md — 6-company compare (EVLV, VVX, LDO, IDR, 7013, ONDS), /compare skill, 2026-03-02/14.
• defense-tech-monitor-2026-03-01.md — /monitor skill stub, 2026-03-06 (frontmatter only; no body content).
• defense-tech-peer-comparison.md — 7-company peer compare (LPTH, MPTI, MRCY, ONDS, PNG, VVX, OPTT), /compare skill, 2026-03-02/14.
• defense-technology.md — the topic-hub page (thesis, tracked tickers, cross-domain links).
• drone-isr-analysis.md — Defense Stocka tactical-ISR drone review.
• _compare/feim-vs-lasr-showdown.md — FEIM vs LASR showdown, /compare skill, 2026-03-22.
• fiber-laser-directed-energy-primer.md — /primer skill deep-dive on fiber lasers and directed energy (IPGP, LASR, COHR, LITE), 2026-03-21.
• four-defense-stocks-ranked.md — Leonardo / IHI / Ondas / LightPath risk-reward ranking (claude-chat sync, 2026-04-05).
• maser-military-landscape.md — Mawer (Samuel/Campbell) European-defense thesis.
• _compare/MPTI-vs-CODA-vs-ESP-and-more-showdown.md — micro-cap showdown (MPTI, CODA, ESP, CVU), /compare skill, 2026-03-15.
• small-cap-defense-deep-dive.md — 94KB Tier 2/3 supplier deep dive (MPTI, CVU, ESP, CODA, MOBX), claude-chat sync, 2026-04-05.
• utr-defense-primers.md — Under The Radar weapons-systems glossary.
• _sources/source-meridian-report.md and _sources/source-mike-multibagger.md — the two source-author pages read for this consolidation.

For company-specific source trails (per-ticker filings, contract announcements, insider-trade records), see the individual ticker pages: ONDS, EVLV, MDA, UAMY, and the LASR deep-dive at ~/claude/output/stocks/LASR/lasr-deep-dive.md.

---

Consolidation queue (merged 2026-05-30 — section-scoped rebuild)

Industry-wide content folded in from these source files. They stay live pending Pink's archive confirm.

• [ ] defense-tech-comparison.md
• [ ] defense-tech-monitor-2026-03-01.md
• [ ] defense-tech-peer-comparison.md
• [ ] defense-technology.md
• [ ] drone-isr-analysis.md
• [ ] _compare/feim-vs-lasr-showdown.md
• [ ] fiber-laser-directed-energy-primer.md
• [ ] four-defense-stocks-ranked.md
• [ ] maser-military-landscape.md
• [ ] _compare/MPTI-vs-CODA-vs-ESP-and-more-showdown.md
• [ ] small-cap-defense-deep-dive.md
• [ ] utr-defense-primers.md
• [ ] _sources/source-meridian-report.md
• [ ] _sources/source-mike-multibagger.md