🔴 FLAGSHIP Defense Tech · Private Markets Published Apr 29, 2026 · Anduril · Shield AI · Saronic · Helsing

Autonomous Warfare Stack:
Defense Tech Q2 2026

$95B in private-market value has accumulated across four venture-backed defense-tech companies in under 36 months. The DoD budget crossed $1T. Autonomous systems got their first dedicated budget line. This is where the capital is actually going.

Private Market Cap
~$95B
DoD Autonomy Budget
$9.4B line
Anduril Army Contract
$20B / 10yr
Companies Covered
4 startups + 2 primes
⚡
00 / 06
Executive Summary
$95B in private-market value has accumulated across four venture-backed defense-tech companies (Anduril, Shield AI, Saronic, Helsing) in under 36 months — a structural shift, not a cycle.
Anduril is projecting $4.3B in 2026 revenue at a $60B valuation (14x forward multiple) and a $1B loss — it is priced like a SaaS compounder but built like a defense prime.
The Pentagon's FY2026 DoD budget crossed $1T for the first time; autonomous systems received a dedicated $9.4B line-item — the first time autonomy has its own budget in DoD history.
CCA Increment 1 is the clearest near-term revenue catalyst: Anduril's YFQ-44A and General Atomics' YFQ-42A are the only two platforms; Shield AI won the autonomy software layer for both — making the Anduril/Shield AI relationship simultaneously competitive and interdependent.
Replicator 1 underdelivered: the Pentagon aimed for "thousands" of autonomous systems by August 2025 and delivered "hundreds." Replicator 2 is now running under a new org (DAWG). The gap between aspiration and execution is where capital is actually flowing.
Helsing is the most exposed: a €12B valuation, an HX-2 that failed frontline tests in Ukraine, and a German procurement subject to parliamentary approval. The market is ignoring material execution risk.
📊
01 / 06
Vitals Table
Company HQ Valuation Revenue (2025E) 2026E Rev Rev Multiple Deployed?
Anduril Costa Mesa, CA $60B* $2.1B $4.3B 14x fwd ✅ Yes
Shield AI San Diego, CA $12.7B ~$300M $540M+ 23x fwd ✅ Yes
Saronic Austin, TX $9.25B ~$400M N/D ~23x 2025 ✅ Yes
Helsing Munich, DE ~$13.4B N/D N/D N/M ⚠️ Partial
Lockheed (LMT) Fort Worth, TX ~$112B mkt cap $67B (total) — — ✅ Yes
Northrop (NOC) Falls Church, VA ~$60B mkt cap $38B (total) — — ✅ Yes

*Anduril $60B valuation from March 2026 reported round (a16z + Thrive Capital, not yet closed as of April 2026). Secondary market implied valuation reached ~$95B in March 2026.

Founder / Investor Profiles
Company CEO / Founder Notable Investors Thesis
Anduril Palmer Luckey (Oculus), Brian Schimpf (CEO) Founders Fund ($1B check), a16z, Thrive Capital, General Catalyst Software-centric defense prime at commercial scale
Shield AI Ryan Tseng (co-founder), Gary Steele (CEO) Advent International, JPMorganChase SRI, a16z, Blackstone ($500M preferred) Hivemind as defense AI OS layer, hardware second
Saronic Dino Mavrookas (Navy SEAL, CEO) Kleiner Perkins, a16z, General Catalyst, 8VC, Lightspeed Autonomous shipbuilding from first principles
Helsing Torsten Reil (co-CEO), Gundbert Scherf Prima Materia (Daniel Ek), Lightspeed, Accel, Saab, General Catalyst European sovereign defense AI; democratic governments only
🏆
02 / 06
Contract Scoreboard
U.S. DOD Awards (2024–Q2 2026)
Award Recipient Value Date Status
U.S. Army Enterprise Contract (Lattice) Anduril Up to $20B / 10yr March 2026 Awarded
CCA Increment 1 — Fury platform (YFQ-44A) Anduril ~$800M+ 2024 Flight testing; production 2026
CCA Increment 1 — Autonomy software Shield AI (Hivemind) Undisclosed 2025 Awarded
IVAS takeover from Microsoft Anduril Up to $22B Feb 2025 Active development
USMC Counter-UAS (Lattice/C-UAS) Anduril $642M / 10yr 2025 Program of record
USMC Bolt-M loitering munitions (OPF-L Phase 2) Anduril $23.9M (600+ units) Jan 2026 Delivery begins Q1 2026
U.S. Navy Corsair USV production (OTA) Saronic $392M / through 2031 2025 Active; $197M obligated
DARPA Pulling Guard (semi-autonomous escort) Saronic Undisclosed Feb 2026 Active
Replicator Scorecard
Initiative Goal Actual (Q2 2026) Key Vendors
Replicator 1 (ADA2) "Thousands" by Aug 2025 "Hundreds" fielded Anduril (Lattice), Swarm Aero, L3Harris
Replicator 2 (C-sUAS) Counter-small-UAS, within 24 months Solicitation live; under DAWG TBD
CCA FY2026 Budget Increment 1 production + Increment 2 awards $804M allocated; production contract expected 2026 Anduril, GA-ASI

The honest Replicator read: The program delivered less than promised at Phase 1. DAWG is the restructured successor. The underlying demand signal — thousands of attritable autonomous systems — is real. The timeline is not. That gap is commercially significant: volume orders everyone is pricing in are lagged by 12–18 months minimum.

Allied & AUKUS Contracts
Award Recipient Value Geography
Bundeswehr HX-2 loitering munitions Helsing €269M (options to €1.46B) Germany
€900M Bundeswehr drone program (tri-vendor) Helsing (~€300M share) ~€300M Germany / NATO Lithuania
Ukraine HX-2 supply (Germany-funded) Helsing 4,000+ units contracted Ukraine (operational)
UK MoD Project TALOS force protection Anduril £17–21M UK
AUKUS / Australia Navy (Corsair ASV) Saronic Trials underway Australia
⚙️
03 / 06
Tech Stack Diff — What's Actually Defensible

The Autonomy Stack: Three Layers, Three Strategies

LAYER 3: MISSION AI / AUTONOMY PLATFORM Anduril: Lattice OS — C2 + sensor fusion + autonomous teaming Shield AI: Hivemind — AI pilot (GPS/comms-denied flight autonomy) Helsing: Altra + Centaur AI — ISR + air combat autonomy LAYER 2: HARDWARE PLATFORM Anduril: Full stack (Fury, Roadrunner-M, Bolt-M, Dive-LD, Ghost-X) Shield AI: V-BAT ISR drone, X-BAT (in development) Saronic: Full stack (Spyglass → Corsair → Marauder USV fleet) Helsing: HF-1, HX-2 loitering munitions; CA-1 Europa (2027 target) LAYER 1: MANUFACTURING / PRODUCTION Anduril: Arsenal-1 (5M sqft, Ohio; production July 2026), SRM factory (Mississippi) Saronic: Franklin, LA shipyard (300K sqft expansion), Port Alpha (planned) Helsing: Resilience Factory (southern Germany, 450 units/month HX-2, scaling to 2,500) Shield AI: Asset-light; manufactures V-BAT through partners
Lattice OS (Anduril)
Defensible at scale
What it is
Open-architecture AI command-and-control platform. Fuses data from sensors, drones, weapons across domains in real-time. Third parties can integrate via Lattice SDK. Selected by U.S. Space Force, U.S. Army ($20B), SOCOM.
Moat
Vendor lock-in through platform adoption. If a military integrates Lattice for C2, adding Anduril hardware is plug-and-play. The Army's $20B contract creates a procurement gravity well.
Risk
Lattice is an open standard — non-Anduril hardware could run on it. If Lattice becomes a government standard rather than an Anduril proprietary moat, the economics change.
Hivemind (Shield AI)
Defensible on EW-denied ops
What it is
AI autonomy software enabling aircraft to fly missions without GPS, comms, or human-in-loop. Modular: EdgeOS (real-time OS), Hivemind Pilot (flight autonomy), Hivemind Commander (mission planning), Hivemind Forge (dev tools).
Moat
V-BAT flew 200+ missions in Ukraine's most contested EW corridor. That's a training dataset and reputational moat nobody can replicate by writing a check. Hivemind Foundation Model trained on simulation + real operational data.
Risk
Depends on OEMs and primes integrating Hivemind — a software layer on someone else's platform. If Anduril builds its own autonomy stack at the Lattice layer, the relationship turns purely competitive.
Altra + Centaur AI (Helsing)
EU sovereign angle real — product unproven
What it is
Altra: reconnaissance-strike software suite linking drones, ISR assets, and C2. Centaur: AI combat pilot that flew the Saab Gripen E in a real beyond-visual-range dogfight (Project Beyond, June 2025) — the first publicly documented AI vs. human fighter engagement.
Moat
European sovereign AI with NATO/EU trust signal. Germany, UK, Nordic states cannot buy Anduril's Lattice for sovereign warfighting without U.S. political exposure. EU Rearm Europe (€800B) creates a captive market.
Vaporware risk
HX-2 performed only 5 successful strikes in 14 deployments during one Ukraine frontline test (Jan 2026). AI components promised in contract spec were not installed at delivery. That's a contract integrity issue, not a development timeline miss.
Saronic (Echelon C2)
Hardware moat — vertical integration
What it is
Echelon is Saronic's C2/mission planning platform. The differentiated asset isn't the software — it's vertical integration of shipbuilding + autonomy under one roof. Navy went from prototype to $392M production contract in under 12 months.
Moat
Software-defined shipbuilding where each generation is cheaper to build through reusable subassemblies and automated production. Already demonstrated 25% production efficiency gain between Marauder vessel 1 and vessel 2.
Risk
Hardware moats are more fragile than software moats. Saronic can be out-manufactured by shipbuilding incumbents (HII, Austal) if the Navy decides to retrofit legacy yards.
💰
04 / 06
Capital Structure — SaaS Multiples on Defense Hardware

Traditional defense primes trade at 1–2x revenue and 10–15x EBITDA with gross margins of 8–12% on hardware. SaaS comps trade at 10–30x revenue with 70–80% gross margins. Defense-tech startups are being priced as if they'll achieve software economics while building hardware at scale.

Company Valuation Revenue (2025E) Revenue Multiple Implied By Multiple
Anduril $60B $2.1B 14x fwd / 28.6x trailing Assumes SaaS-like software revenue ramp via Lattice
Shield AI $12.7B ~$300M 42x trailing Assumes Hivemind becomes autonomous warfare OS
Saronic $9.25B ~$400M ~23x 2025 Assumes Navy scales to hundreds of vessels
Helsing ~$13.4B N/D N/M Assumes EU rearmament → recurring contracts
Palantir (public comp) ~$170B $2.87B 59x Pure software, government + enterprise
Lockheed Martin (LMT) ~$112B $67B 1.7x Traditional prime

The bull case for these multiples rests on: (1) Lattice/Hivemind achieving recurring software revenue as deployed hardware updates convert to software licenses; (2) Arsenal-1 and Port Alpha dramatically compressing unit costs; (3) These companies capturing the growth of the $1T DoD budget while legacy primes grow at GDP.

The bear case: Anduril is forecasting $1B in losses on $4.3B in 2026 revenue. The path to profitability runs through Arsenal-1 scaling (July 2026, execution risk). Shield AI's Hivemind software is 30% of revenue — the other 70% is hardware. Helsing has no public financials and a production reliability crisis. Fixed-price commercial contracts expose startups to cost overruns legacy primes avoid through cost-plus structures.

Exit Math — What Each Needs to Justify Current Valuation at 2029 IPO
Company Valuation Revenue Needed @ 20x (2029) Plausible?
Anduril $60B $3B by 2029 ✅ Very likely — guiding $4.3B in 2026
Shield AI $12.7B $635M by 2029 ✅ Likely — guiding $540M+ in 2026
Saronic $9.25B $462M by 2029 ⚠️ Possible — depends on Navy scaling Marauder
Helsing ~$13.4B $670M by 2029 ❌ Unclear — no public revenue; HX-2 issues
🎯
05 / 06
Three Asymmetric Calls
Call 1 // Saronic
Saronic wins the next Replicator tranche — the market is underpricing it
The setup
Replicator 1 underdelivered on aerial platforms. Replicator 2 is C-sUAS. Replicator 3 (implicit) will almost certainly prioritize maritime. The Navy Secretary called the Saronic Corsair contract (prototype to production in under 12 months) "the new benchmark for rapid defense procurement."
The edge
Saronic's $392M Navy contract runs through 2031. Corsair production scaling to 2,000 units/year. Navy strategy explicitly targets a "Golden Fleet" of ~500 manned + unmanned vessels by 2045. The addressable market for large USVs alone is in the tens of billions. Software-defined shipbuilding already demonstrated 25% production efficiency gain between vessels 1 and 2.
What the market is doing wrong
Saronic at $9.25B on $400M estimated 2025 revenue is being discounted as a hardware company. That's wrong. Saronic's approach is what Anduril is doing with Arsenal-1, applied to maritime.
Call: Saronic announces a second large Navy production contract (Cipher or Marauder) within 18 months, re-rating valuation above $15B. Current price is cheap relative to the evidence.
Call 2 // Helsing
Helsing is overvalued — the HX-2 crisis is more serious than the market thinks
The setup
Helsing carries a €12B valuation. Its flagship product, the HX-2, hit Ukrainian frontlines in late 2025. An internal German Defense Ministry presentation documented: multiple drone failures to take off, promised AI components not installed in delivered units, vulnerability to Russian EW jamming, strike success rate of 5/14 (36%) in one documented trial period.
The bear case calculus
Helsing's €269M Bundeswehr contract has options to €1.46B — options conditioned on performance. The €900M tri-vendor drone program (Helsing/Stark/Rheinmetall) requires Bundestag approval, not yet received. If HX-2 performance data enters the procurement record, the option values compress significantly. Ukraine paused additional HX-2 orders in January 2026.
The delta that matters
The gap between "lab test" performance and frontline performance is the entire product. A loitering munition that works in Bundeswehr test conditions but fails in EW-contested Ukraine is not a ready product. The revelation that contracted AI features were missing from delivered units is a contract integrity issue, not a development timeline issue.
Call: Helsing is worth €6–8B on current evidence — roughly half its stated valuation. Investors buying at €12B are pricing in contract execution that hasn't been proven.
Call 3 // Shield AI
Shield AI is the most likely Lockheed/RTX acquisition target in the next 18 months
The setup
Hivemind is now the autonomy layer for CCA Increment 1 — the U.S. Air Force's most important drone program with $804M in FY2026 funding alone. V-BAT has 200+ operational missions in Ukraine. The X-BAT (stealth VTOL fighter, first flight late 2026) is the world's only autonomous fighter jet in development that doesn't require a runway.
The strategic logic for Lockheed
Lockheed was not selected for CCA Increment 1. Their self-funded Vectis CCA won't fly until 2027. To stay relevant in the autonomous systems era, Lockheed needs an autonomy software stack — and Hivemind is the only battle-proven AI pilot in existence. Acquiring Shield AI ($12.7B) would cost less than two quarters of Lockheed's operating cash flow.
Alternative acquirer
RTX (Raytheon) was also awarded CCA Increment 1 autonomy contracts and is a Shield AI partner. An RTX acquisition would be more accretive but faces more antitrust risk. The CCA Increment 1 production contract award in 2026 is the trigger — whoever wins production control has the most urgent need for a certified autonomy layer.
Call: 40% probability Lockheed acquires Shield AI within 18 months, at a 20–30% premium ($15–17B). CCA production contract award is the catalyst.
📅
06 / 06
Catalysts to Watch — Next 90 Days
Catalyst Date Range Companies Affected What to Watch
CCA Increment 1 full production contract Q3 2026 Anduril, Shield AI, GA-ASI Production quantities and unit pricing. Air Force targeting $25–30M/unit vs. F-35 at $81M. Volume = Anduril's path to Fury profitability.
Arsenal-1 first production July 2026 Anduril First weapons off the Ohio factory line: Fury airframes and Roadrunner-M interceptors. Moment Anduril transitions from contract-funded to industrially scaled. Watch for production cadence disclosure.
X-BAT first flight Late 2026 Shield AI $12.7B valuation rests partly on X-BAT. First VTOL demonstration is the near-term validation. Delay = re-rating risk. Success = clear path to CCA Increment 2 bid.
Helsing Bundestag vote on €900M contract Q2–Q3 2026 Helsing HX-2 reliability data in the public record makes this a politically contested vote. Pass = valuation support. Fail/condition = significant downside.
Replicator 3 / DAWG maritime tasking Q2–Q3 2026 Saronic, Anduril Maritime is the expected next Replicator tranche given Replicator 1's aerial focus and China's USV advantage in the Pacific. A maritime Replicator announcement = largest single demand signal Saronic has ever seen.
Saronic Marauder first splash Q2–Q3 2026 Saronic The 150ft Marauder hull was flipped in December 2025. Waterborne testing validates the large-vessel production thesis and unlocks the Marauder sales pipeline.
Anduril $4B funding round closing Mid-2026 Anduril The March 2026 reported round (a16z + Thrive, $60B) has not closed as of writing. Close timing and whether secondary market pricing at $95B triggers renegotiation matters for comps.
AUKUS Pillar 2 production award 2026 ongoing Anduril, Saronic First large AUKUS Pillar 2 production award to a startup would be a major signal. Anduril deployed across all AUKUS nations; Saronic has Sydney and UK offices.
IPO filings / S-1 registration statements 2026–2027 Any of the above S-1 filing reveals actual gross margins, customer concentration, and contract pipeline — eliminating the asymmetric information supporting current private-market premiums. Expect valuation resets similar to Palantir's IPO.
📋
A / Appendix
The Incumbent Counter-Narrative
Lockheed Skunk Works
Lost CCA Increment 1 with its Vectis prototype. Self-funding Vectis development (first flight 2027). Built MDCX autonomy platform, demonstrated F-22 commanding drones from cockpit (November 2025), tested AI missile evasion on X-62A Vista. The legacy primes aren't standing still — but they're running a fast-follower strategy in autonomy while defending cost-plus programs that won't be disrupted overnight. Lockheed's AI spend is substantial; its AI product is not competitive with Anduril or Shield AI.
Northrop Grumman
Selected for CCA Increment 2 round 2 with YFQ-48A (Project Talon). Building B-21 Raider (classified; program of record). Not a primary beneficiary of the autonomy shift but not disrupted either — B-21 is a decade-long program insulated from attritable competition.

The prime contractors' real advantage: 40 years of DCSA clearances, cleared workforce, government customer relationships, and security protocols. The startups have to build this infrastructure from scratch. Anduril is doing it faster than anyone thought possible. This is why Saronic opened a DC office and hired retired flag officers — the contract is the product as much as the vessel.

Research conducted April 28–29, 2026. Primary sources: company funding announcements (TechCrunch, Reuters, WSJ), SEC/company disclosures, Congressional Research Service reports, DefenseScoop, Breaking Defense, Sacra Research financial estimates, Calibre Defence, Tectonic Defense, Army.mil and DIU.mil contract announcements, ForbesMIL, Fortune, The Information. Financial estimates where marked are third-party analyst estimates (Sacra Research); unverified by Vektor independently. This brief reflects publicly available information. Classified contract details are not included or estimated.

Vektor uses a structured research methodology: primary source triangulation (minimum 3 independent sources per factual claim), recency weighting (>60% of sources from last 90 days), and explicit distinction between verified facts and forward estimates. Track record grading for this brief will be posted at /track-record following CCA production award, Arsenal-1 launch, and X-BAT first flight.
// Companies in this brief
PLTR
Palantir Technologies
The public-market defense AI comp. PLTR vs. Anduril — who owns the government AI layer.
Custom
Silicon
AI Semiconductor Brief
The autonomy stack runs on chips. Lattice and Hivemind need inference silicon — the ASIC map.
Model
Lab Race
Foundation Model Labs
Defense AI runs on foundation models. Anthropic, xAI, and others are bidding for government contracts.
Humanoid
Robotics
Humanoid Robotics Q2 2026
Apptronik has NASA/DARPA heritage. DoD interest in humanoids for logistics and tactical applications — the defense-to-commercial robotics bridge.
⚡ Vektor Intelligence

Get briefs like this
for any company — $29/mo

Companies, markets, and trends. Drop in any name and get a structured intelligence brief in 40 seconds — earnings previews, sector trackers, competitive deep dives.

Start for $29/month → — or get your first brief free —
✓ You're in — we'll send your free brief shortly.

No credit card. Cancel anytime.