MEMORANDUM
To: Investment Committee From: [Venture Partner] Date: October 26, 2023 Subject: Anduril Industries — Series A Analysis (Hypothetical Retrospective)
Thesis: The Software-Defined Defense Prime
Anduril represents a rare opportunity to back a generational shift in the $2+ trillion global defense market. The thesis is not merely that defense budgets are increasing, but that the nature of warfare is shifting from legacy, hardware-centric platforms (aircraft carriers, manned fighters, $100k missiles) to software-defined, autonomous, and attritable systems. Anduril’s ambition is to become the “prime contractor for the AI era,” owning the software stack (Lattice OS) that serves as the central nervous system for a proliferated web of low-cost, high-intelligence hardware.
This is a potential fund-returner because defense primes historically trade on decades-long moats of procurement relationships and manufacturing infrastructure, not software iteration speed. Anduril has inverted this model. By solving an acute, urgent need for the DoD—real-time situational awareness and autonomous effectors—Anduril is building a data flywheel. Every sensor deployed, every drone flown, improves the Lattice OS model, creating a network effect that hardware-only competitors cannot replicate. The upside is not just a successful defense contractor; it is the potential to own the "Windows of the modern battlefield" operating system.
Product & Wedge: The Lattice Loop
Anduril’s strategy is a classic hardware-enabled software wedge. The hardware is impressive—Ghost autonomous helicopters, Altius loitering munitions, and the Sentinel tower. However, the strategic asset is Lattice OS.
Lattice ingests data from thousands of heterogeneous sensors (radar, EO/IR, sonar, satellite) and fuses it into a 3D geospatial command-and-control interface. It applies computer vision and autonomy algorithms to identify threats autonomously, removing the "swivel chair" integration problem that plagues the DoD.
The wedge is the Counter-UAS (C-UAS) and perimeter security market. The DoD urgently needs to protect bases from cheap drones, a problem expensive systems like the Patriot cannot solve economically. Anduril sells the complete stack: the Sentinel tower detects the drone, Lattice identifies it, and the Anvil interceptor (a kinetic drone) takes it down. This end-to-end capability is forcing the government to buy the software platform to operate the hardware, seeding Lattice as the default UI for future autonomous missions.
Market & Competition
The Total Addressable Market spans the DoD budget across land, air, sea, and cyber. The immediate market is the "urgent operational need" budget, which bypasses traditional multi-year acquisition cycles.
Established Primes:
- Lockheed Martin, Northrop Grumman, RTX: These are the incumbents. They own the relationship and the certifications, but their software is notoriously clunky, cost-plus contracts disincentivize speed, and their hardware lacks native autonomy.
- General Atomics: The dominant drone maker (Predator/Reaper). They are slow-moving and reliant on proprietary data links rather than open architecture.
New-Space/Defense Startups:
- Shield AI: The most credible direct competitor. Their "Hivemind" autonomy stack focuses heavily on GPS-denied navigation and swarming F-16s. They are a software-first threat.
- Palantir: A partner/rival. Palantir has the data fabric (Foundry/MetaConstellation) for command, but they lack the physical effector layer. Anduril competes by offering "sense + decide + act" in one loop.
- Skydio: Dominant in small autonomous drones, but focused on the tactical edge (short-range recon) rather than the operational OS layer.
Traction & Business Signal (Publicly Known)
- SOCOM Contract: Early success with the Autonomous Surveillance Towers (AST) for border security, demonstrating non-DoD revenue diversification (DHS).
- USSOCOM SIB: Awarded a $1B ceiling contract for counter-UAS systems, validating the technical approach against larger incumbents.
- USAF Bidding: Anduril publicly bid the Fury (autonomous fighter) against legacy primes for the Collaborative Combat Aircraft (CCA) program. While the outcome of specific tranches is mixed, the fact that the Air Force allowed a Series E startup to bid for a program of record is a massive signal of intent.
- Scale: The company has announced major manufacturing expansions ("Arsenal-1"), signaling that backlog is sufficient to justify capex usually associated only with incumbents.
- Revenue: Unknown specifically. The company has stated it is not a "science project"; public estimates suggest DoD obligations are in the hundreds of millions annually, growing rapidly, but specific ARR and gross margin data are private.
Risks: The Three Deal Killers
If this investment fails, it will be for one of the following three reasons:
1. The Government Procurement Moat is Real (The "Valley of Death") Anduril is a commercial-velocity company trying to sell to a bureaucracy. The DoD "colors of money" are notoriously slow. The risk is that the government uses Anduril for successful "urgent" pilots and experiments but reverts to Lockheed/Raytheon for "Programs of Record" (long-term contracts) due to political inertia, lobbying, and classified facility requirements. If Anduril cannot convert pilots into 10-year multi-billion dollar IDIQs, it will be trapped as a subcontractor to the very primes it seeks to displace.
2. The "Attritable" Margin Trap The thesis relies on proving that high-tech can be high-margin, not just high-volume. Anduril must compete against companies like AeroVironment (Switchblade) and others in a race to the bottom on price for loitering munitions. If the market forces them to sell hardware at commodity margins to win share, the "software story" breaks. They need to charge for Lattice even when the hardware is cheap. If the government refuses to pay for software licenses (traditionally they buy hardware only), this becomes a low-margin aerospace machining business, not a software business.
3. Autonomy Hype vs. Actual Efficacy Anduril claims AI can process sensor data better than humans. If the AI fails in a contested environment—say, a false positive causes a friendly fire incident, or the system is easily spoofed by electronic warfare—the trust in the software is destroyed. Defense is a zero-failure industry for certain use cases; unlike a mobile app, a "bug" in Lattice can have geopolitical consequences. A single high-profile failure could freeze adoption.
Verdict: INVEST (Hypothetical Series A)
At a Series A price, the risk/reward is asymmetric. The technical talent density (Palmer Luckey, ex-Palantir, ex-SpaceX) and the early SOCOM validation de-risk the product. The macro environment (Ukraine, Taiwan) forces a doctrinal shift toward autonomous systems regardless of internal DoD politics.
The One Question Required Before Wiring Funds: "Is the Lattice OS architecture truly open and interoperable enough to win 'Programs of Record' away from Lockheed, or are we just selling a better proprietary drone?"
We need to know if the data model is being built to become the connective tissue for other vendors' hardware, or if it is just a closed-loop accessory for Anduril's own drones. If Lattice can ingest a Boeing drone as easily as an Anduril drone, this is a trillion-dollar software company. If it cannot, it is merely a very good defense contractor.