← Back to list
Colossus (Invest Like the Best / Business Breakdowns)Podcast25 May 2022Source: joincolossus.comHost: Colossus

Anduril: Building the Future of Defense - [Business Breakdowns, EP. 59]

In plain words

This piece explains how defense startup Anduril is challenging traditional giants like Lockheed Martin by using Silicon Valley's fast-iteration approach. CEO Brian Schimpf argues the old 'cost-plus' model (government bears risk, contractors get 7%-12% profit) makes projects slow and expensive—the F-35 fighter jet costs over $1.5 trillion. Anduril's strategy: self-fund R&D, build software and hardware together, find customers with urgent problems (like drone threats), and prototype quickly. Key holdings: Anduril itself (its anti-drone system went from concept to official program in about 2.5 years, a federal record); Lockheed Martin (traditional model, F-35 cost overrun); SpaceX (successful fixed-price contract model).

AI SummaryAI-generated · may contain errors · verify against the original

Anduril is a high-tech defense systems company founded in 2017 by Oculus founder Palmer Luckey, aiming to challenge the traditional defense industry at Silicon Valley speed. The report's core argument is that Anduril has successfully delivered products to the U.S. Department of Defense (DoD) and its

~11 min full read · 9 sections
Deep Analysis

Anduril: Building the Future of Defense – Commentary

At a Glance

Brian Schimpf (CEO and co-founder of Anduril, previously responsible for engineering at Palantir) deconstructs how this high-tech defense company, founded in 2017, challenges the traditional defense industry at Silicon Valley speed. Core thesis: The "cost-plus" model of the defense industry is no longer sustainable. Anduril is reshaping the underlying logic of defense procurement through self-funded R&D, integrated hardware-software development, and the pursuit of early adopters for "urgent missions"—a shift that would have been nearly impossible just a decade ago.


Theme 1: The Chronic Ailment of the Defense Industry — From Cold War Innovation to Bureaucratic Stagnation

Brian Schimpf argues that the U.S. defense industry has degenerated from an "innovation engine" into a "bureaucratic behemoth," with the root cause lying in distorted incentives.

  • Historical Context: From the 1950s to the 1970s, defense was at the forefront of technological innovation — the Pentagon (the world's largest office building) was built in 13 months, ICBMs went from zero to deployment in just a few years, and Skunkworks designed the most advanced fighter jets using pen and paper. During the Cold War, the defense industry attracted the best talent, and the pace of R&D was astonishing.
  • Turning Point: In the late 1970s to early 1980s, the DoD shifted its focus to cost control and predictability, and bureaucracy began to spread. The "Last Supper" in the 1990s — when the then-Secretary of Defense called for the consolidation of roughly 50 major defense contractors — led to the industry concentrating into 5-6 giants (Lockheed Martin, Northrop Grumman, Raytheon, General Dynamics, etc.), capturing the vast majority of the DoD budget.
  • Current Data: The F-35 program began in the early 1990s, with a total lifecycle cost exceeding $1.5 trillion; the new nuclear attack submarine is expected to be launched in 2035 and remain in service until 2085. Brian notes: "I can't imagine working in the technology sector on that kind of timeline."
  • Mechanism Breakdown: Under the cost-plus model, the government bears all the risk, while contractors earn a fixed profit (typically 7%-12%). This creates "no incentive to speed up or deliver on time" — contractors are instead motivated to extend timelines and increase complexity. In contrast, SpaceX uses firm fixed-price contracts, delivering on time with all parties satisfied.

> Reader's Note: As the CEO of Anduril, Brian has a vested interest in highlighting the flaws of the traditional model to underscore the superiority of his own. However, the F-35 cost data and industry concentration he cites are public facts, and his analytical framework (incentives determine behavior) holds up in business analysis.


Theme 2: How to Penetrate the DoD — Finding Early Adopters with "Urgent Missions"

Brian argues that the key to entering the defense market is to find early adopters with "urgent missions," rather than attempting to overhaul the entire system.

  • Strategy: Instead of trying to completely transform the DoD, the focus is on finding customers who "have an urgent problem to solve and are willing to take risks." Anduril's first breakthrough was in counter-drone systems — an emerging and urgent threat where traditional methods (spending five years writing requirements, soliciting bids, and developing) were entirely ineffective.
  • Specific Case: The counter-drone system went from concept validation to a Program of Record in approximately 2.5 years, one of the fastest in federal government history. Brian emphasized: "You need a system integration partner that can understand the threat, the technology landscape, and continuously evolve — just like we treat our phones or Tesla Autopilot."
  • Data Support: Anduril performed best in a drone countermeasure competition hosted by the DIU (Defense Innovation Unit), "shooting down most of the incoming drone threats." This was followed by pilot deployments, iterative improvements, and ultimately scaling into mass production.
  • Falsification Condition: If the DoD cannot establish faster procurement channels for emerging threats (such as drone swarms), Anduril's growth ceiling will be limited to "only being able to find a few urgent missions."

Theme 3: Software-Hardware Integration — Why Anduril Must Build Its Own Hardware

Brian explains that Anduril’s decision to develop both software (the Lattice platform) and hardware is driven by both technical necessity and business model considerations.

  • Technical level: Co-design of software and hardware enables optimal performance. For example, Anduril’s Ghost drone is designed as “a flying stick,” with a modular structure that allows rapid payload swaps. Edge computing (running AI inference in harsh conditions such as deserts with solar power) requires deep coupling between hardware and software.
  • Business model level: The DoD “struggles to understand software business models” — the SaaS model is virtually nonexistent in the defense sector, as the government prefers custom development and ownership. Hardware, by contrast, “is straightforward: it didn’t exist before, now it does, and they understand why it’s worth the price.” Hardware contracts enable rapid scaling.
  • Comparison with legacy: Boeing’s multiple aircraft programs have “completely independent software stacks, with shared software being almost accidental.” Anduril’s Lattice platform, on the other hand, serves as a common infrastructure layer applicable to drones, underwater vehicles, border surveillance, and other scenarios.
  • Key data: The Lattice platform went from concept to prototype in just 3 months, with a pilot deployment at the southern border; it became a formal program of record in approximately 3 years — “one of the fastest in federal government history.”

Theme 4: The Future of Warfare — From "Day One" to "The Long War"

Brian argues that future conflicts will unfold in two phases: high-end power projection on Day One, followed by a protracted, highly dispersed tactical struggle — the latter being the scenario Anduril's technology is designed for.

  • Day One: Requires overwhelming force (stealth bombers, long-range precision munitions) to rapidly destroy key enemy infrastructure. This is the stronghold of traditional defense primes.
  • The Long War: Conflicts in Ukraine and Nagorno-Karabakh demonstrate that cheap drones, precision munitions, and electronic warfare will dominate the battlefield. Brian notes: "I wouldn't want to be in a tank in a modern conflict — that's not a good place to be." Azerbaijan used Turkish Bayraktar TB2 drones and Israeli loitering munitions to destroy Armenia's Russian-made air defense systems and armored units with near-zero casualties.
  • Anduril's Positioning: Focuses on "defensive, short-range, and distributed" capabilities — making an invader "the hardest nut to crack." The Ghost drone (fits in a gun case, range of approximately 12 miles) is designed around the concept: one person can operate 10 drones, rather than 20 people operating one.
  • Comparison with Predator: The Predator/Reaper requires approximately 14 people (2 pilots + sensor operator + 12 data analysts) and extensive ground infrastructure; Ghost aims for 1-2 people to operate, with autonomous flight and automatic target recognition.

> Unique Insight: Brian proposes the rule that "new unmanned systems should no longer use joysticks and rudders" — all drones should default to autonomous flight, with humans only responsible for mission assignment and final decision-making. This stands in stark contrast to the DoD's current procurement practices.


Theme 5: Ethics, Politics, and Open Innovation — Anduril's "Counter-Intuitive" Strategy

Brian outlines Anduril's stance on ethics, politics, and competition, with the "open innovation" strategy being particularly counter-intuitive.

  • Ethical Framework: Brian explicitly rejects the "killer robot" narrative — "No one is proposing that, the government doesn't want it, and it wouldn't even work." Anduril's principle is that "humans are always responsible for system behavior." AI cannot handle "contextual" judgments (political objectives, proportionality, etc.), so humans must retain decision-making authority. Anduril refuses to engage in facial recognition because "it may not achieve the intended effect and cannot be used responsibly."
  • Political Neutrality: There are "no political discussions on Slack" within the company. Brian describes himself as a "lifelong Democrat," while co-founder Palmer Luckey holds different political views. However, national defense is a "highly bipartisan" issue.
  • Open Innovation: Anduril proactively publishes its "playbook" for working with the DoD, even educating potential competitors. Brian explains: "If the DoD becomes a better buyer with more successful models, that benefits all of us. We won't solve every problem, and this field needs fresh blood." This strategy is based on the belief that the defense market is large enough and "there are easier ways to make money than this."

Mentioned Positions

Position Guest Stance Key Data
Anduril (itself) Bullish—disrupting the traditional defense model Counter-drone system went from concept to program record in ~2.5 years; Lattice platform from prototype to formal program record in ~3 years (one of the fastest in the federal government)
Lockheed Martin Risk warning—representative of the traditional cost-plus model F-35 lifecycle cost exceeds $1.5 trillion, originating in the early 1990s
Northrop Grumman Risk warning—beneficiary of industry concentration Along with Lockheed and 5-6 other giants, captures the vast majority of the DoD budget
Raytheon Risk warning—same as above Same as above
General Dynamics Risk warning—same as above Same as above
Boeing Risk warning—case of fixed-price contract losses Incurred losses due to strategically low-bidding fixed-price contracts
SpaceX Bullish—successful model reference Went from zero satellites to the largest constellation operator in just 2 years; fixed-price contract model
Palantir Neutral—learning reference (Brian's former employer) Once had to sue the government to obtain contract opportunities
Bayraktar TB2 (Turkish drone) Case reference—demonstrating drone effectiveness Destroyed Armenian Russian-made air defense systems during the Nagorno-Karabakh conflict
DJI Background reference—small commercial drone threat Used as an IED-style threat (dropping grenades)

Judgments Worth Remembering

1. “The cost-plus model gives contractors no incentive to move faster”——Brian Schimpf: Fixed profit margins (7-12%) mean the longer and more complex the project, the higher the contractor's profit. This stands in stark contrast to SpaceX's fixed-price contracts.

2. “The DoD’s budget is decided 2-3 years before a technology is invented”——Brian Schimpf: Anduril’s counter-drone technology, after a successful pilot, was told to “wait 3 years to enter the budget cycle,” exposing systemic barriers to innovation.

3. “New unmanned systems should no longer use joysticks and rudders”——Brian Schimpf: Anduril’s design principle is one operator controlling 10 drones, not 20 people controlling one. Autonomous flight and automatic target recognition are core.

4. “No one is proposing killer robots, the government doesn’t want them, and they wouldn’t even work”——Brian Schimpf: AI cannot handle contextual judgment (political objectives, proportionality); humans must retain decision-making authority. Anduril refuses to do facial recognition.

5. “Hardware companies should test product-market fit as cheaply as possible”——Brian Schimpf: Anduril’s strategy is rapid prototyping (within 12 weeks), letting customers “see something fly” to get real feedback. This contrasts with the DoD’s traditional “write a 5-year requirement first” approach.

6. “Open innovation benefits the DoD as a whole”——Brian Schimpf: Anduril proactively publishes its “playbook” for working with the government because “if the DoD becomes a better buyer, that benefits all of us.” This is a counterintuitive competitive strategy.

7. “China can field new capabilities in 3-5 years, while we need 7-10 years”——Brian Schimpf: Even if the U.S. starts from a higher baseline, slower iteration cycles mean it will be overtaken. Speed is “the most important factor.”

8. “The ‘last supper’ of the defense industry led to 5-6 giants capturing the vast majority of the budget”——Brian Schimpf: The industry consolidation of the 1990s stifled innovation, creating extremely high structural barriers for new entrants.