← Back to list
Colossus (Invest Like the Best / Business Breakdowns)Podcast26 May 2022Source: joincolossus.comHost: Patrick O'Shaughnessy

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

In plain words

This episode explains how Anduril is disrupting the defense industry with a Silicon Valley approach. The founder argues that traditional contractors are stuck in a 'cost-plus' model (government pays all costs plus a guaranteed profit), making projects slow and expensive. Anduril instead invests its own money, uses fixed-price contracts, and puts software first. Three key holdings: Anduril itself (took only 3 years to turn a border security prototype into a formal program, very fast); Lockheed Martin (its F-35 jet started in the 1990s and costs over $1.5 trillion, a classic example of the old model); SpaceX (praised as a role model for using fixed-price contracts and on-time delivery, building the world's largest satellite network in 2 years).

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

Anduril is a defense technology company founded in 2017 by Palmer Luckey (who previously founded Oculus and sold it to Facebook), aiming to challenge the traditional defense industry at Silicon Valley speed. In the interview, CEO Brian Schimpf discussed the history of the defense industry, emphasizi

~12 min full read · 9 sections
Deep Analysis

Anduril: Building the Future of Defense — Interpretation

At a Glance

Brian Schimpf, CEO and co-founder of Anduril, former engineering lead at Palantir. The main thread of this episode: how Anduril challenges the business model and technological paradigm of the traditional defense industry at Silicon Valley speed. Brian Schimpf argues that the defense industry has degenerated from an innovation pioneer in the 1950s–70s into a bureaucratic system locked in by the cost-plus model, and that Anduril's core disruption lies not in the technology itself, but in rebuilding the incentive structure with a "self-funded, fixed-price, software-first" business model — the most fundamental transformation in the entire industry.


Theme 1: The Decline of the Defense Industry — From "Manhattan Speed" to "F-35-Style Delays"

Brian Schimpf points out that the U.S. defense industry was once the fastest and most innovative sector globally from the 1950s to the 1970s, but has since been dragged into a "slow trap" by the cost-plus model and excessive consolidation.

  • Historical Contrast: In the 1950s, the Pentagon (the world's largest office building) was completed in 13 months; intercontinental ballistic missiles went from zero to deployment in just a few years; Skunkworks built the most advanced aircraft with pencils and paper. Today, the F-35 program began in the early 1990s, with a lifecycle cost exceeding $1.5 trillion; the next-generation nuclear attack submarine is expected to launch in 2035 and serve until 2085 — "I can't imagine doing technology on that timescale."
  • Turning Point: The "Last Supper" of the Department of Defense in the 1990s — the then-Secretary required roughly 50 major defense contractors to consolidate on their own, ultimately shrinking to 5-6 giants including Lockheed Martin, Northrop Grumman, Raytheon, General Dynamics, and Huntington Ingalls, which now carve up the vast majority of the budget.
  • Perverse Incentives of the Cost-Plus Model: Under this model, the government reimburses contractors for all costs and pays a fixed profit (typically 7%-12%). Contractors bear no delivery risk, so there is no incentive to accelerate or reduce costs. In contrast, fixed-price contracts (such as those used by SpaceX) transfer all risk to the contractor — "If you can do it cheaply, great; if not, you lose money."
  • Key Judgment: Ten years ago, people still believed existing players could solve the problem; today, the consensus has shifted — "Traditional companies are not lacking in skills, but their processes, mindset, and construction methods are tailored for large aircraft and ships, making them unable to adapt to the era of software and AI."

Theme 2: Anduril’s Business Model — "Self-Funding" vs. "Cost-Plus"

Brian Schimpf emphasizes that Anduril’s most fundamental differentiator lies in its business model: using its own money for R&D, delivering at fixed prices, and making software the core asset — a stark contrast to traditional defense contractors, where "R&D is paid for by the government and profits come from costs."

  • Software Platform Lattice: From day one, Anduril decided to build a universal software layer (similar to AWS) applicable to various scenarios such as detection drones and underwater targets. This upends traditional practices — Boeing’s aircraft projects each have completely independent software stacks, and "shared software is almost an accident."
  • Why Build Hardware In-House:

1. Control Over Destiny: Only by controlling both software and hardware can deep optimizations be achieved in areas like edge AI inference, battery management, and autonomous flight.

2. Business Model Reality: The Department of Defense lacks a mature software procurement model (unlike SaaS) but understands hardware costs — "Hardware is countable; it wasn’t there before, now it is, and they know why it’s worth the price."

3. Avoid Persuading Traditional Vendors to Change Their Business Models: If Anduril only provided software, it would need to convince hardware vendors to abandon their cost-plus model — "This is nearly impossible."

  • R&D Investment Comparison: Most defense companies allocate a very low percentage of revenue to R&D, and this cost is reimbursed by the government — "Profit is calculated on top of R&D costs, which I find insane." Anduril, in contrast, insists on "self-funding" — "We won’t take on a project unless we are willing to put our own money into it."
  • Data Support: Anduril’s speed from prototype to Program of Record — approximately 3 years for the border security system and about 2.5 years for the counter-drone system — is described by Schimpf as "one of the fastest in federal government history."

Theme 3: Future Warfare—From "Day One Overwhelming Force" to "Decentralized Attrition Warfare"

Brian Schimpf argues that future conflicts will split into two phases: high-end power projection on Day One, followed by a protracted, decentralized attrition war centered on small drones and precision munitions—the latter being the direction Anduril's technology system targets.

  • Day One Framework: The U.S. needs to project overwhelming force over long distances (stealth bombers, long-range precision munitions) to quickly decide the battlefield and establish deterrence. This is where "sophisticated, expensive systems" like the F-35 come into play.
  • Protracted Phase Framework: Once a conflict enters the ground phase, it becomes "a very tactical, decentralized fight." Key data point: In the 2020 Nagorno-Karabakh conflict, Azerbaijan used Turkish Bayraktar TB2 drones and Israeli loitering munitions, even converting old An-2 biplanes into remote-controlled decoys to lure Armenian air defense systems into exposing themselves—Armenia, despite possessing tanks, armored vehicles, and Russian-made air defense systems, was "completely destroyed with almost zero losses."
  • Lessons from Ukraine: "I wouldn't want to be sitting in a tank in a modern conflict—that's not a good place to be." What works is highly decentralized small units equipped with Javelin missiles, Stinger missiles, and small drones for reconnaissance and strikes. Electronic warfare (GPS jamming, communications suppression) has become the norm.
  • Machine-on-Machine Warfare? Schimpf believes machine-on-machine combat will increase significantly, but humans remain unavoidable—"the ultimate goal of conflict (changing political leadership, controlling territory) often requires ground forces." The key is to make aggressors understand that "you cannot succeed"—"Russia miscalculated that, and I don't think it would be any easier for them against any other neighbor."

Theme 4: Ghost vs Predator — How Autonomy Changes the Economics of Warfare

Brian Schimpf uses the comparison between the Ghost drone and the Predator/Reaper to illustrate how a "software-first, autonomous operation" approach fundamentally alters the deployment logic and cost structure of unmanned systems.

  • Predator/Reaper operating model: Essentially a "remotely piloted aircraft"—one pilot flies with a joystick, one sensor operator controls the camera, and roughly 12 people behind them process data and produce analysis reports. Each drone requires a massive ground support team and infrastructure.
  • Ghost design philosophy: A small helicopter that fits into a gun case, with a range of about a dozen miles. How it operates: "You tell it what you want to see—for example, 'go check that building'—and it automatically plans the route, takes off, captures footage, and returns." The goal is one person operating 10 drones, rather than 20 people operating one.
  • Core principle: "There should be a rule that no new unmanned system should ever be built with a joystick and rudder again."
  • Economic implications: When autonomy reduces manpower requirements, the cost logic of the system changes—no longer requiring systems to be "as expensive and sophisticated as possible" (because manpower is a scarce resource), but instead enabling the deployment of large numbers of cheap, expendable systems. "I don't mind losing them, they don't need to be that specialized, and I don't even mind if the enemy knows they are there."

Theme 5: How to Break into the Department of Defense — "Find Early Adopters, Don't Ask Them to Admit Failure"

Schimpf shared Anduril's core strategy for entering the defense market: identify early adopters with urgent mission needs, focus on "entirely new problems" rather than "replacing existing solutions," and avoid triggering the institutional "admit failure" penalty mechanism.

  • Early Adopter Strategy: The first breakthrough was counter-drone systems — "This was the first time the U.S. military faced a threat where the enemy could attack bases from a distance." Special operations forces felt this threat first; they had urgent needs and a tolerance for risk.
  • Institutional Inertia: Even after technology is proven effective, the budget cycle still takes 2–3 years to respond — "Budgets are written before the technology is even invented." The more fundamental issue is: If officials admit that existing solutions are wrong, they are penalized, rather than rewarded for choosing a better option.
  • Key Lesson: Do not try to replace existing systems (which requires the other party to admit failure); instead, target areas where "technological change makes entirely new solutions possible." Both Palantir and SpaceX had to sue the government to get fair treatment — "If something is locked into an existing path, it is almost impossible to get them to change direction."
  • Open Strategy: Anduril publicly shares its "playbook" for entering the Department of Defense — "This field needs more success stories. We won't solve every problem; others need to succeed too. If the Department of Defense becomes a better buyer, that benefits us as well."

Mentioned Positions

Position Analyst Stance Key Data
Anduril (itself) Bullish (under construction) Border security system: from prototype to deployment in 3 years; Anti-drone system: 2.5 years; SOCOM contract: approximately $10 billion
Lockheed Martin Risk warning (representative of traditional model) F-35: started in the early 1990s, total lifecycle cost exceeds $1.5 trillion
Boeing Risk warning (losses on fixed-price contracts) Losses due to "strategically lowball bidding" on fixed-price contracts
SpaceX Bullish (model benchmark) From zero to the largest satellite constellation operator in just 2 years; fixed-price contracts, on-time delivery
Palantir Neutral (learning reference) Once had to sue the government to obtain fair treatment
Bayraktar TB2 (Turkish drone) Neutral (case cited) Played a decisive role in the Nagorno-Karabakh conflict

Judgments Worth Remembering

1. "The biggest problem with the cost-plus model is that contractors have no incentive to speed up or reduce costs—they get a fixed profit regardless." —Brian Schimpf. This incentive structure has stretched defense project timelines from years to decades and inflated costs from hundreds of millions to trillions.

2. "There should be a rule that no new unmanned system can be built with a joystick and rudder again." —Brian Schimpf. Anduril's Ghost drone aims for "one operator controlling 10 units," while the traditional Predator requires 20 people to operate one—autonomy changes the entire economics of warfare.

3. "Don't try to replace existing systems—that requires the other side to admit failure, and the system punishes those who admit failure. Target areas where technological change makes entirely new solutions possible." —Brian Schimpf. This is the fundamental reason Anduril chose counter-drone (a new threat) over replacing the F-35 (an existing system).

4. "The core challenge for hardware companies is not technology, but how to cheaply test product-market fit—let customers see a prototype, say 'this is what I need,' and then they will walk through brick walls for you." —Brian Schimpf. Anduril built a counter-drone prototype in 12 weeks for a competition and a border security prototype in 3 months—speed itself is a competitive advantage.

5. "If China can acquire new capabilities in 3-5 years, and we need 7-10 years—even if we start from a higher base, they will quickly surpass us. Iteration speed may be the most important thing." —Brian Schimpf. This is the fundamental driver of Pentagon reform and the reason Anduril exists.

6. "Hardware requires dozens of specialized skills—electrical design, mechanical design, manufacturing, supply chain, logistics. It's not for the faint of heart." —Brian Schimpf. Coming from a software background into hardware, his most important lesson was: the cost of hardware innovation is far higher than software, but the returns are also greater.

7. "As technologists, we have a responsibility to tell the government what is possible, what can be used responsibly, and what cannot—not just do something because you can." —Brian Schimpf. Anduril refused to do facial recognition because "even if we give it to you, you cannot use it responsibly."

8. "Both Palantir and SpaceX had to sue the government to succeed. This tells us: if the system is locked into existing paths, you can almost never get them to change direction—unless you target an entirely new problem." —Brian Schimpf. This is the core lesson behind Anduril's strategic choices.