Key points
- Anduril lands a multi-hundred-million-dollar program of record 22 months after founding and is now building a submarine shipyard with General Dynamics Electric Boat.
- Trae Stephens says 95% of companies are uninvestable and $100M+ pre-revenue rounds need trillion-dollar outcomes to beat the S&P 500.
- Trae Stephens says a good quest finds you, and the best companies begin with an obsession, not an entrepreneur identity.
Why Anduril Bet on Scale Over a Defense Niche
Anduril is betting that the next great defense prime will sell across domains, because the Department of Defense will not build a $100bn company out of a single category. Trae Stephens started it after failing to sell Palantir into the intelligence community, and the venture math he now works with requires trillion-dollar outcomes to beat the S&P 500.
Takeaways
- Defense rewards multi-domain scale. Trae Stephens says no single-category company reaches a $100bn market cap with the Department of Defense. Anduril set out to be multi-domain like Lockheed and Northrop, and is building a submarine shipyard with General Dynamics Electric Boat.
- Fund math breaks at current AI prices. Trae Stephens says a pre-revenue round above $100m means a $100bn outcome cannot return a fund, so trillion-dollar results are needed. OpenAI and Anthropic may exit above $1.5tn, and the baseline to beat is the S&P 500.
- Returns come from a dozen companies. Founders Fund is 20 years old. Twelve companies, among them Facebook, SpaceX, Airbnb, Spotify and Stripe, produced 95% or more of its returns, so he needs $60bn to $100bn outcomes.
- A good quest finds you. Trae Stephens compares it to Monty Python and the Holy Grail, where King Arthur wanders until God hands him a quest. Companies start from an obsession rather than an entrepreneur identity, and SpaceX did not build Starship on day one.
- Logos and hires do not sell to government. Palantir hired retired generals and former ambassadors, and struck partnerships with Boeing, Booz Allen, Deloitte, Accenture and SAP. None won business, and Trae Stephens's agency told him it builds rather than buys.
- Venture conviction arrives in minutes. Trae Stephens knows within five minutes whether he wants a ten-year relationship with a founder, and 95% of the companies he meets are not investable. Brian Singerman told him to take 500 pitch meetings in his first year.
How Anduril got started
- Trae Stephens scraped federal contract data and met a couple hundred companies.
- He backed Expanse, originally Cadium, and concluded the gap was a software-defined, hardware-enabled company for low-cost autonomous systems.
- Peter Thiel told him to build it. Palmer Luckey first refused, then called back saying he had been fired.
- They built Sentry, a counter-intrusion tower, deploying a test unit in September and towers in December, and reached a multi-hundred-million-dollar program of record 22 months after founding.
The valuation disagreement