The Billionaire Building America's AI Military | EP #158
Recorded live at the Abundance Summit, Peter Diamandis interviews Anduril founder Palmer Luckey about building a defense products company instead of a defense contractor, funding R&D with Anduril's own money, and the recent $22 billion IVAS contract handover from Microsoft. Luckey walks through Anduril's hardware (the Fury/F-Q-44 unmanned fighter jet, Anvil, Sentry Towers, Dive LD/XL submarines, Roadrunner, Menace, Ghost drones) and his four-part test for which problems Anduril takes on: the Pentagon must deeply care, Congress must deeply care, Anduril must be able to do it well, and nobody else must already be doing a good job. He talks about founding Oculus as a hobby, getting fired from Facebook, his bitterness toward specific people (not Zuckerberg), his health longevity stack (Fountain Life, Viome, OneSkin), and closes with audience Q&A covering flying cars, military AI autonomy limits, following skills over dreams, and whether humanity will ever stop fighting wars.
Defense product company vs. defense contractorGeopolitics▶ 16:40
Luckey explains Anduril's core strategic bet: spend your own money to build a product that works, then sell it, rather than taking cost-plus government contracts that reward slowness and padding.
Cost-plus contracts, a post-WWII artifact meant to control profiteering, instead reward higher costs; 80% of major defense acquisition programs go to five companies and 30% have a single bidder.
Anduril took over the $22 billion Integrated Visual Augmentation System contract (and Microsoft's mixed-reality business) after Microsoft's HoloLens-based heads-up display effort stalled; Luckey expects to finish in 6 months what would take others 8 years.
Luckey predicts near-term military humanoid robots will first replace bored human operators of existing manned systems (missile silos, vehicles) rather than door-kicking special forces.
Anduril's four-part problem filterGeopolitics▶ 28:50
Anduril only works on a problem if the Pentagon deeply cares, Congress deeply cares, Anduril can do it well, and others aren't already doing a good enough job.
Luckey discusses Anduril's self-funded detection and suppression tech for the Wildfire X Prize, developed after devastating wildfires displaced his own family, and the perverse insurance-industry incentives that discourage prevention.
Luckey describes starting Oculus purely out of gamer passion, turning down a $1 billion acquisition offer at 13 months (Facebook later paid $2.3 billion), Zuckerberg's $10 billion VR R&D commitment (which became $60 billion), and being fired and ripped out of his own company.
Longevity and health stack (ad segment)Health▶ 32:08
Diamandis describes his three personal health protocols: Fountain Life annual full-body diagnostic scans, Viome metatranscriptomics gut-microbiome testing, and OneSkin's peptide skincare for senescent skin cells.
Luckey argues legacy media companies are running out of zero-interest-rate-era money, know their business model is dying as audiences move to independent journalists, Substacks, and YouTubers, and are lashing out as a result.
Luckey argues war will never be fully autonomous because humans must remain accountable for lethal decisions; outsourcing responsibility for violence entirely to machines removes the weight of consequence that should fall on a person.
Luckey says eVTOL wasted time on pure-electric designs instead of hybrids (batteries are most of the mass), predicts Dallas and New York as early hubs, and notes government/public-works vehicles can fly under regulatory waivers that commercial eVTOL cannot.
Luckey says deterrence through unipolar or stable multipolar power balance can produce long stretches without war, but history's repeated 'end of history' predictions have always been wrong, and asymmetric actors (extremist groups, bio-weapons) remain hard to deter.
Predictions made
openPalmer Luckey: Humanoid robots will appear in defense applications relatively soon, first replacing humans who operate existing manned systems (missile silos, vehicles) rather than performing special-forces roles.
“I'm going to be able to get done in about 6 months what other companies would take eight years to do”
Your call:
openPalmer Luckey: Any Anduril problem that passes its four-part test (Pentagon cares, Congress cares, Anduril can do it well, others aren't doing a good job) will show up as a fielded product within a year or two.
“you're going to see it in the andal showroom within a year or two”
Your call:
openPalmer Luckey: eVTOL aircraft will reach the mainstream, with cities like Dallas and New York becoming early hubs, despite current regulatory obstacles.
“I still believe that that is going to happen now there's obstacles in the way”
Your call:
openPalmer Luckey: Legacy media companies will collapse as their zero-interest-rate-era money runs out and audiences keep shifting to independent journalists, Substacks, and YouTubers.
“I don't think it'll ever be fully autonomous because the gains aren't there and the negatives are”
Your call:
openPalmer Luckey: There will eventually be an extended period of time without major war, driven by stable unipolar or multipolar deterrence, but that period will eventually end.
“I think there will be a period where war ends I think it will go on for a long period of time”
Your call:
Numbers that matter
Anduril took over Microsoft's $22 billion IVAS contractIncludes the rest of Microsoft's mixed-reality business, which Microsoft was otherwise winding down.
80% of major defense acquisition programs (MDAPs) go to just five companies; 30% have a single bidderLuckey's argument for why cost-plus contracting produces zero-competition markets.
Oculus sold to Facebook for $2.3 billion after Luckey turned down a $1 billion offer at 13 monthsFacebook later committed roughly $60 billion total to VR R&D, versus the original $10 billion ($1B/year for 10 years) pledge.
Anduril's Sentry Towers cover about 35% of the US southern borderUsed alongside towers on military bases and nuclear/energy infrastructure worldwide.
Roughly $250 billion in damage from recent Southern California wildfiresCited by Luckey/Diamandis as the scale of loss that controlled-burn funding could have prevented for a fraction of the cost.
FAA-certifiable eVTOL drivetrain components can run 15,000 flight hours without a maintenance intervalLuckey cites this as a civilian-sector advance Anduril wants to pull into military aircraft.
The UK plans to reduce its Navy's size by 30% over the next 10 years through automationLuckey's example of realistic automation gains (thousands to hundreds of people), contrasted with the unrealistic jump to zero.
Worth digging into
🕳️ Cost-plus vs. fixed-price defense contracting
Luckey frames this as the single biggest structural reason the defense industry can't innovate, comparing it directly to residential renovation incentives.
🕳️ IVAS program history and the Microsoft-to-Anduril handover
An $22 billion program moving from a tech giant to a 10-year-old defense startup is a rare, significant shift in how the Pentagon buys technology.
🕳️ Heinlein's 1945 'Solution Unsatisfactory' and AI-piloted fighters
Luckey says Anduril has directly copied Heinlein-era sci-fi concepts for human-AI co-piloting into real products.
🕳️ eVTOL hybrid-vs-electric mass tradeoff
Luckey's claim that pure-electric eVTOLs mostly haul batteries, not people, is a concrete engineering critique of the entire industry's design direction.
🕳️ Legacy media's financial collapse thesis
Luckey predicts media companies are structurally doomed from losing zero-interest-rate-era subsidy money, which is a checkable claim against specific company financials.