← Back to list
Colossus (Invest Like the Best / Business Breakdowns)Podcast9 Mar 2021Source: joincolossus.comHost: Patrick O'Shaughnessy

Josh Wolfe and Tony Thomas - The Past, Present, and Future of Defense – [Invest Like the Best, EP. 216]

In plain words

This interview discusses US defense tech gaps and geopolitical threats. The guests argue the US is losing information warfare and autonomous weapons, with AI beating human fighter pilots 5:0. They highlight Anduril (anti-drone tech), Saildrone (autonomous ocean drones), and TSMC (critical chip maker, but US production is tiny).

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

Josh Wolfe (Co-founder of Lux Capital) and retired four-star General Tony Thomas (T2), former commander of U.S. Special Operations Command, discussed the frontiers of defense technology and geopolitical threats on the Invest Like the Best podcast. The core argument: The United States faces critical

~12 min full read · 8 sections
Deep Analysis

At a Glance

Josh Wolfe (Co-founder of Lux Capital) and retired four-star General Tony Thomas (T2), former commander of U.S. Special Operations Command, discuss the frontiers of defense technology and geopolitical threats on Invest Like the Best. Core thesis: The U.S. faces critical challenges including information warfare, autonomous weapons systems, and semiconductor competition. The traditional defense procurement system is slow to iterate on technology and urgently needs to accelerate the introduction of venture capital and commercial innovation. T2 argues that AI-enabled autonomous systems could become the ultimate deterrent—making war between nations futile, because adversaries would know that even if communications and satellites are destroyed, AI can still execute retaliatory strikes.


Theme 1: Information Warfare — A Battle the U.S. Is Losing

T2 argues that the U.S. is "not even competing" in the information space, while adversaries are exploiting this gap.

  • Russia focuses on creating chaos and disinformation rather than promoting specific narratives; China systematically manipulates the global information environment — the U.S. Department of Defense receives a weekly "China version" news briefing, showing how Beijing repackages the same event for global audiences.
  • T2 points out that the U.S. faces a "truthiness" dilemma: the West pursues factual accuracy, but in information warfare, "speed is more important than perfection" — by the time verification is complete, the narrative is already set, leaving only reactive damage control.
  • Josh Wolfe adds: China has achieved information operations through strategic acquisitions of Hollywood assets (acquiring AMC for nearly $3 billion in 2012, purchasing Legendary Entertainment, etc.), resulting in almost no "Chinese villains" in American pop culture — a stark contrast to the 1980s cultural narrative where the Soviet Union was a clear enemy.

> T2's original quote: "We in the Western world suffer from the intent to be factual to our fault... better to be fast, get something out there in the information space and recover off that."


Theme 2: Autonomous Weapon Systems — From the "Day of the Drones" to AI Air Combat

T2 argues that autonomous systems have fundamentally changed the nature of warfare, yet the U.S. defense procurement system remains stuck in old thinking.

  • The "Day of the Drones" Warning: During the 2015 Battle of Mosul, ISIS used modified DJI drones (carrying 40mm grenades) to launch over 70 sorties against coalition forces within 24 hours. Despite having absolute air superiority, coalition forces could only counter these "swarms" with small-caliber weapons.
  • DARPA AI Air Combat Test: AI defeated a human ace pilot 5:0. T2 notes that the AI employed "suicidal tactics" (trading one for one), but human pilots dismissed this — "AI knows no fear, feels no fatigue; it might take you out with a suicide move, resulting in a 1:0 score for the machine over the human."
  • T2's Core Argument: The question "why does it need to be manned" should be challenged — from helicopters to fighter jets, moving pilots from the cockpit to ground control stations eliminates the "pucker factor," allowing AI to handle decision-making in high-stress environments.

> T2's Original Quote: "I don't want my kids going up against some AI-enabled fighter capability in the future. We need to leap ahead right now. Make the sky black with automated AI-enabled capabilities."


Theme 3: Semiconductors — The Most Fragile Geopolitical Node

Josh Wolfe points out that TSMC is "the most geopolitically important company in the world," and the U.S. reliance on chip manufacturing is a critical vulnerability.

  • Concentration risk: 75% of chips are manufactured in Asia; only TSMC and Samsung can mass-produce advanced nodes below 7 nanometers. TSMC holds a 30% market share for nodes below 40 nanometers, with annual revenue of $50 billion.
  • Severe shortage of U.S. domestic capacity: TSMC's Arizona fab can produce only 20,000 wafers per month at full capacity, while global demand stands at 12 million wafers — "that's like a speck of dust on a chip." The U.S. currently has only one pure defense foundry (in Minnesota), with near-zero output.
  • "Chiplets" are the future direction: Wolfe argues that instead of pursuing monolithic system-on-chip (SoC) designs, a modular chiplet architecture offers higher yields and lower design costs. This could represent a disruptive opportunity on the scale of "TSMC 20 years ago."
  • Supply chain fragility: The critical laser in ASML's EUV lithography machines is exclusively supplied by a single German company (Trumpf), making the entire production chain highly concentrated.

Theme 4: Space Competition — The Area Where the U.S. Is Falling Behind

T2 stated bluntly: "Space is the area I worry about most. We are too far behind."

  • Commercial Analogy: Wolfe compares space development to the 19th-century railroad — companies like SpaceX are "vertical railroads," reducing launch costs from $50,000 per kilogram for the Space Shuttle to approximately $3,000 per kilogram for the Falcon 9.
  • Infrastructure Race: Low Earth orbit satellite communication constellations (thousands of satellites), in-orbit manufacturing (Lux invested in Varda, founded by former SpaceX team members), and satellite repair robots are forming a complete ecosystem.
  • Military Dimension: China has demonstrated anti-satellite capabilities (laser jamming, kinetic destruction) and visited the far side of the Moon. T2 notes that the U.S. is still debating "whether to weaponize space" — "the genie is out of the bottle, the ship has sailed."
  • Wolfe adds: China is lobbying at the United Nations to define facial recognition standards, while they have already exported related technology to dozens of countries — "Whoever defines the standards first dominates the rules."

Theme 5: The Moral Dimension — The Ethical Advantage of Precision Strikes

Josh Wolfe proposes a framework of "technological precision as ethical precision," arguing that more accurate weapons result in less collateral damage.

  • Historical Evolution: From one-on-one hand-to-hand combat, to weapons of mass destruction with nuclear capabilities, and then to precision-guided munitions—the ability to distinguish between "a farmer returning home with a hoe" and "a terrorist holding an AK-47" is itself a moral advancement.
  • T2's "God Call": At the peak of the campaign against ISIS, T2 once authorized a strike—he was told he could accept collateral damage of 10 non-combatants (NCV=10), yet among thousands of authorizations, he used this authority only once. The target was in a taxi, and he could never confirm whether the driver was an accomplice or an innocent civilian.
  • Legal Safeguards: Wolfe was astonished to find, in the theater command center, that a lawyer sat beside the drone operator—"This is one of America's greatest technologies: the rule of law, rules of engagement, democratic oversight."
  • T2's Bottom Line: He advocates that "AI-assisted decision-making should be pushed to the very edge, but lethal decisions should not be fully unmanned"—"When you lose that little devil of doubt, it's time to leave the profession."

Mentioned Positions

Position Guest View Key Data
Anduril Bullish (mentioned by Wolfe; Chris Brose as Strategy Officer) Defense technology company, focused on counter-drone systems
Saildrone Bullish (mentioned by both T2 and Wolfe) Completed autonomous Antarctic circumnavigation (196 days/12,000 nautical miles); first bidirectional transatlantic crossing (68 days/3,400 nautical miles)
Performance Drone Works Bullish (mentioned by Wolfe) Counter-drone company backed by Lux
Applied Intuition Bullish (mentioned by Wolfe) Simulation technology company, co-invested by Lux and Andreessen Horowitz
Varda Bullish (mentioned by Wolfe) Space manufacturing company, founded by former SpaceX team, co-invested by Lux and Founders Fund
Relativity Space Bullish (mentioned by Wolfe) Rocket launch company, backed by Lux
Rocket Lab Bullish (mentioned by Wolfe) Rocket launch company
SpaceX Bullish (mentioned by Wolfe) Falcon 9 launch cost approximately $3,000/kg
Blue Origin Neutral mention (mentioned by Wolfe) Rocket launch company
TSMC Risk warning (Wolfe) Holds 30% market share in sub-40nm nodes, annual revenue of $50 billion; Arizona fab produces 20,000 wafers/month vs global demand of 12 million
Intel Neutral/Risk warning (Wolfe) Lagging in advanced process nodes, needs government push for foundry business
ASML Bullish (Wolfe) Market cap of $250 billion, sole supplier of EUV lithography machines
Cadence Bullish (Wolfe) Market cap of $40 billion, chip design software
Synopsys Bullish (Wolfe) Market cap of $40 billion, chip design software
LAM Research Bullish (Wolfe) Leading in 3D chip stacking technology
KLA/Tencor Bullish (Wolfe) Semiconductor equipment company
Applied Materials Bullish (Wolfe) Semiconductor equipment company
Trumpf Risk warning (Wolfe) German company, sole supplier of key lasers for EUV lithography machines
General Atomics Neutral mention (Wolfe) Predator drone contractor
DJI Risk warning (T2) ISIS uses its modified drones to launch attacks

Judgments Worth Remembering

1. T2: Information warfare determines everything — "You can win every kinetic battle, but lose the information war and you lose it all." The U.S. is "not even competing" in the information domain, while Russian and Chinese information operations have become systematic and scaled.

2. T2: AI air combat has already defeated human ace pilots 5:0, but the military is making excuses to deny it. "AI knows no fear, no fatigue — it might take you out with a suicide tactic, and the result is machine 1:0 human." He argues for questioning "why a human pilot is needed at all."

3. Josh Wolfe: Technical precision equals moral precision. From one-on-one hand-to-hand combat to nuclear weapons to precision guidance, the ability to distinguish combatants from non-combatants is itself moral progress — "greater technical precision brings greater moral precision."

4. T2: AI autonomous systems may be the ultimate deterrent — making war between nations futile. If an adversary knows that even after communications and satellites are destroyed, AI can still execute retaliatory strikes, the cost of war becomes unacceptable.

5. Josh Wolfe: TSMC is "the most geopolitically important company in the world," and U.S. chip manufacturing capacity is a "fatal weakness." The Arizona fab produces 20,000 wafers per month, while global demand is 12 million — "like a speck of dust on a chip."

6. Josh Wolfe: The future of semiconductors lies in "chiplets" — modular, distributed architectures. This could become a disruption opportunity on the scale of "TSMC 20 years ago," lowering design costs and improving yields.

7. T2: The U.S. spends $13 billion on an aircraft carrier, but China can destroy it with a hybrid weapons system (hypersonic glide vehicles, etc.). "Why keep buying more carriers?" He questions the allocation of the U.S. defense budget — it should shift toward autonomous systems, AI, and space capabilities.

8. Josh Wolfe: The disconnect between Silicon Valley and defense is a historical anomaly — Silicon Valley's roots lie in 40 years of electronic warfare and defense contracts. In 1956, Silicon Valley's first IPO was Varian (military microwave tubes), and Fairchild Semiconductor's first revenue came from a military contract.