Christopher Kirchhoff on Military Innovation and the Future of War

Christopher Kirchhoff, founding partner of the Pentagon’s Defense Innovation Unit and co-author of Unit X, talks with Tyler Cowen about drone warfare’s rapid takeover of modern battlefields, hypersonic weapons and EMP threats, the AI arms race with China, why post-Cold War defence consolidation produced slow-moving ‘utility companies’ instead of innovative firms, and why building a single unit inside the Pentagon to buy Silicon Valley technology faster required fighting bureaucratic resistance down to a dog-friendly office policy.

Key ideas

1. Drones have ended man-mechanised land warfare, not just changed a tactic. A quarter of the M1A1 Abrams tanks the US gave Ukraine were destroyed by Russian kamikaze drones; nearly 300 Ukrainian drone start-ups now out-iterate Western defence primes by rewriting targeting software overnight in response to shifting Russian jamming.
2. Hypersonic weapons break the assumptions built into nuclear deterrence. They defeat ballistic-trajectory-based launch detection by manoeuvring mid-flight and compress presidential decision time far below the 20–25 minute window an ICBM allows — with no current defence against them.
3. DIU’s breakthrough was procedural, not technological. A 29-year-old acquisition specialist’s discovery of an obscure ‘other transaction authority’ clause let DIU contract with start-ups in 30–60 days instead of 18–24 months, driving $70 billion in Pentagon technology acquisition since.
4. Post-Cold War consolidation turned defence primes into utility companies. Bill Perry’s 1993 push to shrink the defence-industrial base to five large firms left them so burdened by requirements and audit regimes that newer product-oriented entrants — Anduril, Shield AI, Palantir, SpaceX — are now winning contracts once assumed to go to the primes.
5. Reformers who work against institutional grain get punished, not protected. DIU’s second director withdrew his nomination as Pentagon acquisition chief after an internally cleared whistleblower complaint resurfaced in the media on the eve of his confirmation hearing — an inspector-general system Kirchhoff says is being ‘weaponised’ against effective public servants.

Content

Drone warfare and the end of mechanised land war

Kirchhoff opens by arguing that Ukraine has settled the ascendancy question: drone warfare is not a passing phase but the next generation of military technology, and the loss of a quarter of the US-supplied M1A1 Abrams fleet to Russian kamikaze drones marks the effective end of man-mechanised land warfare. What makes the Ukrainian drone ecosystem remarkable is not simple kamikaze drones but a full battle-management stack — surveillance drones that feed targeting data through open-source fusion software to attack drones — replicating, at a fraction of the cost, the kind of coordinated system the US Air Force fields. Nearly 300 Ukrainian start-ups can rewrite drone software overnight to counter shifting Russian electronic-warfare tactics, an iteration speed no Western defence contractor matches. Kirchhoff generalises the pattern beyond Ukraine: drones favour whoever is trying to defeat a conventional military operating conventionally — Hamas’s use of quadcopters to disable Gaza border-tower generators on 7 October, Houthi harassment of Red Sea shipping, and Hezbollah’s displacement of 85,000 Israeli civilians from northern Israel despite the IDF’s modernity. He rates US preparedness for a surprise drone attack on domestic soil as weak: there is no early-warning radar system tuned to detect small drones, the same gap that let the Chinese spy balloon go undetected.

Hypersonic weapons, EMP, and the collapse of warning time

Cowen presses Kirchhoff to unpack why hypersonic weapons unsettle deterrence built on mutual assured destruction. Two features matter: manoeuvrability — a hypersonic weapon can change course mid-flight, defeating the ballistic-trajectory assumption built into space-based missile-launch detection — and speed, which compresses the decision window far below the 20–25 minutes an ICBM allows and strains the entire continuity-of-government apparatus built around that timeline. There is currently no defence against a hypersonic strike. China has demonstrated fuller flight cycles in public testing than other states; Russia is pursuing a wider portfolio of exotic weapons, including a proposed nuclear payload mounted on a satellite to disable GPS and communications, and supercavitating torpedoes capable of destroying an aircraft-carrier group. On EMP, Kirchhoff distinguishes the threat from mass casualties: an electromagnetic pulse over an inhabited area would disable the electrical grid and most microelectronics in its radius — cars, gas pumps, computers — without the direct lethality of a conventional detonation, a Cold War-era weapon category tested extensively in the 1950s but never used by a major power.

The AI arms race, dual use, and China

Kirchhoff addresses the ‘Stop Killer Robots’ movement — Human Rights Watch and computer scientists including Stuart Russell — with respect for its motives but scepticism of its strategic logic: drawing on Cold War strategic-stability literature showing that nuclear proliferation among states has, on the whole, lessened great-power conflict, he argues that autonomous-weapons restraint cannot be verified the way nuclear or chemical restraint can, because it is software. Unilateral restraint would not produce mutual restraint; it would simply cede advantage — a lesson he thinks Ukraine’s experience confirms. On the US-China AI competition, he attributes the current US lead in generative AI to structural advantages (freer markets, less state interference) rather than superior state direction, while noting China’s Military-Civil Fusion doctrine channels dual-use commercial technology into the People’s Liberation Army without a dedicated acquisition unit like DIU, and without the domestic-censorship constraints that slow China’s population-facing AI deployment. He argues against proposals to nationalise frontier AI labs, contrasting national-labs culture unfavourably with commercial tech-sector dynamism, and favours instead close FBI and intelligence-community cooperation to protect private labs from nation-state theft — modelled on the Enduring Security Framework built for critical-infrastructure cybersecurity in the 2000s and 2010s. He is ‘firmly in favour of dual use to the absolute extent we can’, applying the same net-benefit logic he applies to encryption, and does not expect AI to replicate nuclear weapons’ sudden strategic discontinuity — he expects a more gradual, capability-graded diffusion that gives states time to find new equilibria.

Unit X — DIU’s founding and its procedural breakthrough

Asked to explain the book’s title, Kirchhoff traces DIU (Defense Innovation Unit Experimental) to Ash Carter’s 2001 article ‘Keeping the Edge’, which observed that federal R&D spending was flatlining exactly as the commercial technology economy took off; by the time Carter became Secretary of Defense in 2015, Apple, Microsoft, Google, and Amazon individually exceeded the entire US defence industry’s market capitalisation. Carter tasked Kirchhoff with designing DIU, then installed him and Raj Shah to run it starting in 2016. Its central innovation was procedural rather than technological: acquisition specialist Lauren Dailey, then 29, found a clause buried in the National Defense Authorization Act enabling ‘other transaction authority’ — an Apollo-era contracting mechanism letting DIU sign deals with start-ups in 30–60 days rather than the 18–24 months typical under standard federal acquisition rules. Kirchhoff and Shah took the idea to Washington within days and secured departmental sign-off within two weeks; $70 billion of Pentagon technology acquisition has used the mechanism since. DIU staffs on rotation, DARPA-style, deliberately to resist the institutional ossification Cowen presses him on — a pattern Kirchhoff frames as a recurring ‘Hegelian dialectic’ in defence-agency history, noting DARPA itself spent years in relative obscurity after its founding before becoming the dynamic institution it is today. China, by contrast, has no single DIU-equivalent unit; its Military-Civil Fusion doctrine instead mandates dual-use disclosure to the PLA across the whole commercial sector — a structurally different, state-directed route to the same end.

Defence-industry consolidation and the rise of product-oriented start-ups

Kirchhoff traces the current defence-industrial base to Bill Perry’s 1993 ‘Last Supper’, which pushed consolidation into five large primes to preserve an industrial base under shrinking post-Cold War budgets. The result, he argues, has been a disaster: those primes are now so burdened by requirements, auditing, and accounting regimes that they function more like utility companies than private firms, producing slow innovation and chronic cost overruns even where individual technology is impressive. A newer generation of product-oriented defence start-ups — Anduril, Shield AI, alongside Palantir and SpaceX — is not burdened by the same requirements architecture and is winning contracts once assumed to go to the primes: Anduril and General Atomics, not a traditional prime, are the two finalists for the Air Force’s 10,000-aircraft Collaborative Combat Aircraft programme under Secretary Frank Kendall, the largest doctrinal shift in Air Force history since the Wright Military Flyer. Kirchhoff’s prescription is not simply more defence spending but reallocation — away from demonstrably vulnerable legacy platforms (tanks, aircraft carriers) and towards this experimental, product-oriented ecosystem.

Accountability, whistleblowing, and the Mike Brown case

Kirchhoff describes executive authority inside government as having eroded to the point that even a lunch talk at a think tank required him, as a National Security Council aide, to file an ethics waiver over the value of the meal. His central case study is DIU’s second director, Mike Brown, former CEO of Symantec, nominated by President Biden to become the Pentagon’s top weapons buyer. A whistleblower complaint alleging Brown misused hiring authorities to pay Silicon Valley engineers near-market rates, and flagged a minor office-snack-fund accounting irregularity, was investigated by DIU’s general counsel, cleared in an 82-page report, and closed by the inspector general. It resurfaced in the media on the eve of Brown’s Senate confirmation hearing and forced his withdrawal despite no substantiated wrongdoing — an example, Kirchhoff argues, of an inspector-general system being weaponised against exactly the reformers government most needs. He recounts smaller instances of institutional resistance from DIU’s own founding, including government credit cards cut off out of spite before a first Washington trip, and a colonel (‘Colonel Fear’) threatening eviction over DIU’s dog-friendly office policy, resolved by reclassifying the office dogs as HIPAA-protected service animals.

Government commissions as an instrument of reform

Kirchhoff’s Cambridge political-science dissertation on government commissions — informed by early-career work on the Space Shuttle Columbia accident investigation and an official history of the Iraq reconstruction — taught him that expertise and ‘settled notions of how business should be done’ are frequently wrong, especially when they get people killed. He divides commissions into three types: disaster-management commissions, created mainly to deflect political blame and rarely effective; agenda commissions, like the National Security Commission on Artificial Intelligence, built to concentrate expertise and bipartisan support on a complex ongoing issue rather than respond to a single crisis; and crisis commissions, like the 9/11 Commission, created after an acknowledged rupture, when openness to fresh diagnosis is highest and recommendation-adoption rates are correspondingly high. He remains a strong advocate for commissions as a form of government, noting the British system uses them more freely than the American one.

See also

• Christopher Kirchhoff — speaker
• Tyler Cowen — host
• Unit X — the book this episode discusses
• Jennifer Pahlka on Reforming Government — the bureaucratic rigidity cycle diagnosed from the civilian-government side, naming DIU as its own case study