Summary

The 80-year security bargain is being repriced because Europe collected a vast “peace dividend” while the American taxpayer funded global order, and Washington has reached a “guns and butter moment.” Darren Farber sees the administration’s transactional stance as substantively justified: allies that did not fund the system weakened their claim on unlimited protection. Europe’s force-projection capacity is limited, its nuclear capability is largely strategic rather than tactical, and rebuilding will take years.
The Pentagon gets better technology when it adapts commercial products instead of specifying bespoke systems from scratch. Farber contrasts a bulky military handheld that lasted two hours with an iPhone app a Marine built over one weekend, and credits commercial-off-the-shelf legislation with helping Palantir fend off an Army replacement program funded at 5X Palantir’s level. For drones, the value may migrate from the $150 airframe toward coordination software, navigation and the shaped charge: “It’s not clear to me that making the drone will be this fabulous profit center.”
America cannot match China’s conventional mass, so Farber sees technological superiority as the path to winning, with capitalism and joint warfare as major current advantages. China has built stockpiles that Farber says would take decades of U.S. production to replicate and possesses “two hundred and fifty times our naval construction capacity today.” American companies have profit incentives, the armed services know how to fight jointly, and private markets can finance adaptation—advantages China might try to neutralize by handing command and control to AI.
Defending Taiwan conventionally may be prohibitively difficult, but occupying it could be harder than invading it. A former defense secretary answered “I don’t know” when Farber asked whether America has a moral obligation to intervene, suggesting political will may weaken when victory looks costly. Farber’s alternative defense scenario is a prepared, armed population exploiting Taiwan’s mountainous terrain as an insurgency; China’s likelier paths may instead involve an ambiguous false flag, political co-option or a slow “Hong Kong-esque” absorption without a shot fired.
Ukraine shows that warfare is neither a drone-only revolution nor a vindication of legacy hardware alone—it is “always a combination of old and new.” Javelins, HIMARS and continuously upgraded missile systems work, while several sophisticated new drones failed against a rapidly changing theater and Ukraine converged on locally printed, cheap platforms. Even if autonomous swarms dominate the opening exchange, taking territory still brings back troops, tanks and, as Gaza illustrates, “up-armored bulldozers.”
Defense venture may be directionally right yet financially resemble the fiber buildout of 1999–2000: strategically essential infrastructure paired with substantial capital destruction. The roughly $858 billion budget is not an open TAM; most funding is already assigned across programs and distinct “colors of money,” so displacing an incumbent can take far longer than venture underwriting assumes. The best openings are genuinely new requirements, while durable execution requires a long-tenured “responsible party,” not program owners rotating every 24 months.
In an industrial market, balance-sheet scale may matter more than having the cleverest prototype. Farber would generally back “the largest rounds, the largest raises, the largest balance sheets,” because winning ultimately means manufacturing millions of units through design and economic cycles. Anduril’s million-square-foot facility captures the logic: aggregated capital becomes industrial capacity, while undercapitalized companies can fail irrespective of their technology.
Farber’s preferred defense investment is a one-of-one input near the immutable physics of the system, purchased far below replacement cost and fortified through reinvestment. His former missile-fuel company had 10-year take-or-pay contracts and a Pentagon-estimated replacement cost of $2–$2.5 billion; war supplied an unpriced “toggle in it called war.” The governing principle is elemental: “My top of the mountain is to be as close to the molecule as possible,” supplying every competing platform rather than betting on which finished system wins.

Deep dive

1. The postwar security contract is being repriced

Farber frames the present as the end of an 80-year period in which Western borders generally did not change through force. His lawn-mowing analogy makes the administration’s logic explicit: if one party pays only sporadically, “who’s invalidated the contract?”
Europe built healthcare and education systems partly from a defense “peace dividend,” he argues, while the United States bore the enormous expense of protecting the free world. Now an executive tired of “freeloading” meets a national balance sheet that cannot sustain the same largesse: “We’ve reached the guns and butter moment.”
Farber says European defense capability and force projection are extremely limited, and its nuclear capability is concentrated in strategic ICBMs rather than tactical weapons. Europe can power up, he says, but “it’s going to take them a while.”
Patrick’s North Star question surfaces the risk in retrenchment: Washington traditionally planned to fight two simultaneous near-peer wars while maintaining world order. Farber accepts that a leadership void could produce chaos, but calls burden-sharing “a very necessary moment”; the Europeans gave Ukraine a loan guarantee and expected repayment, while America supplied money “no strings attached.”
The portfolio implications depend partly on what allies can build themselves. Farber points to AUKUS and Australia as a new strategic jump-off point for a China conflict, while the United States will likely retain slower, more exquisite capabilities such as aircraft carriers, nuclear submarines and the triad.

2. Commercial markets routinely outrun bespoke Pentagon development

Farber divides the Pentagon into a front office—warfighting, lethality and force projection—and a back office containing every business function. The back office has improved too slowly; the front office owns “pristine, exquisite weapon systems” but increasingly needs inexpensive autonomous mass.
His Afghanistan specimen is devastating: the military built a belt-mounted contracting device with an expensive battery, thermal printer and two-hour life. When the first iPhone appeared, a Marine built a superior app over a weekend and asked, “Why don’t we use this?”
Commercial-off-the-shelf-first legislation required the department to try existing products before developing military substitutes and limited demands for military-style cost visibility. Farber says it “enormously” helped Palantir when the Army was funding an internal alternative at 5X Palantir’s level and trying to turn Palantir off.
He expects commercial autonomy from companies such as Joby, Waymo and Tesla eventually to be adapted into defense applications. The department does best, he argues, when it lets adjacent commercial technology advance and then pulls it in.
Procurement is less uniformly cost-plus than its critics imply: Farber puts cost-plus at 15% of contract volume and roughly 30% of the budget, while the “supermajority” of contract volume is lowest-price-technically-acceptable. He calls DOGE a meritorious, roughly 20-year “extinction event”: unusual in public form, not unprecedented in substance.

3. A zero-based force would still mix exquisite systems with old weapons

Farber’s honest non-answer to Patrick’s zero-based-budget challenge is that doctrine has not resolved what next-generation warfare requires. Ukraine combines drones with 155-millimeter shells, small-caliber ammunition and mines; its minefield, he says, is roughly twice Florida’s size. “War is really a combination of old and new.”
He worries that exclusive preparation for China could leave America exposed to nonlinear threats, including non-state actors using advancing laser techniques to make fissile material. That does not justify preserving everything—the littoral combat ship is a failure—but it makes wholesale deletion of carriers, submarines and frigates reckless.
Enterprise carriers cost about $11 billion and take too long to build, yet Farber still sees a role for them. Five Israeli F-35s, he says, effectively eliminated Iran’s aerial defenses; at below $110 million per aircraft versus roughly $85 million for a Gulfstream 800, the unit economics look less absurd than detractors suggest, though logistics must improve.

4. America cannot out-mass China, but it retains institutional advantages

The two-war standard is no longer credible because “massing” means producing enough physical capability to win. China’s conventional stockpile would take decades of U.S. production to reproduce, Farber says, and its naval construction capacity is 250X America’s: “You’re not possibly gonna catch up” on a legitimate timetable.
His answer is technology plus capitalism. The department manufactures little itself, so it can tap private capital, commercial innovation and profit incentives; Farber contrasts that system with Chinese generals being jailed after ICBMs were found filled with water, and with a Chinese businessman’s observation that local entrepreneurs would rather access American customers than enter PLA defense work.
Joint warfare is another edge: Goldwater-Nichols forced senior American officers into cross-service assignments, creating integrated doctrine and practice. China has never fought a joint war, Farber argues, while low trust and corruption complicate command both within and across its service branches.
The inversion worries him: China’s institutional weakness may accelerate AI adoption. A dispassionate system could coordinate branches that do not trust one another, while America’s successful doctrine and “fighter pilot union” resist surrendering the yoke. The laggard institution may therefore have the stronger incentive to automate.

5. Taiwan may be invaded conventionally but defended as an insurgency

Farber thinks stopping a Chinese invasion would be “extremely hard.” When he asked a former defense secretary whether America had a moral obligation to defend Taiwan—not whether it could—the answer was “I don’t know.” He inferred that bipartisan will might exist for an easy victory but not for a costly conventional contest.
China can devastate Taiwan, but occupation requires preserving something worth governing. Farber points to America’s estimated 300–600 million firearms and asks how anyone could occupy it; if each Taiwanese adult had a long rifle, mountainous, lush terrain could support small groups “hiding in plain sight.”
Patrick’s supply-chain interruption matters: China could encircle and starve the island. Still, America spent 20 years being ground down by an underfunded Afghan insurgency, and distinguishing Taiwanese loyal to Beijing from resisters could become nearly impossible if the population supports resistance.
Farber rejects a binary forecast. He considers an ambiguous false flag and “cooperative peacekeeping force” more consonant with Chinese strategy than overt war; another increasingly discussed branch is political co-option through carrots, sticks and second-passport flight, followed by a gradual “Hong Kong-esque” degradation of autonomy.

6. Ukraine validates American missiles while commoditizing the drone

Farber calls Ukraine “a love letter to the American defense industrial base.” Baltic officials repeatedly ask him for Javelins, Taiwan has received HIMARS, and intercepted Russian communications warn, “They have the American stuff.” His prescription is more production and incentives that let primes drive unit costs down.
The drone airframe follows the flat-panel-TV arc: Ukrainians print platforms in theater for roughly $150, while value concentrates in coordinating software, navigation and shaped charges. Low-cost manufacturers can lead, so Farber doubts that building the drone itself becomes “this fabulous profit center.”
Several sophisticated drones from next-generation companies failed because the theater evolved too quickly. Farber links that fragility to DeepSeek: a novel approach can trigger an abrupt phase shift, and being first does not secure the franchise when wartime learning curves move faster than commercial ones.
Patrick tests the drone-swarm thesis with a battlefield no soldier can cross. Farber calls it a false choice: after one swarm wins with perhaps 75% casualties, occupation still requires old tools. Gaza’s subterranean defenses and Israel’s “unsung hero”—up-armored bulldozers followed by tanks—show why technology changes warfare’s opening, not its entire sequence.

7. Defense venture is constrained by budgets, time and accountability

Farber compares today’s defense enthusiasm with Global Crossing: the market correctly foresaw indispensable infrastructure, but the capital structure produced carnage. Next-generation defense may likewise be “a good directional view” paired with a poor weighing machine for deployment, timing and investor returns.
The headline defense budget is not an addressable market. Roughly $858 billion, excluding black ops or special appropriations in the example Farber cites, was already divided among programs and distinct “colors of money”; a startup can move quickly only when a requirement is genuinely new. Otherwise it spends investor capital arguing, “You’re wrong, and my taste is better than your requirement.”
The structural repair is a “responsible party” owning both fame and blame. Most uniformed leaders rotate after roughly 24 months and never see a program mature; the nuclear Navy instead uses roughly eight-year leadership billets descended from Admiral Rickover’s model and, Farber emphasizes, has never had an incident in its nuclear Navy.
Capital depth is industrial power. After raising heavily, Anduril announced a million-square-foot facility because the mission is not making one novel widget but millions at scale. Farber would generally choose “the largest rounds, the largest raises, the largest balance sheets,” while conceding that a true technological discontinuity could break the rule.

8. The primes are capable; their incentives are the binding constraint

Farber rejects the caricature of uniformly sclerotic primes. Lockheed is, in his telling, a top-five global patent owner with generations of engineering talent; give these companies permission to deliver their best judgment rather than satisfy overdetermined requirements, and “these guys can do a lot.”
Capital allocation is already formidable. He describes Northrop Grumman as compounding roughly 15% for 20–25 years and Lockheed’s CEO as someone who took American Tower from about $2 billion to $100 billion in market value over a decade.
Farber admires assets inside TransDigm and HEICO, but TransDigm’s roughly $3 billion of EBITDA at 50% margins creates downstream resentment. The ideal asset is “the cup holder in the Joint Strike Fighter”—too trivial to requalify—yet the primes’ institutional shareholders eventually notice when suppliers earn 2.5X the primes’ margin.
Scale therefore contains its own ceiling. Boeing once contemplated recapturing supply-chain margin and excluding expensive suppliers from future designs; as niche economics become material, vertical integration and new competitors appear. “When you’re Pluto and the market is Jupiter, it works. I just don’t know how long.”

9. The best defense annuities sit closest to the molecule

Farber looks for structural impediments to entry, not fashionable growth. His pink-diamond thesis was simply that the sole mine was running out and replacement supply was impossibly expensive; defense’s one-customer monopsony similarly creates natural one-of-one suppliers because factories rationally match only funded demand.
Albion bought the nation’s only missile-fuel producer, including its intermediate chemicals, for a small fraction of the Pentagon’s estimated $2–$2.5 billion replacement cost. Ten-year take-or-pay contracts supported immense leverage. The asset also carried contingent risk: the former plant was the site of the largest non-nuclear explosion in U.S. history, and the department contemplated moving the factory to the Utah desert.
The unpriced option was “a toggle in it called war”: rising danger increases consumption, but no conventional base/upside model charges the buyer for that outcome. Albion added commercial space-launch customers and sold after value was recognized—“probably too early,” consistent with Farber’s refrain, “We always sell too early.”
His firearm-accessories company can make 100,000 magazines daily, expanded share from 60% to 90%, and verticalized into nylon. Its product ran 300,000 rounds before testers exhausted their ammunition versus roughly 3,000 for the predecessor. Reinvestment simultaneously lowered government pricing and deepened a sole-source moat.

10. Network denial is the first fulcrum of a near-peer war

Net-centric warfare connects drones, ships and soldiers for shared awareness—but connectivity is also the kill switch. In what Farber recalls as MC02, a retired flag officer repeatedly won an expensive war game by destroying the network and reverting to smoke signals: “Everything turns into nothing.”
Starlink is his “unsung hero” of Ukraine; without its hard-to-crack communications and navigation layer, he believes the contingency would already have been lost. The next systems therefore need autonomy and inertial navigation sufficient to operate without GPS, ground links or a mesh network.
Farber’s best case is symmetrical capability producing another Cold War in which weapons become effectively unusable through mutual deterrence. His governing rule is unforgiving: “You just can’t be second,” because a temporary lead creates the temptation to flex.
The worst case is technology putting capabilities once reserved for nation-states into companies of 100 people while removing humans who might refuse to turn the key. Farber views some dangerous diffusion as nearly inevitable; governance must focus on controlling the physical substrates and molecules, while technological superiority remains “the best insurance policy” against plans that routinely fail.

March 2025 Fintwit Book Club: Diary of a Very Bad Year with Byne Hobart from The Diff Market Overview — April 1, 2025