Palantir CTO Shyam Sankar launches 'Mobilize': America needs 800 days of weapons stockpile, has 8

Summary

Shyam Sankar, CTO of Palantir, argues that America faces a critical deterrence gap. The nation needs 800 days of weapons stockpile to credibly deter China but currently has only 8 days on hand. The shortfall reflects not a lack of spending—the U.S. spends $1 trillion annually on defense—but a failure of industrial capacity and organizational structure.

Sankar traces the problem to a historical break. In 1989, 94% of major weapon systems were built by dual-purpose manufacturers like Chrysler (missiles and minivans), Ford (satellites), and General Mills (torpedoes). Today, defense is dominated by specialized contractors run by financial engineering rather than founders. "Our entire industrial base was made up of founders," Sankar says. "What's happened? Financialization. Buybacks, dividends, financial engineering over real engineering."

The reindustrialization effort is real but early. Sankar estimates America is roughly 5% of the way toward the industrial capacity needed. The path forward requires both new entrants building entirely new classes of weapons and legacy manufacturers retooling existing capacity at speed. Retooling took 18 months in World War II, a timeline constraint that makes starting early critical. General Motors produces a new Escalade every 90 seconds. The nation needs to produce SM-6s, SM-3s, and Tomahawks at the same rate.

AI and worker productivity

Sankar rejects the framing that AI will displace workers. Instead, he argues AI is an opportunity to give American workers "superpowers"—making them 50 times more productive than workers anywhere else. A submarine parts manufacturer used AI to automate planning and quoting, reducing the process from hours to 10 minutes. Instead of idling during planning cycles, they added a third shift and hired more workers because demand increased.

At Panasonic Energy's battery factory inside Tesla's Gigafactory in Nevada, AI compressed training for battery technicians from three years of apprenticeship to three months. The workforce is composed of former casino workers, showing that reskilling is possible with the right tools and intent.

Sankar emphasizes that the future of AI will be written not by inventors but by users. Galileo did not invent the telescope; he used it to discover planetary motion. The impact of AI on society is determined by the people who wield it, and that's increasingly American factory workers and subject-matter experts who are not formally trained computer scientists.

Innovation against institutional resistance

Sankar draws historical parallels to innovation that succeeded despite institutional opposition. Andrew Higgins built the boats that won World War II—92% of all landing craft were Higgins boats—but the Navy initially refused to let him compete. Hyman Rickover's first office was a women's restroom; the Navy wanted to humiliate him into quitting. Billy Mitchell sunk a stationary ship from an aircraft without permission, proving air power was not just for communications but for warfare.

The pattern holds today. Colonel Kukor, the father of Maven, the Defense Department's most consequential AI system, faced institutional hostility. Services threatened by his work filed IG investigations against him and claimed he was "housing Iranians in his basement" despite being a devout Mormon with four daughters in a 1,400-square-foot home with no basement.

Sankar's core message to young people and innovators is to cultivate agency. "Do you really believe that your human effort can make a dent on the planet?" He argues that AI will reverse the managerial revolution of the 20th century. Power that was centralized in middle management can now flow back to frontline workers who know what they're doing, because bureaucracy is being automated away.

Strategy for defense startups

Sankar tells private defense tech companies they must hold two contradictory ideas. Run toward operational pain and real-world results, but recognize the Defense Department is not your buyer. Companies earning validation in actual conflict—he cites Shield's work in Ukraine—won't get paid for it immediately. That validation is what enables programs of record and long-term contracts. If you optimize only for bureaucratic procurement, you lose the heresy that drives innovation.

On fast-following versus invention, Sankar notes that Operation Paperclip offers the most consequential example. Stealing German technical papers failed completely, but acquiring Wernher von Braun and other Nazi engineers delivered ICBMs, the space program, and strategic offset. The lesson is primacy of winning; moral discomfort is subordinate to capability. A North Korean MiG defector allowed the U.S. to reverse-engineer and build its own fighter, restoring air dominance. "You don't wanna be like, hey, I didn't steal anything from the adversary, but I died nobly."

Sankar's book "Mobilize" makes the case for why the American industrial base must act now, why AI is a tool for worker agency rather than displacement, and why the heretics rather than the managers will determine whether America can rebuild the deterrence it needs.