The Smart Money Was Always on David

Children hear the story David vs. Goliath as a great, and surprising, victory for the underdog. The truth is that David was the odds-on favorite from the outset. David could be sure of launching a stone at Goliath before the giant ever got within sword’s length. Even more important, if David missed, he could shoot again. The slingshot is an iterative weapon. The shooter learns as he goes. Misses provide essential data.

We are living in an age of David vs. Goliath, wherein the advantage has shifted decisively from mass and power to speed, and especially speed of learning from rapid iterating failure.

Why does SpaceX keep embarrassing NASA and Boeing? For those august institutions, failure is intolerable and feared. They put enormous resources into the avoidance of error. As a result, they launch less than 1% as often as Space X.

Launch Cadence, SpaceX vs. NASA

To be fair, the statistics for SpaceX versus NASA are not entirely apples to apples. NASA does not build rockets; it contracts them out. It often buys space in other vehicles, including even SpaceX’s Falcon. What the table illustrates is not so much the source of NASA’s failures, but rather of SpaceX’s success: frequent failure. Everything about the SpaceX program is oriented towards learning through failure. Nasa’s infrequent launches cost about $2-4 billion apiece. The advertised cost of SpaceX launches is less than $70 million, allowing SpaceX to consciously and deliberately exploit the learning curve. Every launch is a test of hardware. Every launch generates new telemetry data. The results of every launch feed back into design.

As a result, over the course of 15 years, crucial statistics—how many booster stage rockets can be reused and how frequently flights can be cadenced—have improved dramatically. In 2025, SpaceX managed one launch roughly every two to three days, and launched more orbital missions than NASA has launched heavy rockets in the past half-century.

SpaceX had some inherent advantages, the biggest being that most of its flights are unmanned, which makes tolerance of failure infinitely greater. Still, as a test of mass vs. speed of learning, the contrast is powerful. Goliath has a bigger budget, a vast array of contractors to do his bidding, great political support, and vast institutional scale. David has faster iteration and rapid feedback loops.

In our time we see this pattern repeatedly. Ukraine has fought Russia to a standstill, albeit at a terrible cost. If this remains a war of attrition, Russia may yet prevail. Still, Ukraine has already overperformed against fearful odds. Much of that is down to the difference between the two nations’ drone programs.

Both armies used drones extensively and for similar purposes. Drones are the best current example of asymmetric warfare wherein a weapon costing hundreds or a few thousand dollars can destroy tanks costing millions.

In Russia, however, the drone ecosystem remains positively Soviet, dominated by state-owned or state-aligned firms and Ministry of Defense procurement channels. Changes must be centrally approved. Losses are treated as embarrassments. Failures are denied and covered up rather than valued as learning opportunities. Design updates are tied to procurement cycles, not field improvisation.

Ukraine’s drone system modifications are driven not by headquarters committees, but by warriors in the field. Frequent failure becomes rapid learning. Redesigns are decentralized and often implemented within days. Losses are treated as inputs to redesign.

Reports from the Royal United Services Institute, the Institute for the Study of War, and the Center for Strategic and International Studies all affirm that Russian UAV systems tend to evolve through formal model updates rather than rapid field-level iteration. These same reports describe Ukraine’s drone ecosystem as decentralized and adaptive, with frontline feedback feeding quickly into design changes—highlighting how learning velocity, not inventory size, has become the decisive variable.

SpaceX’s advantage over traditional aerospace did not come from larger budgets or better initial designs, but from orders-of-magnitude faster learning driven by an unprecedented launch cadence. Ukraine’s drone advantage did not come from scale or inventory, but from learning speed: designs changed faster than Russia’s defenses could adapt, turning rapid iteration into battlefield power.

Crucial for investors today is to assess companies not only by the magnitude of their assets or even their current technology lead, but by their speed of learning.