Abstract

Successive US defense strategies stake American advantage on adapting faster than the rival, yet the competition with China is still measured chiefly in platforms and inventories. This essay argues that strategic competition is, at its base, a contest of institutional learning rates, and that the United States is on course to lose that contest not for lack of spending but for lack of capture: the federal enterprise cannot convert what it trains into durable, applied capability. It introduces Compound Learning, an analytic framework that treats workforce development, learning architecture, and organizational transformation as three multiplicative inputs to a repeating cycle, and it derives a stability condition that explains why incremental training is not merely insufficient but a managed retreat. The framework offers defense planners a small set of net-assessment metrics for the workforce dimension of the pacing problem and a concrete agenda for force development.

The Learning Race Starts Now

The 2018 National Defense Strategy declared that success in war "will increasingly depend on how well each side can integrate" new capabilities and adapt its way of fighting, warning that the Department had to deliver performance "at the speed of relevance."¹ Its successors retained the judgment and named the benchmark: the People's Republic of China is the Department's pacing challenge, the competitor against which American capability is measured.² These are, on their face, claims about hardware and tempo. Read more carefully, they are claims about learning.

A force that integrates new capability faster, fights in new ways sooner, and retains those gains longer is a force that learns faster than its rival. The strategy documents have correctly identified the dependent variable. They have been less clear about the independent one. This essay makes a single argument: strategic competition with China is, at its foundation, a contest of institutional learning rates, and the United States is on a trajectory to lose that contest for reasons that have nothing to do with the size of its defense budget.

The competition is usually narrated in tonnage and inventories—shipbuilding capacity, magazine depth, hypersonic test cadence. Those measures matter. But they are downstream of a deeper variable: the rate at which each side converts investment in its people into capability that lands in the operating force and survives there. On that variable, the trend lines are not favorable, and the cause is not parsimony. It is capture. The American defense enterprise spends heavily on training, education, and talent, and it cannot reliably say what those expenditures buy. A competitor that captures more of what it builds will, over enough cycles, outpace a wealthier competitor that captures less. The arithmetic is unforgiving, and it does not reward the bigger budget. It rewards the higher rate.

Decision Advantage Is a Learning Rate

For more than a decade the joint force has organized its modernization around decision advantage—the condition, in the Department's own formulation, of being able "to make and implement decisions faster and more effectively than an adversary, based on superior data, information, and understanding."³ The concept is sound and its lineage is distinguished, running back through the observe-orient-decide-act loop to the recognition that tempo in decision is itself a weapon. The error lies not in the concept but in its operationalization.

The deeper variable is the institution's capacity to learn: to recognize new external knowledge, assimilate it, move it into applied practice, and retain it against the constant erosion of attrition, rotation, and adversary adaptation. Economists call this absorptive capacity.⁴ A command saturated with sensors but unable to absorb what they reveal does not hold decision advantage; it holds data.

This is the Red Queen condition, named by the evolutionary biologist Leigh Van Valen for the figure in Lewis Carroll who must run continuously merely to remain in place.⁵ In a coevolutionary contest, the relevant question is never whether an institution is improving in absolute terms. It is whether it is improving faster than its rival. A force that adds capability at a steady rate while its competitor compounds capability faces a gap that widens every cycle. Decision advantage, understood this way, is not a posture an enterprise adopts once. It is a rate it must sustain.

The Rival Is Building a Dynamic Learning System

The People's Republic of China is investing in the inputs to a national learning rate with a deliberateness that the platform-centric debate tends to obscure.

Begin with the raw stock of talent. Chinese universities were projected to graduate roughly 77,000 STEM doctorates per year by 2025, against approximately 40,000 in the United States.⁶ The quality tier is moving in the same direction. By the Nature Index, China surpassed the United States for the top position in 2023 and widened its lead in 2024.⁷

More telling than the floor is the rate, and here Beijing is engineering the institution, not only the inputs. The Department's most recent assessment of Chinese military power describes the People's Liberation Army's pursuit of "intelligentized warfare"—the integration of artificial intelligence, big data, and autonomy across planning and operations—and its stated intent to accelerate the combined development of mechanization, informatization, and intelligentization by 2027.⁸

A competitor that treats workforce design, education reform, and doctrinal learning as co-equal lines of modernization—on a par with shipbuilding—is engineering its own learning rate on purpose.

The American Absorption Gap

The United States does not have a training-investment problem. It has a capture problem. The Government Accountability Office (GAO) has repeatedly found the Department's professional military education and acquisition training programs deficient in assessing front-end competencies and back-end results.^{9 10} These are not findings about insufficient funding; they are findings about uncaptured value.

The applied literature on training transfer quantifies the leak that such measurement gaps conceal.¹¹ A force that does not capture and reinforce what it trains is not standing still; it is decaying at a measurable rate between training events. In a contest against a competitor that compounds, decay is defeat deferred.

A Framework for the Learning Race

Compound Learning offers a discipline for the workforce dimension of strategic competition. It treats institutional capability as the product of three engineered inputs operating on a repeating cycle:

1. Stock (S): The floor of the system—the cumulative capability resident in the people the institution already has.
2. Throughput (T): The rate at which new capability moves from instruction into applied practice.
3. Retention (R): The structural conditions—incentives, decision rights, and operating cadences—that let a behavior persist.

The decisive claim is that these inputs combine multiplicatively:

Capability(t) = S · μᵗ, where μ = 1 + α(S · T · R).

Because the inputs multiply, a near-zero value in any one of them collapses the multiplier toward 1.0.

Genuine compounding requires that μ > 1 + λ, where λ is the decay rate. If the per-cycle multiplier does not exceed the decay rate, the institution is conducting a managed retreat.

Conclusion

Successive defense strategies have located American advantage in the ability to adapt faster than the rival. But adaptation is not an attitude; it is an institutional rate. China is engineering its learning rate as a deliberate line of national effort. The United States is allowing its own to be set by default. The pacing challenge is a contest of learning rates, and the arithmetic does not care which side spends more. It cares which side keeps more of what it builds, and how fast.