A Bottleneck

The United States did what few countries in the world can afford. It mobilized tens of billions of dollars through the CHIPS Act, drew TSMC, Intel and Samsung to build plants in Arizona, Texas and Ohio, and raised chip reindustrialization to a matter of state policy with bipartisan backing. The capital was available and the political will was aligned.
This week's Bloomberg report documents the variable that fell behind: people. The Semiconductor Industry Association projected in 2023 a shortfall of 67,000 workers by 2030; the study now circulating, in July 2026, raises the figure to 157,000 vacant skilled positions. Engineers, cleanroom technicians, process specialists. The vital input that no budget allocation can materialize within a fiscal cycle.
This figure deserves a careful look from the Caribbean, because it coincides with a local move. The Dominican Republic has its National Strategy for the Promotion of the Semiconductor Industry, the ENFIS, backed by decree 324-24 and coordinated by the Ministry of Industry and Commerce, betting on capturing links in the chain that nearshoring is pulling out of Asia: design, assembly, testing, packaging, logistics.
And this month, once again, the emphasis of the public conversation returns to innovation. The word reappears with the regularity of a season. Every government, every summit, every plan reintroduces it as if it were debuting it, and the national narrative spins on its own axis while the international thermometer reads a different temperature. In our discourse, innovation behaves like potential energy: it is declared, spread, celebrated and turned into the apex of a pyramid of intentions. The important thing to keep in mind is that potential energy does work only when there is a medium to conduct it. Here, the narrative has become the medium, when it should be merely the announcement of the end.
In the semiconductor industry the real medium has a name and a gestation time: a cleanroom technician takes years to mature, and a process engineer requires more than a decade of accumulated training. The United States, as a nation, recently discovered that it can buy the factory and finance the machine, and still run into the one variable that demands biological patience. The scarcity that burdens its revival is one of trained biographies.
Our country recycles its narrative apart from that context and runs a precise risk: it repeats the right word and dodges the right question. If the world's largest economy, with its university muscle and its national laboratories, faces a six-figure shortfall in the workforce that ENFIS would also need, Dominican competitive advantage is decided in the variable that enrollment exposes: the proportion and speed with which the education system can train its people in STEM.
We now face terrain different from the link we aspire to capture. This one will be decided by the density of people able to sustain it, and that density is built in the technical education system, in the polytechnic schools, in the engineering curricula, in the continuing education of a workforce that today is counted in thousands per year. Real performance lies in the accumulated figure: between 2012 and 2023 the country trained 90,740 STEM professionals, according to MESCyT, an average close to 8,700 per year, while these programs make up less than 17.5% of Dominican university enrollment. The shortfall the United States projects for a single decade, 157,000 people, comfortably exceeds everything our ecosystem trained in the last twelve years.
It helps to ground that scale in a concrete case. ITLA, a public talent pipeline ideally suited to this technical profile, graduated its record class in 2026: 642 professionals, according to the Ministry of the Presidency. In twenty-five years of existence it has trained 7,161 graduates. A sustained institutional achievement that, set beside global demand, measures exactly the distance between what we produce and what the chip ambition requires.
Here an asymmetry is at work: innovation understood as a declarative act is cheap and runs out fast; a speech is enough to produce it and a legislative term is enough to forget it. And there is plenty of that on every institution's social media.
Installed talent capacity is expensive, slow and cumulative, and for exactly that reason it pays off for generations. A country accumulates talent the way an organism accumulates immunological memory: exposure after exposure, cohort after cohort, until the learned response becomes structure.
The semiconductor industry punishes economies that confuse the two registers with particular harshness, because its value chain tolerates very little improvisation. Every test station, every packaging line, every design lab needs trained hands before the first machine is switched on.
The public entrepreneurs who designed ENFIS know this distinction; the technical evidence is in the document. The challenge is for the national narrative to internalize it before the cycle dilutes it again into a new slogan.
And here a geopolitical observation is in order, one that usually gets left out of the enthusiasm. The same nearshoring wave that opens the Dominican window opens simultaneous windows in Mexico, Costa Rica, Vietnam, India and Malaysia, all competing for the same scarce talent.
When global demand for technicians outstrips world supply for half a decade, the country that graduates faster captures the investment, and the one that graduates slower watches the factory settle in another jurisdiction. Read calmly, the American shortage works as a price signal to be paid: it announces that the most contested asset of this decade is the trained person, and that its value will keep rising as long as the gap persists.
The useful horizon then becomes quantitative. The mature question stops being how much we celebrate innovation and becomes how many technicians we graduate per year, with what curriculum, aligned with which link in the chain, and whether that curve comes anywhere near the one the strategy itself presupposes. A serious industrial plan includes an explicit training target, a schedule of graduates, and a stable alliance among universities, free trade zones and the multilateral organizations already supporting the process. That scaffolding turns the word into capacity, and capacity into jobs that last.
The United States has just shown, with its checkbook open, that the bottleneck of this era is cultivated person by person, cohort by cohort. Reading that data point in time marks the difference between an industrial policy with roots and one more season of the same word that every year is presented as new.
In global indicators, whoever does the most wins, ahead of whoever ‘posts’ the most.