The factories are funded, the subsidies are in place, and the chips are designed. What is missing, according to a report McKinsey published with the SEMI Foundation on September 17, is the people. The study projects that the U.S. semiconductor industry will face a talent gap of as many as 157,000 workers by 2030.
The number is the headline, but the report’s supporting figures tell the sharper story. Only 3 percent of American graduates entering engineering jobs each year choose the chip industry, and 73 percent of chip employers say they have clear difficulty hiring engineers. The pipeline, in other words, was not built to feed the industry Washington decided to rebuild.
Jon Taylor, an executive vice president in Samsung’s semiconductor unit in Austin, put the problem plainly to CNBC: the technical talent coming through the pipeline is simply not enough. His voice matters because Samsung is one of the companies spending billions on U.S. fabrication plants that will need workers the moment they are finished.
The mismatch has its origins in a decades-long drift. Chip manufacturing left the United States over a generation, and with it went the engineering culture, the training programs, and the habit of choosing a career in a factory that most Americans stopped thinking about. The CHIPS Act money arrived before the workforce did.
The buildout is real. Fabs are rising in Arizona, Texas, Ohio, and New York, financed by a mix of government subsidies and private capital, and each one is a years-long project that ends in a hiring requirement measured in the thousands. The industry has planned for the machines; it is only now confronting the people.
The talent gap is not uniform. The shortage is most acute in the roles that require both engineering depth and cleanroom experience: process engineers, equipment technicians, and the specialists who keep the fabrication lines running. Those skills cannot be hired off the shelf, and they take years to develop.
The report frames the problem as a recruiting failure as much as a training failure. The chip industry, its authors argue, has not made a case to young engineers that it is a growth industry with careers to match, and the perception gap is as wide as the skill gap. Students who might build a fab choose software instead, where the salaries and the stories are more visible.
The universities have responded, but slowly. Semiconductor programs are expanding, and the industry has begun to fund them, but a student who enrolls today will not be a process engineer for several years, and the factories are hiring now. The lag between the buildout and the workforce is the industry’s central problem.
The competition makes it harder. The same engineers the chip industry needs are wanted by defense contractors, by the AI companies building their own silicon, and by every manufacturer facing the same demographic squeeze. A scarce pool of technical talent is being bid up by more buyers than it can serve.
Analysts said the shortage is a risk to the entire reshoring project. A fab that cannot be staffed cannot run, and a fab that runs with undertrained staff yields poorly and costs more. The billions spent on the buildings will underperform if the people are not there to run them.
The report stops short of prescribing a single fix, but the direction is clear. The industry needs more students entering the pipeline, more mid-career workers retrained into it, and more immigration pathways for the skilled workers who already exist elsewhere. Each of those is a political choice as much as an industrial one.
The immigration question is the most sensitive. The U.S. chip industry has always depended on foreign-born engineers, and the current visa system makes it hard to keep the graduates the country itself has trained. The debate over whether to open the door wider is the one the industry is least eager to have in public.
The workforce issue also has a regional dimension. The new fabs are concentrated in states that must build housing, schools, and infrastructure for the workers they will import, and the communities are learning that a factory is a commitment that reaches far beyond the factory fence.
For Samsung, TSMC, Intel, and the other builders, the hiring is already underway. The companies are competing for the same small pool, and the early signs suggest wages and signing bonuses are rising, a preview of the broader labor market the report describes.
The 157,000 figure is an upper bound, and the report’s range reflects uncertainty about how many fabs actually get built and how much automation offsets the need for workers. The lower bound is still a large number, and either end of the range describes an industry short of the one input money cannot buy.
The contrast with the past is what gives the report its weight. The United States once trained the world’s semiconductor engineers and built the industry that trained them. The country is now rebuilding the industry and discovering it must rebuild the pipeline at the same time, and the two projects are not moving at the same speed.
The next several years will show whether the gap can be closed. The report is a warning dressed as a forecast, and its message is that the factories are the easy part. The hard part is convincing enough people that a career inside them is worth choosing.
