A post I wrote a few weeks ago about the semiconductor talent shortage picked up more traction than almost anything else I've posted this year, and the most useful thing to come out of it wasn't the reach. It was the comments. Reading through them, what struck me wasn't any one correction, it was how narrow the general understanding of this industry seems to be, mine included until I looked closer. All the industry reports frame the shortage as an engineering talent problem, which is the version most people seem to carry around. The semiconductor supply chain runs through materials science, specialty chemicals, mechanical and electrical trades, equipment manufacturing, and a dozen adjacent skill sets before a chip ever gets designed. Calling it an "engineer shortage" flattens a much more complicated picture.

That instinct held up. The reporting backs it up once you read past the headline number.

Where the headline number actually comes from

The figure making the rounds this month, a shortfall of 127,000 to 157,000 US semiconductor workers by 2030, comes from a McKinsey & Company analysis done with the SEMI Foundation and the National Network for Microelectronics Education (NNME), reported by CNBC C and other outlets on September 17, 2026. It's worth noting this number has been climbing. McKinsey's original 2024 semiconductor talent research put the US gap at "more than 100,000 engineers" by 2030, with Europe similarly short and Asia-Pacific needing over 200,000. A SEMI-backed estimate from mid-2025 put the US-specific number at around 67,000 workers, as part of a broader estimate that the global industry would need roughly one million additional skilled workers by 2030. The newest figure, 127,000 to 157,000, is technicians and engineers combined, and it's nearly double the regional estimate from just a year earlier. Reshoring has moved faster than the training pipeline has caught up.

That escalation is itself a data point: this isn't a stable, well-understood gap. It's a moving target that keeps getting revised upward as CHIPS Act-funded fabs actually break ground.

It's not one shortage, it's several

Read the reports closely and the "worker shortage" splits into pieces that don't share a training pipeline:

Chip design and process engineers. This is the piece that gets almost all the media and policy attention. Only about 3% of US engineering graduates take a job in the semiconductor industry, a figure reported by multiple outlets covering the September 2026 analysis. Process engineer roles typically require a bachelor's degree, specific coursework, and 18 to 36 months of on-the-job ramp time before someone is fully productive, according to workforce data compiled by staffing firm AMTEC using BLS employment figures for the sector. Entry-level process engineers start above $90,000.

Fab technicians. This is a larger, less visible gap. The work is physical and equipment-specific: photolithography, chemical-mechanical planarization (polishing wafers to atomic-level flatness), thin-film deposition, ion implantation. None of it requires a four-year degree, but it does require hands-on training on production equipment that most community colleges didn't offer until very recently. The Semiconductor Industry Association's 2026 workforce blueprint estimates roughly 60% of new manufacturing jobs in the industry won't require a four-year degree at all. The SEMI Foundation is only now building the standardized credential for this: a national Equipment Maintenance Technician Certification, piloting regionally in spring 2026 and rolling out nationally in fall 2026, developed with educators from 12 community colleges. Entry-level fab technicians start above $50,000.

Skilled trades. Someone has to build the fabs before anyone can staff them. Deloitte's semiconductor workforce analysis flags a need for electricians, pipefitters, and welders specifically tied to new facility construction, a category that barely shows up in "semiconductor jobs" coverage because it's construction labor, not chip labor, even though the industry can't function without it.

Equipment and materials. The tools that make chips are their own industry: companies like ASML (whose EUV lithography systems can cost upward of $200 million each), Tokyo Electron, and Onto Innovation need their own engineers and technicians. Materials and specialty chemicals, the rare-earth and chemical inputs that make wafers possible, are the least quantified part of the entire chain. I looked through McKinsey's, SEMI's, SIA's, and Deloitte's workforce reporting for this piece and none of them publish a shortage estimate specific to materials or chemicals supply. That's not proof there isn't a gap. It's more likely a sign that this layer isn't being tracked with the same rigor as the more visible parts of the chain, which is close to what the commenters on my original post were pointing out.

Assembly, test, and packaging (ATP). Deloitte puts ATP at roughly 15% of the global semiconductor workforce, spread across roughly 500 facilities worldwide, and it's rarely broken out separately in shortage coverage even though advanced packaging is becoming more technically demanding, not less, as chipmakers move to 3D stacking and heterogeneous integration.

The mismatch

Line these up and the imbalance is obvious. The four-year engineering degree pathway, the smallest slice of the actual job openings, has had a dedicated national program since 2022, when Purdue launched the country's first comprehensive Semiconductor Degrees Program. The technician pathway, which the SIA says covers roughly 60% of new manufacturing jobs, is only getting its first standardized national certification in 2026. The trades pathway barely gets discussed as part of "the semiconductor shortage" at all. And the materials and chemicals layer isn't being measured in a way that would let anyone say whether it's short-staffed or not.

None of this means the current approach, more engineering scholarships, university partnerships, CHIPS Act research funding, is wrong. It's aimed at a real gap. It's just aimed at the smallest and most visible one, while the larger technician and trades gaps are getting formal training infrastructure years later, and an entire layer of the supply chain isn't being measured at all.

Key lessons

  1. A single headline number usually hides several different problems. "157,000 worker shortage" sounds like one crisis. It's closer to five, each with a different training timeline, entry requirement, and starting salary.
  2. The most visible gap isn't always the biggest one. Engineering degrees get the university partnerships and the press coverage. Technician and trades roles make up more of the actual job count and are only now getting standardized training paths.
  3. What doesn't get measured doesn't get funded. Every major workforce report I reviewed for this piece had numbers for engineers and technicians. None of them had a specific figure for materials and specialty chemicals talent, a genuine ecosystem dependency that's effectively invisible in the current data.
  4. Escalating estimates are a signal, not noise. The US shortfall estimate roughly doubled from a 2025 regional figure to the 2026 McKinsey/SEMI Foundation number. That tells you the pipeline is falling further behind construction, not catching up.
  5. "More engineers" is necessary but not sufficient. Fixing this needs parallel investment across four-year degrees, technician certifications, trades apprenticeships, and better visibility into the materials and equipment layers, not just one pipeline scaled up.

Where this leaves the conversation

Reading the comments, and the questions people have sent me since, one thing is clear: a lot of people think of this industry as engineers who design chips. In reality it's bigger than that, and the reports back it up. The semiconductor shortage is a systems problem before it's a staffing problem. Job seekers, community colleges, and policymakers are mostly working from a four-year-degree angle, and that's the smallest part of the actual gap. The more honest starting point is the one the industry already knows: this is an ecosystem of design, process engineering, technician trades, equipment, materials, and packaging, and it needs to be measured and taught like one, not reduced to a single headline number.

That also means changing how the industry talks about itself. Not as one narrow path in, get an electrical engineering degree or don't bother, but as a set of genuinely different routes that all keep you close to the silicon: a trades apprenticeship into fab construction and maintenance, a community college certification into process or equipment technician work, a materials or chemistry background into the supply side, none of which require the same four years in the same classroom as the design track that gets all the attention.

References