NotedSource

The quiet erosion of scientific fluency inside industrial companies — and why it matters more than most R&D budget conversations do.

There’s a version of the innovation problem that gets talked about constantly: companies aren’t investing enough in R&D. Budget cuts, short-termism, the relentless pressure of quarterly earnings. That narrative is real, and the data supports it.

But there’s a second version of the problem that almost no one is talking about — and it’s the one that makes the first problem substantially worse. It isn’t just that companies have reduced what they spend on science. It’s that they’ve spent decades systematically reducing their ability to understand it.

That distinction matters. A lot.

 

What four decades of withdrawal actually produced

The departure of large corporations from basic scientific research is well documented. A landmark study published in the 

Strategic Management Journal by researchers at Duke, UEA, and Oxford — Arora, Belenzon, and Patacconi — tracked publications by company scientists across a range of industries between 1980 and 2006 and found a clear, sustained decline. The value attributable to scientific research inside firms dropped measurably over that period, even while patenting activity held steady. The authors described it precisely: large firms still valued “the golden eggs of science, as reflected in patents, but seem to be increasingly unwilling to invest in the golden goose itself.”

That framing — golden eggs without the goose — captures something important. Companies retained the commercialization infrastructure. They kept the IP counsel, the patent teams, the licensing apparatus. What they let atrophy was the scientific intelligence that tells you which eggs are worth incubating in the first place.

The shift wasn’t random. It tracked with globalization, competitive pressure from low-cost manufacturers, and the rise of a shareholder-value model that treated long-horizon basic research as an inefficiency. Over time, industrial labs that once employed substantial populations of PhDs — materials scientists, polymer chemists, metallurgists, corrosion engineers — were restructured around applied development teams focused on near-term commercial problems. Research that didn’t have a product attached to it within a reasonable planning window got defunded.

The strategic logic was coherent, even if the long-term consequences weren’t.

Source: Arora, Belenzon & Patacconi — Strategic Management Journal (2018)  |  NBER Working Paper w20902

The talent picture in industrials and materials

ManpowerGroup’s 2025 Talent Shortage Survey found that 72% of employers in the industrials and materials sector report being unable to find workers with the skills they need. That’s not primarily a manufacturing floor problem. It’s a research and engineering problem. The National Science Board has described what it calls an “accelerating STEM talent crisis,” specifically flagging chronic shortfalls in advanced manufacturing and materials research — areas where the supply of specialized expertise has not kept pace with demand even as overall STEM employment has grown.

In additive manufacturing specifically, 58% of business leaders reported in 2025 that limited access to skilled talent in materials science is constraining their ability to grow — up from 43% just the year before.

The semiconductor industry is facing a related version of this: roughly a third of the most experienced engineers, process technicians, and materials specialists are approaching retirement age, and the tacit knowledge they carry — the kind that doesn’t live in documentation — is leaving with them. Training new engineers to work at that level of specialization isn’t a course. It’s a decade.

What this reflects is a workforce shaped by decades of hiring decisions that prioritized application over research, engineering over science, product development over discovery. The people who could evaluate a novel alloy’s behavior under thermal cycling, or assess whether a new polymer formulation would meet long-term fatigue requirements, were not replaced when they left. In many cases, they weren’t even identified as critical until they were gone.

Source: ManpowerGroup 2025 Talent Shortage Survey  |  MRINetwork — Materials Science Talent Gap (2025)  |  National Science Board — STEM Talent Crisis  |  Sourceability — Semiconductor Talent Shortage

The absorptive capacity problem

There’s a concept in innovation research called absorptive capacity: the ability of a firm to recognize the value of new external information, assimilate it, and apply it to commercial ends. The foundational work on this, by economists Wesley Cohen and Daniel Levinthal, was published in 1990 — and it articulated something that has only become more relevant since.

Absorptive capacity is not passive. It doesn’t accumulate by exposure. It has to be maintained through active engagement with the underlying science — through hiring people who understand it, through internal R&D activity that keeps that understanding current, through the organizational practice of reading, evaluating, and applying external research. When you stop doing those things, you don’t just slow down. You lose the capacity to engage with the frontier at all.

This is where the withdrawal from basic research creates a second-order problem that most companies haven’t named correctly. The issue isn’t only that internal R&D has declined. It’s that declining internal R&D erodes the ability to use external R&D effectively. A company that has lost its population of senior materials scientists doesn’t just struggle to generate new discoveries internally. It struggles to evaluate discoveries that come from outside — from suppliers, from startups, from university research programs, from published literature. It can’t ask the right questions. It can’t assess technical claims. It can’t structure partnerships that would actually produce something commercially useful.

The academic literature on R&D outsourcing has documented this pattern empirically: firms that maintain internal scientific expertise extract significantly more value from external research relationships than firms that don’t. The translation layer — the internal expert who can bridge what a university lab has produced and what a commercial product needs — doesn’t disappear quietly. When it goes, the entire pipeline of external knowledge becomes much harder to access.

Source: Cohen & Levinthal — Absorptive Capacity (1990)  |  ScienceDirect — R&D Outsourcing & Absorptive Capacity

Why this matters specifically for materials and industrials

The materials and industrials sectors are facing a period of compounding scientific demands that make this erosion particularly costly.

Consider the convergence of pressures currently bearing down on industrial companies: decarbonization requirements that demand new material formulations and processing approaches; supply chain reshoring that requires domestic mastery of processes previously offshored; the rise of additive manufacturing that depends on materials science expertise to deploy effectively; and electrification across transportation and energy that is driving urgent demand for novel battery materials, thermal management compounds, and lightweight structural components.

These are not incremental engineering problems. They are research problems. They require understanding how new polymer composites behave under extended mechanical stress, how novel alloys perform across temperature ranges, how surface coatings interact with corrosive industrial environments over years of service life. They require the ability to evaluate what university labs and national laboratories are producing in materials science — and to commission, structure, and manage collaborative research that can move discoveries into commercial applications.

The global advanced materials market is on a trajectory toward roughly $580 billion by 2030, growing at a CAGR above 8% through the decade. The companies positioned to capture that growth are the ones that can translate emerging materials science into real products. That translation requires people who understand the science — not just the product specs.

The companies that have spent decades thinning their scientific staff are now trying to execute an extraordinarily technically demanding strategic pivot with organizational capabilities that were designed for a different era.

Source: PS Market Research — Advanced Materials Market  |  Deloitte 2025 Manufacturing Industry Outlook  |  DOE — Advanced Materials & Manufacturing

The access problem compounds the literacy problem

There’s a compounding factor that makes this harder: even companies that recognize the gap struggle to bridge it effectively.

University materials science departments, engineering schools, and national laboratories are actively working on many of the problems that industrial companies need solved. Research on corrosion-resistant coatings, high-temperature alloys, sustainable polymer systems, lightweight composites, and advanced battery materials is happening across dozens of institutions — often at precisely the depth and over precisely the timeline that internal commercial teams couldn’t sustain.

But finding the right researcher, establishing a working relationship, and structuring a collaboration that actually translates into commercial application is not straightforward. Industrial companies regularly report that navigating academic institutions is a significant barrier in itself — fragmented faculty directories, slow institutional processes, mismatched expectations about IP and timelines, and a fundamental absence of the relationship infrastructure that used to exist when companies ran large internal labs with meaningful academic connections.

A 2025 report exploring industry-university partnerships described the core problem simply: companies have a challenge, they have a topic, and their next step is figuring out how to identify the right people. That identification step — which sounds trivial — turns out to be a substantial friction point. For companies that have also lost the internal scientific fluency to even articulate what they’re looking for in technical terms, the friction is compounding.

Source: Research Information — The Hidden Challenge of Industry-University Partnerships (2025)  |  PMC — Systematic Review on University-Industry Collaboration Barriers

The path forward isn’t only more spending

The obvious prescription — spend more on R&D — is true but insufficient on its own.

The companies making the most meaningful progress on genuinely difficult materials and industrial research challenges share a different pattern: they are rebuilding scientific intelligence inside the organization, not just writing bigger checks to external parties. They are hiring researchers who can read the literature and evaluate what’s in it. They are maintaining active relationships with universities not as a one-time project exercise but as an ongoing function. And they are investing in the organizational capacity to translate external research into something their engineering teams can actually use.

That last piece — translation — is the underappreciated one. It’s not enough to have a relationship with a university materials lab if no one inside the company can evaluate what that lab is producing, understand how it maps to a commercial need, or communicate the commercial constraints back in a way a researcher can work with. The partnership only works if both sides can speak the language. And for a growing number of industrial companies, rebuilding that capacity is a more urgent priority than the budget conversation usually acknowledges.

The firms that figured this out first didn’t do it by accident. They recognized that the decades of scientific withdrawal had created not just a resource gap but a comprehension gap — and that closing it required investment in people and relationships, not just projects.

The ones still working through that recognition are running out of time to get there.

Source: Global Innovation Index 2025 — WIPO  |  MIT Sloan — Best Practices for Industry-University Collaboration