Quality-system Control Modes and Technological Transitions in Six Global Semiconductor Firms

Purpose: Leading firms in the same industry espouse near-identical quality philosophies, yet their performance diverges starkly, a gap the field has not adequately explained. We examine it in six global semiconductor firms (Samsung, TSMC, Intel, Micron, Texas Instruments, and GlobalFoundries) and ask which dimension of a quality system still differentiates firms once the content of their declarations has converged.Methods: A longitudinal multiple-case study combines a thirty-year trajectory comparison (1996–2025) with a fully comparable five-year panel (2021–2025). Disclosures, quality policies, supplier handbooks, and academic case studies are triangulated to reconstruct each firm’s espoused quality philosophy and its espoused control mode, which are set against performance trajectories and a supplementary crisp-set QCA. The coding was independently replicated by a third coder (κ = 1.00) and subjected to sensitivity analysis.Results: The content of the six philosophies has converged, through institutional isomorphism, on “zero defects through variability elimination”; their espoused control modes have not. Three firms specify only the outcome to be achieved (TSMC, GlobalFoundries, Micron) and three specify the methods to be followed (Intel, TI, Samsung). This residual difference co-varies with performance, which splits into four trajectory types. Within the sample, outcome-specified control combined with committed transition resources is necessary and sufficient for the realization of technological transitions.Conclusion: Four propositions condense the evidence: espoused content converges (P1); espoused control modes remain divergent and carry the explanatory variance (P2); outcome-specified control permits but does not produce transition realization, becoming effective only with committed transition resources (P3); and accumulated quality capability conditions the conversion of technology and AI investment into performance, AI acting as a complementary asset rather than a direct cause (P4, exploratory). In isomorphic industries, what should be measured is not the content of quality declarations but their control mode.