Report Shows Traditional Competition Metrics Misguided for Semiconductor Industry
PORTLAND, Ore. (Nov. 12, 2025) — The traditional antitrust frameworks used to analyze static markets are fundamentally inadequate for the task of evaluating the semiconductor industry’s competitive dynamics, a new white paper by scholars from the International Center for Law & Economics (ICLE) concludes.
Authored by David Teece of the University of California-Berkeley Haas School of Business, along with ICLE President Geoffrey A. Manne, Chief Economist Brian Albrecht, and Senior Scholar Mario Zúñiga, the paper notes that semiconductor markets are driven by the powerful and interconnected forces of Moore’s Law and Rock’s Law.
Moore’s Law, named for Intel Corp. cofounder Gordon Moore, predicts that the number of transistors in a dense integrated circuit will double every two years. Rock’s Law, named for famed Silicon Valley investor Arthur Rock, predicts that the cost to construct a modern semiconductor-chip fabrication plant will double every four years. Together, the authors write, this relentless pace of technological advancement and exponential rise in the cost of manufacturing capacity create a high-stakes, high-risk environment in which firms must continuously invest tens of billions of dollars to remain competitive, and where market leadership is just a temporary reward for successful innovation.
The paper analyzes this “competition for the market,” detailing how these unique economics have guided the industry’s evolution from vertically integrated device manufacturers to the modern fabless-foundry ecosystem, a specialized structure that promotes efficiency and democratizes chip design. The authors conclude that competition enforcers could be misled by static metrics into misdiagnosing the industry’s health, and that resultant policy interventions could inadvertently stifle the innovation that has powered the digital economy.
Manne adds:
“The semiconductor industry is a high-stakes competitive arena where firms make billion-dollar bets on unproven technologies years before demand is certain. The cadence of Moore’s Law forces companies to deliver a new generation of chips on a tight timeline, while Rock’s Law exponentially raises the financial ante for each new fab. This creates a recurring ‘race for the market,’ in which firms compete not through static price wars, but by making enormous, front-loaded investments to achieve a temporary technological lead. These temporary rewards, often misdiagnosed as monopoly profits, are in fact entrepreneurial returns necessary to justify and fund the next round of innovation. It is a healthy process, not a sign of a broken market.”
To speak with Geoffrey, contact Jim Fellinger at [email protected]. Download the paper here.
Moore’s and Rock’s Laws in Semiconductor Competition
- Moore’s Law (The Relentless Clock): the expectation of predictable, compounded gains in chip performance (doubling on a fixed cadence) forces an intense, relentless schedule of innovation, or “beats.” Falling behind by even a single beat can cost a firm its customers for years.
- Rock’s Law (The Exploding Cost): as chips get denser, the cost of manufacturing capacity rises exponentially.
- Data: the cost to build an advanced-fabrication plant (“fab”) now tops $10–20 billion. A single high numerical-aperture (NA) scanner using extreme ultraviolet (EUV) lithography—a critical piece of equipment—can cost more than $400 million.
- The Investment Stakes: leading firms like Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC) routinely commit to capital expenditures and R&D costs that total 40-60% of their annual revenue (peaking above 60% in 2021). These aren’t routine expenses; they are “existential bets.”