Legacy Process Nodes: Why 28nm and Older Remain Essential to the Industry

Written by

in

It is easy to assume, from the steady drumbeat of news about ever-smaller process nodes, that the semiconductor industry is in a constant race toward the smallest possible transistor, and that anything less than the leading edge is obsolete. In reality, the overwhelming majority of chips manufactured worldwide, by unit volume, are built on process nodes that are a decade or more old. Nodes like 28nm, 40nm, 65nm, and even older ones remain not just alive but essential, running at high utilization in fabs around the world. Understanding why requires looking past transistor density and toward the broader question of what a given chip actually needs to do.

Not Every Chip Needs to Be Cutting-Edge

Leading-edge process nodes are optimized for maximizing transistor density and minimizing power consumption in enormously complex, high-performance chips such as smartphone processors and data center accelerators. But a huge share of the world’s electronics does not need anything close to that level of transistor density. A microcontroller managing a car’s window motor, a power management chip regulating voltage in a household appliance, or a sensor interface chip in industrial equipment typically requires only a modest number of transistors, and the chip’s function depends far more on reliability, analog performance, or power handling than on raw digital processing density.

For these applications, a legacy node is not a compromise; it is often the better engineering choice. Mature nodes offer well-characterized, highly predictable transistor behavior refined over many years of production, lower manufacturing cost per chip, and design tools and intellectual property libraries that have been thoroughly validated across a huge range of previous designs.

Where Legacy Nodes Dominate

Automotive Electronics

Modern vehicles contain dozens to well over a hundred individual chips, controlling everything from engine timing to power windows to infotainment systems. Many of these chips, particularly power management, sensor interface, and basic microcontroller functions, are manufactured on mature nodes, often 40nm, 65nm, or older, chosen specifically because automotive electronics demand extremely high reliability over long operating lifetimes and wide temperature ranges, qualities that mature, well-understood process technology is better positioned to deliver than a newer, less proven node.

Power Management and Analog Chips

Analog and power management circuits, voltage regulators, battery management chips, and power delivery components, generally do not benefit from the transistor density improvements that drive demand for leading-edge digital logic. Many analog circuit techniques actually perform better, or are easier to design reliably, on larger transistor geometries, where device behavior is less sensitive to the kind of variability that becomes more pronounced at very small dimensions.

Microcontrollers and Embedded Systems

The vast and growing world of embedded electronics, industrial sensors, smart home devices, appliance controllers, and simple IoT endpoints typically runs on microcontrollers that need only modest computing power but enormous production volume at low cost. Legacy nodes, with their lower per-wafer manufacturing costs and fully depreciated equipment, are generally the most economical way to meet this demand.

Display Drivers and Specialty Chips

Chips that drive display panels, manage radio-frequency functions, or interface with specialized sensors often have requirements, high voltage tolerance, specific analog performance, or compatibility with non-standard materials and processes, that are better served by mature, specialized process technology than by the newest digital logic nodes.

The Supply Crunch That Changed Perceptions

For much of the 2010s, legacy nodes received relatively little public attention, overshadowed by coverage of leading-edge process milestones. That changed dramatically during the global semiconductor shortage of the early 2020s, when automakers and other manufacturers discovered that a shortage of comparatively simple, mature-node chips, rather than cutting-edge processors, was constraining production of entire vehicles and appliances. The shortage highlighted an important truth: a finished product often depends on dozens of different chips, and a single missing legacy-node component can halt production just as effectively as a missing leading-edge processor.

This episode drove renewed investment in legacy node manufacturing capacity, as both chipmakers and governments recognized that mature process technology was not a declining backwater of the industry but a critical and under-invested foundation for enormous segments of the global economy.

The Economics of Mature Nodes

Legacy node fabs benefit from a different economic logic than leading-edge facilities. The equipment used to manufacture mature nodes is typically older and has often been fully depreciated on a company’s books, meaning the ongoing cost of operating that equipment is much lower relative to its output than for a newer fab still recovering its initial capital investment. This allows mature-node chips to be manufactured profitably at much lower prices than leading-edge chips, which is essential given that many of these chips sell for a small fraction of the price of an advanced processor.

At the same time, demand for legacy node capacity has remained robust and, in many categories, has grown, driven by the proliferation of connected devices, the increasing electronic content in vehicles, and the ongoing expansion of industrial automation, all of which rely heavily on exactly the kind of chips mature nodes are well suited to produce.

A Durable Part of the Industry, Not a Relic

Far from fading away, legacy process technology continues to see fresh capital investment, including entirely new fabs built specifically to expand mature-node capacity, reflecting a recognition that demand for these chips is structural rather than temporary. Some of this investment goes toward modernizing mature-node fabs with larger wafer sizes or improved automation, even while keeping the core transistor technology at an older node, improving manufacturing efficiency without chasing further transistor density gains that would not benefit these applications anyway.

Conclusion

The semiconductor industry’s public narrative often centers on the relentless pursuit of smaller, denser, faster chips, but the reality of global electronics manufacturing rests just as heavily on mature, legacy process nodes that quietly power the automotive, industrial, and consumer electronics that fill everyday life. Far from obsolete, nodes like 28nm and older remain a durable, economically vital segment of the industry, one that the chip shortage of recent years made impossible to overlook.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *