The global semiconductor shortage that peaked in 2021 and 2022 wasn’t caused by a single event. It was the result of several forces colliding at once: a pandemic that shut down factories and supercharged demand for electronics, a manufacturing process too complex to scale quickly, supply chains designed to carry almost no spare inventory, and a concentration of production in a handful of facilities. The shortage cost the auto industry alone an estimated $210 billion in lost revenue in 2021, with ripple effects reaching consumer electronics, medical devices, and industrial equipment.
COVID-19 Hit Supply and Demand Simultaneously
When governments worldwide imposed lockdowns in early 2020, semiconductor factories in Asia were among the first to feel the impact. Strict containment measures disrupted operations at fabrication plants across Taiwan, South Korea, Malaysia, and China. A 2021 survey by the Institute for Supply Management found that 97% of organizations experienced disruptions in supply availability, production capacity, lead times, and transportation following the pandemic. The world’s 1,000 largest corporations had more than 12,000 facilities, stores, and operations located in quarantine zones by early March 2020.
At the same time, demand for chips surged in directions no one had forecasted. Millions of people suddenly needed laptops, webcams, monitors, and networking equipment to work and learn from home. Data centers expanded to handle the explosion in cloud computing, video streaming, and online shopping. Gaming console sales spiked. Automakers, who had cancelled chip orders early in the pandemic expecting a sales slump, came back to the market only to find their place in line had been taken by consumer electronics companies willing to pay more.
Chips Take Months to Make
One reason the shortage couldn’t be resolved quickly is that semiconductor manufacturing is extraordinarily time-consuming. The average manufacturing cycle runs 12 weeks to more than six months, depending on the complexity of the chip. After fabrication, final assembly and testing adds another 4 to 8 weeks. A simple component like a diode might require 6 to 10 major processing steps, while a high-performance processor or memory chip can require over 100. Each step must be executed with extreme precision on equipment that costs hundreds of millions of dollars.
Building a new fabrication plant is even slower. A modern fab takes three to five years to plan, construct, and bring to full production, and costs upward of $10 billion. That means when demand spikes, chipmakers can’t simply flip a switch. They’re locked into capacity decisions made years earlier. Even running existing facilities at maximum output, the industry couldn’t close the gap between what was needed and what could be produced.
Just-in-Time Inventory Left No Buffer
For decades, manufacturers across industries relied on “just-in-time” inventory management, keeping only the chips they needed for immediate production and ordering more as demand dictated. This approach minimized warehousing costs and waste, but it also meant there was virtually no cushion when supply dried up. When chip deliveries slowed in late 2020, companies had no stockpiles to fall back on.
The pain was most visible in the auto industry. BMW, Volkswagen, Ford, and Toyota all cut output or temporarily closed assembly lines because they couldn’t get enough microchips. AlixPartners estimated that 7.7 million vehicles of production were lost globally in 2021 alone. The crisis forced a dramatic rethinking of inventory strategy. A McKinsey report found that 61% of firms responded by increasing inventory, diversifying suppliers, or localizing production networks. In the UK, roughly 84% of firms planned to move away from just-in-time practices in favor of “just-in-case” models that maintain safety stock and backup suppliers. Even Toyota, the company that pioneered just-in-time manufacturing, began building larger inventories of electronic parts. Huawei took it further, stockpiling chips for its telecom equipment and smart devices.
Disasters Compounded the Problem
The pandemic alone would have been enough to create a shortage, but a series of additional disruptions made it worse. On March 19, 2021, a fire broke out at a Renesas Electronics factory in Naka City, Japan. Renesas is one of the world’s largest suppliers of microcontrollers used in cars. The fire halted production at the plant, and because Renesas chips were already in short supply, the shutdown forced automakers around the world to further suspend or scale back production.
A severe winter storm in Texas in February 2021 knocked out power to several Samsung and NXP semiconductor facilities for weeks. A drought in Taiwan threatened the enormous volumes of ultrapure water that chip fabrication requires. Each of these events individually would have been manageable under normal conditions, but stacked on top of a global shortage, they amplified delays that cascaded through supply chains for months.
Geographic Concentration Created Fragility
The semiconductor supply chain is remarkably concentrated. TSMC, based in Taiwan, produces about 64% of the world’s advanced logic chips (those made at the smallest, most cutting-edge process nodes of 7 nanometers and below). While that figure is often exaggerated to 90% in political debates, 64% is still an extraordinary concentration of critical manufacturing in one company on one island. Any disruption to TSMC’s operations, whether from a natural disaster, a geopolitical crisis, or an equipment failure, would ripple through the global electronics industry.
Raw materials showed similar concentration risks. As of 2022, Ukraine supplied an estimated 50% of the world’s neon gas, and the semiconductor industry accounts for up to 90% of neon laser demand. Neon is essential for the lithography process that etches circuit patterns onto silicon wafers. When Russia invaded Ukraine in early 2022, it raised immediate concerns about neon supply, adding yet another layer of uncertainty to an already strained market. Ukraine’s dominance in neon production traces back to the Soviet-era steel industry, where neon is a byproduct of oxygen separation in steelmaking.
Older Chips Were Hit Hardest
Much of the public conversation focused on cutting-edge processors, but the most painful shortages actually hit mature, less advanced chips. These are components made on older manufacturing nodes (often 28 nanometers and above) that go into cars, appliances, industrial controls, and power management systems. They’re inexpensive per unit, which means the profit margins don’t justify building new factories dedicated to producing them. Chipmakers have little financial incentive to expand capacity for parts that sell for a few dollars each when they could allocate that investment toward advanced chips that sell for far more.
This created a frustrating dynamic: automakers couldn’t get the $1 microcontrollers needed to complete vehicles that sell for $40,000. The scarcity of these mature chips forced some auto assembly lines to shut down well into 2022, long after supplies of more advanced processors had begun to stabilize.
How the Industry Is Responding
Governments and companies have poured enormous resources into expanding semiconductor manufacturing capacity. The SEMI World Fab Forecast, published in December 2024, lists more than 1,500 semiconductor facilities and production lines globally, with 180 new volume facilities and lines expected to start operation in 2025 or later. The United States passed the CHIPS Act to subsidize domestic fab construction. The European Union, Japan, South Korea, and India have launched similar programs.
These investments take years to pay off. New fabs being built by TSMC in Arizona, Samsung in Texas, and Intel in Ohio won’t reach full production overnight. The semiconductor industry operates on long planning cycles, and the lesson of the shortage is that those cycles don’t compress easily, no matter how urgent the need. The companies that weathered the crisis best were those that had diversified their supplier base, maintained strategic reserves, or secured long-term supply agreements before the crunch hit.