Semiconductors: What Are They and Why Is There a Shortage?

It is widely known that there is a shortage of semiconductors on a global scale that is affecting production in a host of industries. Goldman Sachs says 169 industries are affected by the crunch. The poster child for the shortage seems to be the global auto manufacturing industry, where lost production affects tens of thousands of workers and a great deal of lost revenue.

Less publicized than the auto sector is the commercial vehicle industry, which is facing the same problem of strong underlying demand and a supply shortage of semiconductors that have restrained production. The auto manufacturers and commercial truck OEMs share the same supplier bases, so the problems of one industry can affect the other. Before the pandemic, there were rumblings of potential shortages, but the real shortage came when demand roared back in the second half of 2020. For context, global semiconductor sales declined between 2018 and 2019 but increased by 6.8% in 2020 according to the Semiconductor Industry Association. 2021 was the first year since 2018 that the sales number of chips exceeded a trillion. In 2021, 1.15 trillion semiconductors were sold, with the biggest growth among auto-grade chips that can withstand heat and other physical challenges. Sales for that segment rose 34% over the prior year at $26.4 billion. Unit sales rose 33%.

Demand for semiconductors is not just coming from vehicles, the trend towards digitalization ramped up by the pandemic will continue to drive demand higher. Not only cars and trucks, but phones, passports, credit cards, pencils and pens, radios, TVs, blenders, refrigerators, coffee machines, and even lightbulbs, get packed with semiconductors. Demand for consumer products is not the only factor for the shortage. There is a surge in demand from autos and commercial trucks to adopt technologies such as driver assistance systems and autonomous systems. Demand for electric cars and trucks is set to explode, and they use a lot more chips than their gasoline cousins. Car unit sales stumbled last year because of chip shortages, but EV sales were up. The transition to electric vehicles is underway. For commercial trucks, it is a slow start, but the wave is growing. The U.S. has deployed 1,215 zero-emissions vehicles Class 2 through Class 8 as of December, but that number could increase more than ten times within ten years, according to Calstart. There are about 140,000 pending orders for zero-emission trucks, expected to be delivered in the next decade. Models of zero-emission trucks have grown 625% since 2019, a sign the industry believes in the growing market.

Of the current deployments as of December 2021, more than half of the trucks (738) are in California, where state and local officials are aggressively asking the industry to begin the transition. Another 113 deployments are in New York, and 70 are in Texas. California’s role as a launching pad for zero-emission trucks is not surprising. The state wants fleets to cut emissions, with mandates for green-truck sales and operations, such as the California Air Resources Board’s Clean Air Act. The state has also made grants for battery-electric owners and hydrogen fueling stations. New York made a rule at the end of Feb-21, with the aim of shifting to zero-emission heavy-duty trucks. Overall, the national transition appears to be happening first in lighter trucks. Of the deployments, almost half (598) are in medium-duty trucks. Yard tractors have 237, and heavy-duty trucks number 47. Medium duty step vans have 203. Big fleets understand there is a public appetite for a Class 8 transition to zero-emissions units. In September, Schneider announced its plans to add 50 Freightliner Cascadias to intermodal operations in Southern California. The San Pedro Bay ports handle 40% of imported goods that flow into the U.S. The shorter port trips, from drayage to warehouses, are considered ideal candidates for electric trucks. The price of battery packs is projected to come down, as production ramps up. The total cost of ownership will be close to diesel trucks in a short period.

Several questions come to mind concerning semiconductors. What is a semiconductor anyway? Another key item related to the shortage is: What are the industry's near- and middle-term prospects? The answer to the first question is surprising. A semiconductor is a material that isn’t an insulator and isn’t a conductor. It is a hybrid material that, depending on the environment and situation, can act as both insulator and conductor. The most used semiconductor materials are silicon, germanium, and gallium arsenide. Of the three, germanium was one of the earliest semiconductor materials used and has four valence electrons in the outer shell of the atom. The number of valence atoms in a semiconductor material is important as it determines its conductivity. Germanium has largely fallen into disuse, in favor of the current king, silicon. Silicon has four valence atoms and melts at a higher temperature than germanium. The extraction, purification, and crystallization processes for silicon are both efficient and economical.

Gallium arsenide is the second most common semiconductor in use today. Unlike silicon and germanium, it is a compound, not an element and is made by combining gallium, with three valence atoms, with arsenic, which has five valence atoms. Eight valence atoms make gallium arsenide devices respond quickly to electric signals, suitable for amplifying the high-frequency signals seen in television signals. There are limitations: the compound is difficult to manufacture in mass quantities, unlike silicon, and the chemicals used are quite toxic. Silicon Dioxide has characteristics superior to silicon, allowing it to be used as an insulator, passivation layer, and a building layer in metal oxides. The future of silicon is coming into question as it is reaching the physical limit of Moore’s Law, but several other compounds have been developed to ease the industry’s path into the future. There are other prospects in the wing that may fundamentally change how we use semiconductors, like the development of quantum computers. Technology change in the electronic world moves fast.

With most industrial sectors reliant on electronic devices, the demand side of the semiconductor industry is relatively stable. The market for semiconductor materials reached $50 billion in 2018 and is projected to reach $70 billion by the end of 2025. While some materials in the production of semiconductors, like silicon, are inexpensive, the Rare Earth Materials (REE) used in the production of high-K dielectrics and chemical-mechanical polishing can be costly. Processes to separate REEs from the rock are difficult and costly, requiring thousands of stages to extract and purify the finished material. The difficulties have caused mining companies not to pursue their potential profits. China is one country that has focused on REE mining and refining, with the result that China produces 85% of the world’s supply of tungsten and molybdenum. China’s market dominance allows it to set prices and could be used as a political weapon in the future. REE deposits are found evenly across the earth, although it is hard to find in big deposits in one place. However, it takes a willingness to invest in mine extract and refine the materials to become self-sufficient in these materials.

In the beginning, the current semiconductor shortage began due to stronger demand for more advanced chips from the consumer electronics and computer industries. This was sparked because of the pandemic and the stay-at-home work conditions for millions of people. Demand for consumer products was not the only factor as demand for chips surged by the auto industry, driven by the adaption of new technology, kicked into a higher gear. The company Intel projects that semiconductors will account for over 20% of the input costs of a new premium car in the very near future, up from 4% in 2019. Global revenue from the semiconductor market increased by 25.1% in 2021. In this kind of demand surge, it is not surprising that manufacturers are struggling to keep up with supply. Given the supply problems with most other industrial commodities, it is no surprise the chip makers got caught short when demand roared back.

A vast majority of the world’s semiconductor production is being carried out by two companies in East Asia, Taiwan’s TSMC and South Korea’s Samsung. Their foundries account for as much as 70% of the world’s semiconductor production. To make matters worse is the high entry barriers into semiconductor manufacturing. It takes an upfront investment of $10-to-12 billion just to get started. Also, to get a plant up and running takes two-and-a-half to three years and, after all that investment, a new player would not match the incumbent's chip yields because of the steep learning curve. When you add things up, these kinds of upfront costs, while also considering the price pressures in the tech sector where a high percentage of risk to potential profit is, market conditions favor large companies who can spread costs across a large group of customers.

Help is coming. TSMC, the world’s largest semiconductor contract manufacturer, indicates it would spend about $100 billion over the next three years to bolster production capacity. Adding to that number are plans with Sony to build a factory in Japan to provide chips for cameras, cars, and other purposes. Samsung plans to spend $205 billion in the next three years, with a good portion going to its chip business. Intel is setting aside $95 billion over the next decade and aims to double down on contract manufacturing by building new chip-making facilities in Europe, and also wants to spend $20 billion to build two new factories in Arizona. When the semiconductor shortage became worse, the world’s political leaders recognized the need to have the world’s supply chains more diversified. As such, a large percentage of production is in potential conflict places, such as Taiwan. The U.S. only makes about 10% of the chips it uses, one of the reasons Joe Biden pledged his massive bipartisan tech funding bill of $52 billion.

The real trillion-dollar question is in the near term: When will the shortage ease? To answer that question, we have to differentiate between the different types of chips. Our industry of commercial trucks and trailers is closely associated with the global auto industry. They share the same suppliers and OEMs that make medium-duty trucks, who also produce light-duty trucks. They share the same demand for the same kinds of chips that work in adverse climate conditions. In a shortage, everybody competes with everybody, because capacity is limited. Unfortunately, the signs are still unclear as to when the shortage will end. Although most industry analysts see the semiconductor shortage easing slightly in the first half of 2022 with more easing in the second half of the year, a return to a post-pandemic level of supply and demand is not seen until 2023.

JPMorgan projects that 2023 will not see a supply shortage, but that 2022 will be a bit more tricky. Gokul Hariharan, co-head of Asia Pacific technology media and telecom research at JPMorgan, said, “Things could improve in the second half of the year as more supplies come online. But the first months could see pockets of shortage across the industry. There is capacity coming online, not just from the foundry companies, but also from the IDMs- integrated device manufacturers- companies. All the European and U.S. IDMs are also expanding their capacity and a lot of it is slated to come online from the middle of 2022 onwards.” Foundries are companies that are contracted by semiconductor firms to build chips. IDMs, on the other hand, are companies that design, manufacture, and sell chips. Although it is a long-term direction, the shortage of chips drove companies like Apple, Amazon, Meta (formerly Facebook), Tesla, and Baidu to start shunning established chipmakers and bringing certain aspects of development in-house. That won’t alleviate the near-term, but it might down the road.

A majority of the major automakers including General Motors, Ford, and Hyundai predict the shortage will ease in the second half of the year. However, automotive chip producers have forecast that the market will be tight and supplies will continue to be short throughout 2022. While the shortage will endure through 2022, Deloitte Global believes it will be less severe than in the fall of 2020 and most of 2021. Separately, a handful of industry experts believe the shortage will ease in the second half of the year after the automobile-related backlog is addressed. Deloitte Global believes the lead times were 20-52 weeks for multiple kinds of semiconductors, but by the end of 2022, those lead times will be close to 10-20 weeks and close to balance by early 2023. This implies the shortage will have lasted 24 months, similar to the 2008-09 chip shortage. In short, while some automakers are less convinced that the end of their shortage is in sight, there should be an improvement in the second half of the year.

Right now, most global auto manufacturers are operating on a restrained production plan. In the U.S., truck and trailer OEMs are scheduling orders according to their projections of production capability. All this suggests that the OEMs would love to have higher production rates but are proceeding with their capability for production. This also implies that underlying demand is stronger than the actual orders that are recorded. At this point, sales tell production information. With little inventory, customers are buying all the products that can be produced. January is a difficult month to draw comparisons on sales because of the large seasonal influence. In the U.S., car and light truck sales hit 15.2 million, the strongest SAAR since the 15.4 million level in June. However, actual sales fell shy of December but were consistent with Q4 levels. LMC Automotive believes the January sales numbers were strong enough to maintain their forecast of 15.9 million for the year. The company noted that the impact of the shortage appears to be past the Q3 peak of 800,000 units of lost production volume of North America. The current volume disruption in Q1 is projected to be at least half of the level of the peak, and they expect continued improvement for the year. Further improvements will bring the market to 16.9 million vehicles in 2023.

So far this year, there is not enough evidence to support the notion that deliveries of semiconductors are speeding up in a meaningful way for the OEMs of cars and trucks. Progress in semiconductor delivery will continue its slow incremental, but positive, trend. Production plans for transportation equipment seem to be a month-to-month decision for the time being. However, help seems to be on the way if the expected capacity promised for mid-year manifests itself. There are other items other than semiconductors in short supply. Labor is also a production weight. In many industries, plant managers are saying, “We now have the parts, but not the people to boost production.” Sooner, or later, the supply chains will catch up, but labor may prove to be an extended problem.