The Impact Of The Semiconductor Shortages

You’ve probably read in the press about the shortage of semiconductors and the debate about when we might get back to normal. First let’s learn the “lingo.” A semiconductor chip is “an electronic circuit whose components, such as transistors and resistors are etched or deposited on a single slice of semiconductor material to produce a microchip.” Also referred to as a “chip.”

Materials used may be silicon, germanium and gallium arsenide. Microchips are usually less than a square inch and contain billions of transistors. Think of chips as tiny brains. They can be found in cars, computers, hospital ventilators, military jets, etc. In fact, 12% of our Gross Domestic Product (GDP) uses semiconductors. Goldman Sachs estimates that 169 different industries need microchips.

The supply chain is lengthy and complex. At one time there were 25 companies making high end chips but now there are three. Only Intel is headquartered in the U.S. Asia is the location of 75% of all semiconductor manufacturers.

Semiconductor Manufacturing
It is expensive to build a chip plant – called fabrication plants or “fabs.” Intel’s Arizona fab cost $10 billion. Plus, fabrication plants are specialized. You can’t just adjust the line for a different chip type. Apple and Microsoft buy from Taiwan Semiconductor Manufacturing Company (TSMC) because it is the only company that makes the sophisticated, high end chips they require for their products. (There are also low-end chips.) TSMC produces the most advanced semiconductors – 30% faster and more powerful than Intel’s.

Automobile Semiconductors
TSMC makes the chips for American cars. When the pandemic hit, car sales stalled. Car manufacturers cancelled their chip orders. Now, there is a large demand for cars. Dealers can’t keep them in stock. TSMC can’t produce new chips for the cars because the supply chain takes so long to gear up – seven to eight months.

Moving Forward
Global semiconductor sales in 2019 were $450 billion. Capital Group (parent of American Funds) forecasts this market growing to almost $1 trillion by 2030. Why?

In a word, DATA. Data is created all the time. It is generated through social media, pictures and videos, trip logs and more. In 2018, machines exceeded humans as the largest source of data creation. This will only increase. This data creation requires large resources of processing power. R&D by semiconductor companies is to find ways to put more transistors on smaller slices to meet this need.

According to Bloomberg Data, by 2025, the percentage breakdown of sectors requiring semiconductors will be:

• Intel is going to build two new fabs costing about $20 billion
• TSMC is building two additional fabs in Arizona investing $100 billion including more R&D – also in Arizona
• Samsung Electronics plans construction of a new facility in Texas – estimated cost $17 billion

The White House is proposing credits and incentives for R&D for semiconductors. The goal is to build smaller chips being able to hold more transistors in the U.S.