The disruption of supply chains has created a shortage of chips, forcing automobile makers to slow down production. How does the global chipmaking industry work, and how did similar shocks play out earlier?
Mercedes-Benz India has recently issued advertisements appealing to potential customers to bear with “a short waiting period” for vehicle deliveries. It’s an unusual request for the the German luxury carmaker — but it is not alone in asking customers to be patient.
Wait times for cars are currently among the longest in recent times, and a shortage of critical parts such as semiconductor chips are being cited as a key reason. Both global and Indian manufacturers have reported production issues linked to availability of chips, and been forced to apply brakes on production.
A common component across cars, electronic goods, medical devices, and smart appliances is the integrated circuits (IC) chip. ICs are the basis of all modern electronic equipment, and the chip is their brain and nerve centre. The circuit is essentially an aggregation of electronic components — resistors, transistors, capacitors — stuffed into a small silicon chip, and connected together to perform a single or multiple functions.
Before the Covid-19 pandemic, about a trillion chips were being manufactured on average globally every year. That’s about 120 chips per person — which would seem like a lot, considering only a small percentage of the global population uses high-end connected items containing many chips. But the chip shortage today has a lot to do with the pandemic — and how it changed consumer behaviour.
As country after country went into lockdown, people ended up buying more computers, phones, and gaming devices as they stayed at home. And since factories were shut, automakers cut down on chip purchases. This fundamentally reshaped supply and demand.
Now, as demand from automakers comes back on stream, chip companies are scrambling to adjust production and supplies to sectors such as automobiles.
Carmakers use chips in everything from power steering and fuel injector sensors to navigation systems and parking cameras. As cars get ‘smarter’ and more ‘connected’, electronic parts and components today account for 40% of the cost of a new internal combustion engine car, up from less than 20% two decades ago. Chips account for a bulk of this increase.
Why couldn’t carmakers stock up?
With just-in-time deliveries, carmakers typically kept low inventory holdings and relied on an electronics industry supply chain to feed production lines as per demand. There were two reasons for this: a steady decline in input prices and improvements in the processing power of chips. The number of transistors mounted in IC circuit chips has doubled every two years. This phenomenon, widely known as Moore’s Law, meant smaller devices with greater processing speeds.
But much of the progress in chip design has been on the non-auto side. The auto industry has by and large stuck to basic chips that have performed the same functions over the years — mostly aiding power steering or navigation assist. Given the auto industry’s traditional reliance on low-end chips, chipmakers on the cutting edge have focused on the non-auto segment. The basic products used in the car industry such as micro controllers are made under contract at older foundries.
Also, the auto industry only spends around $40 billion a year on chips — about a tenth of the global market. By comparison, just one company — Apple — spends more on chips just to make one product — iPhones.
How is the semiconductor market shaped?
The Big Three of chip manufacturing are Intel (US), Samsung (South Korea) and TSMC (Taiwan). Intel and Samsung are integrated device manufacturers who can design, manufacture and sell chips from end to end. TSMC is a foundry, which makes chips for companies that do not have their own factories or fabs. According to the consulting firm McKinsey, it costs at least $5.4 billion to build a facility with the latest 5-nanometre production lines. One example of the 5-nanometre chip is the A14 Bionic, which powers the iPhone 12. Those are made by TSMC.
Is the chip shortage something new?
Not really. Something similar to the 2021 “chip famine” has occurred at least four times in the last three decades, and the impact has been larger on each successive occasion. Among the reasons have been the Japan earthquake of 2011, and a surge in demand following the PC boom of 1994 and 1995. In almost all cases, shortages arose after the supply side failed to catch up with demand.
One of the earliest chip shortages — in 1988 — was blamed on a 1986 semiconductor trade agreement between the US and Japan. To stop Japanese manufacturers from dumping chips in the US, the Ronald Reagan administration coaxed Japan into a pact to help revive American manufacturing.
A March 12, 1988 report in The New York Times said the pact prohibited Japanese firms from selling below cost, and the Japanese government advised companies to limit output, “which would dry up the excess supply that had kept prices low”. Also, Japan, under American pressure, “became stricter in issuing export licenses for chips, slowing down exports”. While these moves helped curtail chip production and dumping stopped, “prices are [now] so high that the floor prices called for in the agreement are moot”, the report said.
Pulling the reins on the Japanese helped the Taiwanese semiconductor industry — which evolved from being job-shops for other firms into a multibillion dollar industry. Companies such as TSMC were direct gainers — according to TrendForce, TSMC commanded 54% of the global foundry revenues of $86.65 billion in 2020. Taiwan drew in 64%.
“A high dependency of the semiconductor suppliers on a single source in Taiwan for MCUs, combined with a general capacity constraint at IDMs (integrated device manufacturers) and foundry businesses, will cause shortages until the third quarter (July-September 2020). This is exacerbated by general infrastructure deficiencies on older semiconductor processes, as well as high demand for performance chips from adjacent industries,” IHS Markit said in a February 2021 report.
So what happens now?
Countries with or without a footprint in semiconductor manufacturing are working to secure their chip supplies. In May 2020, TSMC had announced a $12 billion factory in Arizona; earlier this month, in a move that is being seen as a challenge to TSMC and Samsung, Intel announced a $20 billion spending to build two factories in Arizona.
The Indian government has renewed nearly decade-old efforts to have companies build semiconductor fabrication facilities in India, or acquire semiconductor fabs outside India. The Centre’s earlier attempts at making India self-reliant for semiconductor fabs have largely been straitjacketed by the companies’ demands of cushioning the unavoidable high capital investments with various sops.