A global shortage of semiconductors is expected to last until at least next year, which – along with geopolitical tensions, COVID-19 and other factors – has shone a spotlight on the vulnerability of the world’s semiconductor supply chain. While that’s encouraging various countries to beef up their domestic chipmaking capabilities – particularly China, which is already ahead of that particular curve – semiconductor players warn that self-sufficiency should restore balance to the global supply chain rather than replace it.
The global chip shortage initially hit the automotive sector hardest but is rapidly impacting other sectors, especially smartphone and PC makers. There are a number of factors contributing to the shortage, not least a massive spike in demand driven by the need for new PCs and devices due to the rise in work-from-home arrangements, as well as growth in the 5G and electric vehicle segments, according to S&P Global. Making matters worse is a spate of factory shutdowns earlier this year. Texas’ statewide power failures in February shut down fabs operated by Samsung Electronics, Infineon Technologies and NXP Semiconductors, while a fire in March closed a Japanese fab run by Renesas Electronics.
There’s also the issue of US export rules that currently prevent any company that uses US technology in any part of the chipmaking process from doing business with Huawei. That includes Taiwan Semiconductor Manufacturing (TSMC), whose C-level executives blame US policies for the current shortage.
In a conference call in April, TSMC chief C.C. Wei said the US crackdown on chip exports to China caused a rush to stockpile chipsets, “creating short-term imbalance in the supply chain”, according to the South China Morning Post. TSMC chair Mark Liu made similar comments to reporters at the end of March: “Uncertainties led to double booking, but actual capacity is larger than demand.”
China’s head start on self-sufficiency
Those polices have already spurred China to beef up its own chipset capabilities to rely less on Western companies, with the goal of developing self-sufficiency in its domestic chipset sector by 2025.
While China has a considerable gap to close to achieve that goal – particularly in terms of advanced chipmaking capabilities below 14nm – recent analyst reports expect China to succeed in that endeavor for all other chipset categories.
The 28-nm node is too big for smartphones, but it’s the most widely-used fab node for ICs and is in high demand for products such as cars and home appliances. With China having already established a presence at every part of the 28nm-technology industry chain, according to Li Ke of the China Semiconductor Industry Association, China’s 28-nm chips are expected to become mature this year.
China is also making progress with smaller nodes, with 14-nm chips produced by Semiconductor Manufacturing International Corporation (SMIC) becoming a growing source of revenue for the firm, and expected to become mature in 2022.
Meanwhile, last October, SMIC announced its FinFET N+1 process, which delivers 57% reduction in power consumption, a 63% reduction in logic area, 55% reduction in System on Chip area (55%), and a 20% gain in performance (all compared to SMIC’s 14-nm node). A Goldman Sachs report in July 2020 predicted SMIC would be able to start producing 7-nm chips by 2023.
The US export controls have been so disruptive that other markets are considering a similar route to protect their supply chains from the effects of geopolitical tensions. But that’s not necessarily a good idea, according to a new report from the US-based Semiconductor Industry Association and Boston Consulting Group.
The report estimates that replacing the global supply chain with self-contained local ecosystems would be disastrous for the semiconductor industry, requiring “at least $1 trillion in incremental upfront investment”, plus anywhere from $45 to $125 billion in incremental recurrent annual operational costs. This would push semiconductor prices up between 35% to 65% overall, which would have to be passed on to consumers for semiconductor players to stay in the black, the report says.
The report also finds that the disruption of global semiconductor supply chains has highlighted the fact that the geographical distribution of the supply chain is heavily skewed towards Asia. For example, literally all of the world’s most advanced semiconductor manufacturing capacity (i.e. below 10nm) is located in just two markets: Taiwan and South Korea.
It’s this overconcentration that makes the supply chain vulnerable not only to geopolitical disputes but also natural disasters, infrastructure shutdowns or other major disruptions. Building up semiconductor capabilities in other regions will help alleviate this, the report says, but with the goal of creating a more resilient global ecosystem, not simply doing everything yourself.
Regardless of how geographical overconcentration of the global chip supply chain is addressed, the chip shortage is the most immediate problem – and it’s one that’s not going away soon. TSMC and Intel have both expressed concerns that the shortage could extend well into 2022 if demand remains high. And while a change in US policy would be the ideal move, there’s no indication that the Biden administration is even considering that possibility for now.
Consequently, semiconductor players are planning sizable investments in new fabs to scale up capacity.
Intel has unveiled a $20 billion plan to build two new fabs in Arizona, while TSMC’s board has approved $2.89 billion for capacity expansion “for the purpose of installing mature technology capacity”. TSMC recently confirmed plans to invest $100 billion over the next three years to increase capacity at its plants.
SMIC is also investing in capacity expansion – in March, the company announced an agreement with the Shenzhen government to invest $2.35 billion to build a new wafer fab focused on 28-nm and above chips.