Chip shortages will be the norm in the future-EEWORLD

Collect

Without a $1 chip, the absence of one could hinder the shipment and sale of a more valuable device, appliance or vehicle, according to a new report from accounting and consulting firm Deloitte.

The semiconductor industry experienced one of its longest shortages, from spring 2020 to fall 2021, as demand surged during the COVID-19 pandemic and recovery. Deloitte expects it to last at least until 2022, driving shipment delays for some components into 2023.

The impact is still being felt in PCs, smartphones, data centers, gaming consoles, other consumer products, and especially the automotive industry. From 2020 to 2022, the cumulative global sales losses from shortages could exceed $500 billion.

The next semiconductor shortage could be as big or bigger than this one. The company said the economic damage could be even greater given the growing importance of chips to multiple industries. So it looked at what semiconductor manufacturers, distributors, customers (the semiconductor supply chain), and governments could do to avoid another potential disaster. The problem is so big that no one company, or even one industry, can have an impact alone.

Some might think today’s shortage is a one-time event. As long as we don’t have a once-in-a-century global pandemic, a fire at a major chip plant in Japan, a freeze in Texas, and a ship stuck in the Suez Canal — all of which come at just the right time — the next shortage won’t happen” and might not be that serious?

Above: Over the decades, the chip industry has experienced six declines and shortages. Image source: Deloitte

But Deloitte said that within the next decade, it is almost certain that a combination of events such as a global recession, major weather events and disruptions near important seaports or straits could occur almost simultaneously. The chip manufacturing industry and supply chain that currently exists are inherently vulnerable to disruption, making shortages inevitable.

Over the past three decades, we have seen six shortages of similar duration or magnitude to today’s. Sometimes shortages occur or are exacerbated by external shocks like the tech bubble or the 2009 recession, but sometimes they “just happen.”

Adding capacity in the chip industry has always been an expensive and cumbersome approach. It happens in waves driven by technology and market forces, and there is a long lead time between the decision to build a fab (or semiconductor manufacturing plant) and when that fab produces its first products (finished wafers). So the real question is not if shortages will come again, but when and how severe will they be?

Breaking the bullwhip

Above: Solutions to chip shortages. Image source: Deloitte

When poor communication in the supply chain causes supply and demand to become out of sync, the bullwhip is the seesaw for sales.

All the different players need to do their parts well and work together without creating excess. Deloitte said companies should choose specific actions or combinations of actions based on the role they play in the broader semiconductor ecosystem and value chain.

The entire chip industry is committed to increasing overall production capacity at an unprecedented level. From 2021 to 2023, capital expenditures by the three major players may exceed $200 billion and may reach $400 billion by 2025.

Governments have also pledged hundreds of billions of dollars. Deloitte expects the number of 200mm wafers (processed in chip factories and cut into individual chips) started globally each month to rise to 30 million by the end of 2023, from about 20 million in 2020.

Capacity will grow for both 200mm and 300mm wafer sizes, with both growing at roughly the same rate. To be clear, 200mm growth will come primarily from capacity additions to existing fabs, rather than construction of entirely new factories, which will account for nearly $12 billion in capital equipment spending between 2020 and 2022.

Above: Deloitte's forecast of chip production capacity. Image source: Deloitte

From a technology perspective, capacity at mainstream and more advanced process nodes (below 10nm, primarily 3nm, 5nm, and 7nm — where nm refers to nanometers, or the width between circuits) will grow faster than at more mature process nodes. Notably, demand is growing for wafer size and all process nodes, not just the most advanced.

Would a broad-based 50% increase in capacity over just three years be enough to make up for future shortfalls? The answer is not so obvious.

In addition to increasing overall production capacity, the chip industry should also build local production capacity. Chip manufacturing is geographically clustered and needs to be distributed in more regions. The concentration in East Asia (including Japan and China, close to 60%) in 2020 has attracted the attention of the governments of the United States, Europe and China, and plans have been made to build new factories in these countries or regions, such as Israel, Singapore and other countries.

Above: The chip industry is concentrated in Taiwan and other parts of Asia. Image source: Deloitte

It is difficult to change the geographic concentration of chip supply. There are more than 400 semiconductor manufacturing plants in the world, and according to plans, 24 new 300mm wafer fabs will be added by 2022, but only 10 new 200mm wafer fabs will be added during the same period.

Some of those are in South Korea and Taiwan. Adding a few dozen new locations outside those clusters could help. Deloitte said the new locations would only cause East Asia’s concentration to drop by a few points, meaning it would still produce more than half of all chips by 2023.

The industry should also become lean strategically -- chip buyers, distributors and retailers need to determine which level of leanness to choose. There is such a thing as too lean.

The demand-side bullwhip needs to be broken. OEMs (system companies), distributors/suppliers, and customers are subject to the bullwhip effect, where delayed communication between stakeholders at all levels of the supply chain is amplified by judgment calls about demand signals. This needs to change.

Above: The U.S. share of chip manufacturing has been declining for decades. Image source: Deloitte

Smart operations capabilities are critical to semiconductor manufacturing, which is inherently complex and sensitive, largely automated, and supported by capital-intensive factories. Functionality that facilitates digital process modeling (e.g., digital twins), operations monitoring, factory operations synchronized with material availability, and responsive factory scheduling adjustments enables factory operations teams to operate efficiently with high asset utilization.

Deloitte says many manufacturers began digital transformation in the spring of 2021. Continuous innovation is needed to better adapt to future supply chain-driven business disruptions. All of this requires close communication.

Demand is growing at about the same pace (or faster) as planned capacity is growing. Demand drivers include 5G, artificial intelligence and machine learning (AI/ML), the intelligent edge, and the Internet of Things (IoT). Some of this is about providing increasingly powerful chips for products that already use a lot of chips, but some of it is about adding chips to products that didn't have them before.

Keywords：chip Reference address：Chip shortages will be the norm in the future

Previous article：Following TSMC and Samsung, will SK Hynix also consider building a wafer factory in the United States?
Next article：Synopsys Acquires Concertio, a Leader in AI-Driven Real-Time Performance Optimization

Popular Resources
Popular amplifiers