When Semiconductor Chips Are Down, What Can Be Done?
Over the past 12 months, car manufacturers have screeched to a halt, consumer electronics plants have gone offline, and home appliances factories have rumbled to a standstill.
Why? Because of the global semiconductor chip shortage, expected to extend well into 2022 and likely into 2023. It would be easy enough to point to the pandemic as the culprit, but this was merely the tipping point, and grossly underrepresents the confluence of different factors impacting the industry and the global economy.
While the semiconductor industry only amounts to 0.3% of the U.S. GDP, semiconductors are required to produce 12% of the U.S. national output. So, as we design and manufacture smarter products, almost every manufacturing industry is dependent on semiconductors.
Here is a summary of the factors that led to the calamity and perspective on how companies throughout the value chain can create greater resilience during times of disruption.
What took chips out of the global supply chain?
A few months ago, I wrote about the semiconductor challenges which explained that at the most fundamental level, demand is out of phase with supply.
At the beginning of the COVID outbreak, as the demand for cars plummeted, the automotive industry cut its forecast by about 13%.
Demand for consumer electronics as “working from home” became the norm. This resulted in increased demand for additional bandwidth, improved audio equipment, new video lighting, and high resolution cameras. The stay-at-home mandates also increased demands for home appliances such as TV’s, washing machines and dishwashers (my household had to invest in two of these three).
This caused a surge in demand for semiconductors, and high-tech OEMs quickly seized the opportunity and increased their forecast to take advantage of the additional capacity.
Next, geopolitical tensions arose between major technology-leading countries. The United States and China have been engaged in severe cross-border trade disputes over the last several years, as have Korea and Japan. In response to U.S. sanctions imposed on China, large Chinese OEMs were stockpiling chips to hedge against impending supply disruptions.
Globalization has exposed risks
For cost-reduction purposes, overt the past decade or more, the production of semiconductors and microelectronics has moved offshore, and supply chains have become more global. In 1990, 37% of chips were made in American factories. By 2020, despite accounting for an estimated 47% of semiconductor sales, that number had declined to just 12%.
Europe’s share declined to nine percent from 40 percent over the same period. Manufacturing is now concentrated in Taiwan, South Korea and China. So, when several foundries and integrated device manufacturers (IDMs) closed or ran reduced shifts early in the pandemic, it created a noticeable impact on global capacity.
Taken in isolation, any of the above events would not create the massive trough in supply. The system has enough slack to absorb one or even a few of the aftershocks. However, the cumulative effect of them all hitting the system in a relatively short time exceeded the system’s capabilities to respond.However, several other noneconomic factors also impacted supply. For example, a prominent semiconductor manufacturer in Japan who produces 30% of the chips used in the auto industry closed their factory for four weeks due to a fire. Additionally, the United States imposed sanctions on the export of semiconductor manufacturing equipment to China in September 2020, resulting in several foundries based in mainland China being unable to purchase necessary equipment required to support the growth in chip demand.
So why can’t we just make more semiconductors?
Even during the best of times, manufacturing semiconductors is a lengthy and variable process. Progressing from wafer fabrication to die bank takes 12 to 14 weeks. Little can be done physically to reduce the lead time without incurring significant expediting costs.
Compounding the problem is the time to add new capacity. To build and qualify a new fabrication takes 3 to 5 years and has an entry cost of $4 billion to $10 billion. Furthermore, the output is highly vulnerable to defects, contamination, and technical factors that can have a major impact on the quality and grade of the semiconductors being produced. Therefore, it is extremely expensive and difficult to pivot in response to major changes in demand from OEM customers.
When the chips are down, what can be done?
It is important for organizations to address the challenges across three different time horizons: the immediate, the near term, and the long term. By doing so, they can help ensure revenue continuity, fulfill customer expectations, and achieve greater resilience.
In the short term, we are in firefighting mode and ensuring customers are constantly updated. Material substitution must be evaluated to fulfill demand, and expectations constantly reset and communicated across the network. To enable this companies, need a best-practice approach to:
- Maximize production capacity through predictive what-if scenarios for yield management
- Smart testing to optimize test flows and costs to meet defect parts per million requirements
- Enable multitier visibility, collaboration, and exception management with key suppliers within their network
Over the 6-12 month time horizon, companies are in the containment phase where they are working to establish tighter processes and deploy improved tools to create more visibility and insight for better decision-making.
In this phase, the Integrated Device Manufacturers (IDMs) and foundries are striving to get ahead of unforeseen problems – for example, by establishing early warning systems for plant maintenance in the fabs. On the other hand, Original Equipment Manufacturers (OEMs) are establishing processes to replan and allocate supply to customers and improve B2B connections across the network. Key goals in this horizon include:
- Increasing uptime of equipment with predictive maintenance based on real time sensor data
- Improving customer satisfaction with supply chain planning based on objective-driven customer allocations
- Extended collaboration across the value chain to facilitate orchestration of material flow with all suppliers
Finally, let’s look at the 12 month and beyond outlook. In this phase, companies need to define their vision and strategy for the value chain with the goal of improving resilience to become more agile, sustainable, and transparent. These traits are required to buffer their value chain from inevitable and unpredictable shocks to the forecast. Each value chain has value chain participants who defined initiatives to improve resilience. For example:
- Foundries may be working towards reshoring manufacturing capacity to domestic markets
- OEMs my look to create demand and supply-side planning systems
- IDMs and fabless semiconductor companies my prioritize developing streamlined planning processes to ensure tighter alignment between customer forecasts and order commitments
It is clear that there is no quick fix to the semiconductor challenges and that this is not the first chip shortage, and it will certainly not be the last. Companies across the supply chain need to have well established processes, current planning, logistics, maintenance management and network tools, and the right talent to help ensure the gap between supply and demand remains small.
To learn more, check out this recent IDC Analyst Connect paper.