Semiconductor Challenges in the Automotive Industry

Wilfried G. Aulbur is responsible for the firm's global Commercial Vehicle, Construction and Agricultural business. Wilfried is based in Chicago and supports the US Automotive, Industrials and Private Equity practice. Wilfried's work includes growth strategies, M&A and due diligence, operational performance improvements and technology strategies. Wilfried has served on numerous industry associations and panels, is a regular speaker at key industry events and publishes extensively.

The automotive industry has been battling massive supply chain challenges in the aftermath of COVID and lost significant revenues in the process. While the intensity of overall supply chain disruptions due to COVID is slowly subsiding, it is still worthwhile to understand the overall dynamics of key supply chain shortages, semiconductors being one of them, and to differentiate temporal from systemic supply chain challenges. In this article, I will focus on the systemic challenges in the automotive semiconductor supply chain.

Four key factors can be identified as main reasons for the automotive semiconductor crisis. The first is the lack of understanding of the semiconductor industry by automotive OEMs and Tier1s. The second is the overall relevance of the automotive industry in the semiconductor industry and the third is the use of older technology in automotive vs consumer goods which is by and large reflected in capacity additions in semiconductor fabs around the globe. A final complication are the increasing tensions between the West and China which threaten supply from a geopolitical perspective.

Automotive companies have learned over the years to buy key commodities, e.g., castings, and have a trained set of engineers and procurement people that understand how these products are being produced, what their quality and cost levels should be etc. Faced with significant loss of incoming orders during COVID, OEMs and subsequently suppliers did what they do best, adjust the supply chain by canceling orders to reflect expected lower demand. Unfortunately, semiconductors do not fit the standard just-in-time or just-in-sequence models of the automotive industry. The production of a semiconductor is a complex, long and iterative process that starts at the production of silicon wafers, moves to front end processing (lithography, deposition, etching, metallization, etc.), back-end processing (back grinding, dicing, flip-chip, etc.), packaging, and components to a system which is then integrated in a vehicle. Building semiconductor fabs is a highly capital intensive process which requires high factory utilization. Hence, if one industry cancels orders, it is in the interest of producers to find replacement customers quickly. With increasing demand in the consumer goods segment during COVID, these customers weren’t hard to find and long-term contracts were signed. This resulted in lack of capacity for the automotive industry when the order books were starting to fill up again.

Combined with a better understanding of semiconductor supply chains and the decision not to reduce volume requirements during the crisis, this led to continued growth for the company

The second challenge for automotive OEMs and suppliers was the fact that in contrast to standard components (castings, plastics, etc.), the automotive industry is not the dominant customer for semiconductors. As shown in Figure 1, 80% of the semiconductor revenue is made in communication, consumer goods and data processing. To make things worse, automotive is a market with very high fragmentation since every ECU often is tailored to a specific task and OEM, vs highly standardized applications such as an Apple iphone. Hence, automotive companies found themselves in unexpected and unfortunate situation that their leverage over suppliers was limited.

In addition, traditional OEMs used E/E architectures that were based on 1990’s technology. As Figure 2 shows, standard internal combustion engine vehicles use dominantly mature nodes that were introduced in the 1990s as well as legacy nodes that are even older. This compares unfavorably with consumer goods and medical devices, areas in which leading edge notes and advanced nodes are much more common.

The challenge with this choice of technology is that hardly any capacity is being built up in the technology nodes relevant to automotive as the business case for investments is challenging. The capacity additions that are happening in mature nodes are predominantly happening in China and 65% of these additions is for local consumption. In addition, even in mature nodes, OEMs and Tier 1s compete with consumer goods companies. Take the Apple iPhone 13 Pro as an example. Compared to the previous version, the iPhone 13 Pro added features such as fast charging, premium audio, and 5G amplification that added demand of mature nodes. The additional yearly demand in 40-100 nm equivalents is estimated to be about 2.0 billion industrial/automotive semiconductors which were not available for automotive players.

In addition, once the crisis hit, many OEMs found themselves not in a position to effectively combat the crisis. Transparency over the supply chain in terms of which chip was used in which component and who produced it where was by and large completely lacking.

"Combined with a better understanding of semiconductor supply chains and the decision not to reduce volume requirements during the crisis, this led to continued growth for the company"

However, not all automotive OEMs are created equally. Tesla, for example, already had an E/E architecture that actively uses advanced nodes and looks more like an iPhone rather than a traditional vehicle. Combined with a better understanding of semiconductor supply chains and the decision not to reduce volume requirements during the crisis, this led to continued growth for the company. Many other OEMs had to accept volume and market share losses instead.
In the short-term, the supply situation is improving for technologies with high overlap between automotive and consumer electronics needs, i.e., 90 nm and newer. This is driven by the overall cooling of consumer demand given high interest rates and a more challenging economic environment. However, in the mid-term we expect key shortages to continue to exist as shown in Figure 3.

In addition, continued tension with China, which together with Taiwan accounts for about 38% of the global semiconductor manufacturing capacity, puts pressure on semiconductor supply chains. Escalating tensions can easily lead to major, systemic disruptions in supply and need to be taken into account from an overall corporate planning perspective.

What are then levers that OEMs and suppliers need to apply? Clearly, semiconductor sourcing is a strategic effort of players that requires professional management in procurement with personnel that understands the intricacies of the market in depth. Beyond task forces, creating transparency and working with suppliers to ensure supply, a key lever for the future is design to risk. This entails moving away from mature nodes towards a more substantial leverage of advanced nodes as well as moving towards more modern, centralized E/E architectures.