The automotive industry is running out of chips. The global semiconductor shortage that began in the first quarter of 2021 has halted assembly lines around the world, as the long lead time for the tiny silicon chips has slowed production of everything from smartphones and home appliances to driver-assistance systems. Major carmakers, including a US-based OEM, have already announced significant rollbacks in their production, lowering expected revenue for 2021 by billions of dollars.
That challenge in the auto industry is the latest in a series of them that began in the early months of the COVID-19 pandemic, when auto sales plummeted as much as 80 percent in Europe, 70 percent in China, and nearly 50 percent in the United States. The lack of demand for new cars shuttered auto factories and sent home millions of workers, while orders for semiconductors—used in myriad ways, including in fuel-pressure sensors, digital speedometers, and navigation displays—dropped off precipitously.
The effects of the semiconductor shortage have extended beyond the auto sector, with other industrial players struggling to secure chips. That highlights the fragility of those supply chains, which largely rely on Asia as a hub of semiconductor manufacturing. Many automakers are now operating in crisis mode, and few expect a rapid resolution. Auto manufacturers and chipmakers alike will need to work together to tackle the imbalance in demand. This article addresses both how the shortage happened and what remedies for it exist.
The effects of the semiconductor shortage have extended beyond the auto sector, with other industrial players struggling to secure chips.
How the shortage happened
No single incident or disruption caused the semiconductor shortage. Instead, a confluence of events contributed to the situation the auto industry now faces.
Struggles during the COVID-19 crisis
In the first half of 2020, the auto industry faced a substantial drop in demand. Moreover, while new-vehicle sales grew in the second half of the year, the highly ambiguous sales outlook at the time meant that automakers didn’t meaningfully increase their semiconductor orders. At the same time, driven by the shift to remote work and the associated greater need for connectivity, consumer demand significantly rose for personal computers, servers, and equipment for wired communications, all of which heavily depend on semiconductors. That meant that even as the auto industry drastically cut chip orders, other sectors faced an increased need.
Our analysis of IHS Market data reveals that the actual demand for semiconductors in the auto industry in 2020 trailed a prepandemic estimate by around 15 percentage points (Exhibit 1). Over the same period, most other segments (with the exception of the industrial sector) experienced rapid expansion, resulting in an average increase from 5 to 9 percent in semiconductor sales beyond the forecasted growth. Because of that, when the auto sector’s demand recovered faster than anticipated in the second half of 2020, the semiconductor industry had already shifted production to meet demand for other applications.
Lack of new capacity
The semiconductor industry has matured in recent years through consolidation and the achievement of greater scale. Its capacity has expanded modestly but steadily—by around 4 percent annually, in line with sales (Exhibit 2). In parallel, semiconductor utilization has been consistently high (at or above 80 percent) in the past decade. In fact, utilization in 2020 was close to 90 percent, which many industry leaders regard as full utilization, since exceeding that level often results in disproportionately longer lead times. Therefore, while the semiconductor industry has increased its production capacity by nearly 180 percent since 2000, its total capacity is nearly exhausted at the current high utilization rate.
Because of geopolitical tensions, some consumer-electronics makers have considerably increased their chip-inventory levels to get through a period of limited access to semiconductor manufacturing. We estimate that such stockpiling caused a surge in semiconductor demand of 5 to 10 percent in the wireless space—the equivalent of one-third of auto-market chip sales.
The typical contracts for sourcing parts in the auto industry differ significantly from other industries, which are more often governed by long-term binding agreements (so-called take-or-pay deals) and provide semiconductor suppliers with purchase orders that go well beyond six to 12 months. Amid an auto supply chain that is complex and often heavily outsourced, the chip-sourcing commitment cycle for the auto industry, however, tends to be shorter term—especially with respect to binding purchase commitments on the order of a few weeks to a few months. While the auto industry has had a good reputation for stable demand in the past, semiconductor manufacturers are now committed to more conventional, longer-term contracts from other fast-acting industries.
Just-in-time manufacturing practices, which can minimize waste and increase efficiency by keeping on-hand inventory low, are widely leveraged in the auto supply chain. In normal times, the reduction of inventory is financially beneficial; however, in the event of an unexpected shortage, the practice causes immediate disruption of the entire supply chain. Since many players didn’t expect the chip shortage in 2020 and 2021, they likely had very limited stock available to weather the crisis.
5G rollout and overlapping chip demand
Industry demand for semiconductors varies by node size. Chips in the smaller size ranges, the most advanced of which are seven and 14 nanometers or smaller, are often used in leading-edge technology applications but aren’t required by many automakers. Our analysis reveals several knock-on effects of large-scale technology adoption that the auto industry must consider. For example, an expansive 5G rollout requires a large number of radio-frequency semiconductors manufactured at the same, larger node sizes as auto chips. The same is true for power-electronic chips needed to boot up servers and PCs (Exhibit 3). That amount of overlap means that as the rollout of 5G occurs over the next few years, automakers might see a continuing scarcity of chips.
Prospects for recovery
The global semiconductor shortage isn’t likely to resolve in the short term, because of factors such as the complexity of semiconductor manufacturing and the increasingly sophisticated chips needed in auto design. Because of that, we offer OEMs some shorter- and longer-term strategies to consider as businesses deal with the imbalance in semiconductor supply and demand.
In the short term, we don’t see any indication that the current supply and demand imbalance for semiconductors will resolve. That’s because typical lead times for semiconductor production can exceed four months for the products that are already well established in a manufacturing line (Exhibit 4). Increasing capacity by moving a product to another manufacturing site usually adds another six months (even in existing plants). Switching to a different manufacturer (for example, changing foundries) typically adds another year or more because the chip’s design requires alterations to match the specific manufacturing processes of the new manufacturing partner. Additionally, chips can contain manufacturer-specific intellectual property that may require alternations or licensing. Also, alternative suppliers in the auto industry must go through a lengthy and complex qualification process.
Our analysis suggests that chip capacity won’t catch up with demand in the short term for the auto industry. That is primarily because of the continued increases in volume and sophistication levels of the chips needed to power new technologies, such as advanced driver-assistance systems and autonomous driving.
Leading companies have taken a variety of measures to deal with the current situation. Many have established dedicated war rooms that combine their supply and demand intelligence to create greater transparency. For instance, automatically generated dashboards combine data from multiple sources on many segments, such as a company’s supply chain and a semiconductor player’s commitments. The use of analytics to match supply with demand helps reduce a cumbersome and error-prone manual effort. The goal is to provide clear input for internal communication and for communication to suppliers and customers. Companies usually view that as a no-regrets move.
Beyond that, many automakers and tier-one suppliers continue to collect and analyze more sophisticated intelligence on the semiconductor value chain and chip-manufacturing locations. To make more informed decisions, company leaders continually reassess the competitive landscape by weighing the technological applications for prioritization at the level of the individual chips required. Several tier-one and semiconductor players have complained about the lack of transparency regarding real demand levels (driven partially by the recent crisis-mode practice of overordering to secure a basic level of supply) and prioritization among individual components.
In our experience, a joint discussion involving an OEM, its tier-one suppliers, and semiconductor suppliers can help align the goals of all participants. Offering extra payments to expedite the production of wafers when capacity amounts to less than 5 percent of the production volume can also help. Other options involve replacing back-ordered components with similar but more feature-rich units (for example, swapping in chips with more memory) and using consumer-grade chip sets that receive additional quality tests.
Solving the long-term problem
In the longer term, the auto industry will need to rethink the way it structures contracts for semiconductor-related sourcing. As a good place to start, OEMs and tier-one players could make up-front volume commitments more binding (for instance, by moving to 12 months on a production- or technology-corridor1 level and six months on a chip-set level). A more balanced risk-sharing plan aligned along the value chain could also help drive adoption rates.
In addition, companies might have to reconsider, at least in part, the current practice of just-in-time delivery and low stock levels along their value chains. There is also a need to align with the current push by various governments for more regional sourcing, since many government leaders are concerned about the fragility of supply chains and the prospect of depending on single suppliers and distant countries for vital needs. A McKinsey Global Institute analysis found that supply-chain shocks affecting global production occur just under every four years, on average, with companies losing 42 percent of one year’s earnings every ten years.
In addition, auto players can revise their strategies for sourcing various chips. While a sourcing decision for a specific type of chip might appear on paper as less expensive than other options, the assessment might change when factoring in the cost of complexity in areas such as sourcing resilience and the software life cycle. In the qualification process, companies might need to reconsider some parameter constraints (for example, temperature range) to create the right balance of broader sourcing opportunities and product reliability.
Companies could also consider making selective investments in supply-chain resilience, with a clear-eyed view of their dependency on selected components and supply uncertainties. Such investments could range from spending on dual-source manufacturing qualification jointly with semiconductor suppliers to adjusting pricing levels with supply guarantees to bundling volumes to achieve greater negotiation power.
The current chip shortage is disrupting auto businesses across the value chain as OEMs and their suppliers rush to procure reliable chip sources. As auto players ponder their next moves and semiconductor manufacturers struggle to keep up with demand, both industries need to align their short- and long-term strategies to weather the supply-chain disruption as successfully as possible.