Supply of electronic components is finally starting to match demand within some verticals, but the automotive industry is still feeling the pressure of shortages. Three million vehicles have been cut from global production levels, and estimates indicate that this number could reach 3.8 million before year end. While some businesses are managing inflated inventories, why is automotive still dealing with the consequences of the chip shortage?
Car builds heavily rely on legacy technology. However, the cons for legacy tech are growing and one point is becoming incontestable: Automotive manufacturers cannot afford to play the waiting game any longer when it comes to innovation. To move the automotive industry forward and out of the ongoing shortage, car manufacturers will need to work alongside chipmakers to increase supply and modernize vehicles where possible.
In 1965, Gordon Moore, co-founder and chairman emeritus of Intel Corporation, hypothesized what is now known as Moore’s Law. This theory stated that the number of transistors on a microchip would double every year. The theory has proven true over the years, as technology like cell phones and computers have gotten smaller, and therefore necessitated advanced process technologies run on fewer nodes.
As nodes have gotten smaller, their output in performance, and power reduction changes. The size of the wafer’s geometry directly relates to a semiconductor’s memory, yield and reliability.
Thanks to their larger size, legacy process nodes have better refresh performance, improved yields with lower failure rates, higher temperature tolerance, fewer fabrication-related issues and superior reliability. These are all very important for performance and safety, which are among the main priorities for the automotive industry’s customer base.
While the automotive industry prefers legacy technology because of its reliability, they are not the only ones; 40nm - 65nm process nodes, used for MCUs, ASICs, and image sensors, are also popular with industries like consumer electronics, industrial manufacturers, smart home and infrastructure as well the Internet of Things. These industries are growing faster and placing larger orders, which means automotive customers often find themselves in long queues for parts.
Prior to the pandemic’s effects on the supply chain, booking windows for car makers were about 12 weeks. Within a year, that window stretched to at least 12 months, with some OEMs forced to book capacity at tier-2 foundries as much as two years in advance. We have already seen how this has affected production levels via substantial cuts to forecasts, but it has also pushed pricing higher.
The components around these aging systems are advancing, which means that the infrastructure of the software needs consistent repair maintenance. Car dealerships have reported that 60 – 80% of IT budgets go towards upkeep for the legacy tech. If the automotive industry phases out legacy technology, the money once used for maintenance could go towards innovation.
According to Intel’s CEO, semiconductors will account for around 20% of a vehicle’s bill of materials by 2030. As these expenses trickle down to the consumer, car makers run the risk of demand suffering.
In accordance with Moore’s law, legacy technology manufacturing is continuously being replaced with newer, faster versions. Despite the sourcing difficulties that can come from using analog-based components, many vehicles still utilize them for a variety of applications. In some cases, you can find process nodes in today’s cars that were once used in PC chips as far back as 2005.
As supply of these components decreases, lead times lengthen, and shortages increase. In the case of integrated circuits built on legacy manufacturing nodes, studies show an anticipated supply deficit of nearly 30%, when lined up with existing demand, as far reaching as 2026.
Cars haven’t upgraded to new kinds of components because the industry lacks innovation. Plus, the benefits of legacy technology, combined with the costs associated with moving away from these process nodes, has kept the automotive industry clinging to the past.
In addition, development cycles for vehicles are lengthy; it takes four or more years to develop new models. Once these vehicles come to market, they have a long shelf life and undergo only minor facelifts. Departmental silos within car manufacturing companies keep product, design, and engineering teams from creating consistent cross-functional product development. Furthermore, the majority of the component manufacturing process is outsourced to chipmakers.
Car makers are comfortable sticking with their tried-and-true, antiquated technology. They know it works and that it produces quality products, but as components become more advanced, the automotive industry steadily increases its chances of being left in the dust.
The chips used in cars are usually over 20nm, and the factories that produce those chips are being rebuilt to focus on advanced chips. Chip manufacturers are focusing on developing newer, smaller nodes, mainly those under 10nm. By 2024, 30% of capacity will be allocated to high-tech process nodes under 10nm.
Initiatives like the CHIPS Act in the United States and European Union financially support this increase in allocation. Yet, companies like STMicroelectronics claim that these bills do not go far enough, as the demand for legacy technology outweighs the need for smaller node chips within industries like automotive. STM has made a point of saying that they will not participate in Europe’s chip project if it involves advanced technologies.
As much as 67% of chips today run on process nodes of 90nm or higher, while 21% are between 22nm and 65nm. McKinsey has also estimated that by 2030, 67% of total demand will still be dedicated to 90nm and above. These numbers support STM’s position that the real goal should be helping the entire electronic component ecosystem instead of hyper-focusing on one area of demand.
If the automotive industry decides to pivot to advanced technology, it will be a long road. It would take between two and five years to design, test and validate these components for traditional car models. Plus, the risks of ramping up production on smaller nodes would heavily affect timelines and success rates.
If cars solely relied on smaller node technology, they would need to work around the industrial temperature issues, as these new process nodes have less stability at higher temperatures. Quality could suffer, and when it comes to the automotive industry, that is simply not an option.
While bills like the CHIPS Act are a good starting point to address the needs of the semiconductor industry and the businesses it supports, there is additional progress that needs to be made. While the methods may vary, the end goal for chip manufacturers should be to alleviate the shortages for car makers.
Companies like Taiwan Semiconductor Manufacturing Co. are taking steps toward innovation by investing $100B over the next three years in new factories, with some specifically focused on automotive customers. As of now, only 14% of TSMC’s chips go towards the global automotive IC market, but with this new expansion, TSMC is hoping to change that.
At their Ireland fab campus, Intel is committing foundry capacity to automotive verticals. Plus, Intel is diligently working to expand factory operations within Ireland.
GlobalFoundries is also investing in expansion within their manufacturing plants in Asia, Europe and the United States, as well as Dresden, Germany. Dresden already uses process technologies on nodes ranging from 55nm - 22nm.
The automotive industry should also chip in to make changes to their builds. Innovation is clearly lacking within their current system, so car makers should encourage cross-team development and further the pace of modernization.
Where possible, advanced chips should be utilized over trailing edge technology. Tesla and electric vehicle makers in China have already seen the benefits of embracing more advanced process nodes, as they have more versatility in designs for chips.
Advanced chips have proven that they can be used in cars for applications like artificial intelligence, infotainment and communications. The industry has also already identified a use for advanced microcontrollers on larger process nodes for braking. If car makers can break down their historical silos and have teams work together to make this sort of innovation possible, it should be done.
Manufacturers are already investing to improve their chip-making capacity. However, the automotive industry will need to make a change in the status quo if they want to see an improvement in the ongoing chip shortage. Although legacy technology has served the industry well, innovation in design and flexibility in supply partnerships would go a long way in helping the automotive industry improve its standing in the supply chain.