Excerpted from MIH EV Professional Journal Vol.07 – The Smart Vehicle Race Reshapes Supply Chain Positioning (original article)

As smart vehicles move into mass production, semiconductor demand is no longer concentrated solely on high-computing-power system-on-chips (SoCs). A broad range of technologies—including low-level control, power conversion, sensing, connectivity, security, memory, and storage—are also seeing growing demand. Rather than driving upgrades to a single type of chip, smart vehicles are expanding the overall semiconductor landscape, from control and power management to sensing, connectivity, security, and data storage. As vehicles generate, retain, and update increasing volumes of data, automotive storage and NAND controllers are becoming important areas to watch. At the same time, amid greater volatility in memory supply and demand, securing a stable supply for automotive applications will also influence the pace at which automakers introduce intelligent features.

Traditional Automotive Chipmakers Adjust Their Strategies as Automakers' Needs Diverge

The traditional automotive semiconductor market has long operated under a relatively stable model. With long product lifecycles and stringent automotive-grade reliability requirements, automakers and Tier 1 suppliers focus not only on specifications, but also on supply stability, quality traceability, and long-term support. Historically, demand has centered primarily on microcontroller units (MCUs), power devices, analog ICs, sensors, and controllers, with relatively well-defined roles for different chips across the vehicle's various control units.

As the automotive industry enters the era of smart vehicles, these demand patterns are becoming increasingly diversified. Advanced Driver Assistance Systems (ADAS), smart cockpits, electrification, connectivity, and software updates are expanding the role of automotive semiconductors from supporting individual control functions to enabling vehicle-wide functionality. As a result, the semiconductor mix required by automakers now spans multiple areas, including control, power management, sensing, security, connectivity, and data processing.

Looking at the current industry landscape, traditional automotive chipmakers are not focusing on a single direction. Instead, they are broadening their existing product portfolios to support the evolving functionality required by automakers. Infineon Technologies, building on its strengths in power semiconductors and automotive MCUs, has expanded into applications related to electrification, body control, and safety. STMicroelectronics continues to strengthen its presence across power, sensing, analog, and automotive control technologies to address growing component demand as vehicles incorporate more electronic functions. Renesas Electronics is enhancing its support for automakers and Tier 1 suppliers in system development, with offerings ranging from controllers and SoCs to development platforms and reference solutions. NXP Semiconductors also remains a major automotive semiconductor supplier, with a portfolio spanning processors, controllers, security, and connectivity. These developments reflect how established automotive chipmakers continue to adjust their market positioning in response to the evolving requirements of smart vehicles.

These developments show that traditional automotive chipmakers are not simply making incremental extensions to their existing product lines. Instead, they are seeking to align their capabilities in control, power management, sensing, connectivity, and security more closely with the functional requirements of next-generation vehicles. The market changes driven by smart vehicles extend beyond high-end artificial intelligence (AI) chips and are also reshaping the product lines that have traditionally formed the foundation of the automotive semiconductor market.

Semiconductor Demand for Smart Vehicles

As Vehicles Become More Data-Driven, Onboard Storage Demand Comes to the Fore

As smart vehicles incorporate more features, semiconductor demand is expanding beyond control and computing to encompass data-related requirements. Vehicles are gradually evolving from mechanical systems that simply execute commands into mobile platforms that continuously generate, use, and update data. While some data must be processed in real time, other data needs to be stored reliably within the vehicle, making onboard storage increasingly important.

According to Nelson Duann, Senior Vice President of Client & Automotive Storage Business at Silicon Motion, two key factors are driving the growing demand for automotive storage. First, as artificial intelligence (AI) and large language models (LLMs) move into the automotive environment, the volume of data is increasing significantly, driving demand for greater storage capacity. Second, as automakers integrate previously separate cockpit, driving, and parking functions into unified systems, multiple low-capacity storage components are being consolidated into a single high-capacity storage device. In other words, while the number of storage devices in a vehicle may decrease, the capacity and reliability requirements for each device are increasing.

However, rising capacity requirements do not necessarily mean that supply can keep pace. When the memory market faces shortages or price fluctuations, automotive demand is often not the first priority, as data centers and smartphones remain the major markets that memory manufacturers prioritize. For automakers, therefore, the challenge presented by growing automotive storage demand goes beyond capacity upgrades. It also involves securing a stable supply of memory through market cycles. As a result, storage is no longer simply a place to hold data. As vehicle functions become increasingly dependent on software and data, automotive storage solutions must provide more than capacity. They must also ensure that data can be stored reliably and accessed consistently throughout the vehicle's lifecycle, while supporting vehicle development, mass production, and subsequent product transitions.

As a result, the role of storage components in smart vehicles cannot be evaluated solely in terms of capacity or cost. As data requirements grow amid fluctuations in memory supply, automakers are placing greater emphasis on whether storage solutions can provide long-term support throughout vehicle development and mass-production cycles.

This shift is creating a new area of demand in the automotive semiconductor market that has received relatively little attention in the past. Discussions around smart vehicles typically focus first on computing performance, sensing capabilities, and control chips. However, as in-vehicle data needs to be stored, accessed, updated, and managed, the role of automotive storage is also evolving. It is becoming increasingly interconnected with vehicle software functions, data lifecycles, and long-term reliability, making NAND controllers another important area for understanding the data infrastructure of smart vehicles.

Automotive Storage Capacity Grows, Putting Reliability and Supply Chains to the Test

Within this shift in demand, Silicon Motion offers a useful example of how automotive storage is evolving from a capacity-focused component into part of the vehicle's data infrastructure. Unlike automotive platform or high-performance SoC suppliers, Silicon Motion focuses on the specialized area of NAND controllers and storage management technologies. For smart vehicles, reliable data storage and access depend on controllers and firmware that can operate consistently over extended periods. As in-vehicle data volumes grow and storage capacities increase, these capabilities—once less visible in the automotive semiconductor landscape—are becoming an increasingly important part of automotive semiconductor demand.

Capacity | From Distributed Low-Capacity Storage to High-Capacity Integration

As demand for automotive storage increases, market changes are reflected not only in storage capacity, but also in product configurations and supply chain requirements. Citing Silicon Motion's product portfolio as an example, Nelson Duann noted that different in-vehicle systems previously relied on multiple low-capacity storage devices. However, with trends such as cockpit-driving integration and the integration of cockpit and parking functions, storage configurations are shifting toward fewer devices with higher capacities. Currently, embedded MultiMediaCard (eMMC) and Universal Flash Storage (UFS) remain the mainstream storage solutions for automotive applications, while the adoption of PCIe solid-state drives (SSDs) is also gradually increasing. Automotive embedded storage capacity varies by application, ranging from low-capacity control modules to higher-capacity cockpits, ADAS, and in-vehicle computing platforms. Some suppliers already offer automotive UFS or embedded storage products with capacities of up to 1TB. Silicon Motion's Ferri storage solutions, along with its eMMC, UFS, and SSD controllers, illustrate this transition from distributed, lower-capacity storage toward higher-capacity and more highly integrated storage architectures.

Endurance | Keeping Pace with Automakers' Long Product Cycles Amid Memory Generation Transitions

Competition in automotive storage is no longer defined solely by chip pricing or capacity specifications. Instead, it has become a comprehensive test of controllers, firmware, testing, supply chain management, and automotive qualification processes. For automakers, once a storage component enters mass production in a vehicle model, considerations extend beyond the initial cost to include whether stable supply can be maintained for years, whether product transitions can be managed smoothly, and whether existing vehicle models can continue to secure components as memory technologies evolve. Long-term supply is one of the key requirements that distinguishes the automotive market. Nelson Duann noted that automakers previously often required 10 to 15 years of supply support. Although product lifecycles have shortened, support of at least five to seven years is still required. For example, as 8GB Multi-Level Cell (MLC) storage products enter the End-of-Life (EOL) stage, Silicon Motion is maintaining safety stock to support legacy demand while also helping customers transition to 64GB Triple-Level Cell (TLC) storage products. This transition illustrates that the challenges facing automotive storage come not only from capacity upgrades driven by new vehicle functions, but also from the gap between the need to maintain supply for existing vehicle models and the continued evolution of memory technology.

Reliability | Ensuring Data Integrity and Retention Across Wide Temperature Ranges

Automakers previously often required 10 to 15 years of component availability. Although product lifecycles have shortened, they still typically require at least five to seven years of supply support. Suppliers therefore need to help customers manage EOL transitions, capacity upgrades, and migrations between legacy and new products. Capacity expansion, however, is only the first layer of change. The greater challenge for automotive storage lies in reliability. Nelson Duann noted that one of the most fundamental reliability requirements for automotive storage is the ability to operate across wide temperature ranges. Depending on installation location and product grade, automotive storage components must meet different operating temperature requirements. Common specifications range from -25°C to 85°C, while some automotive-grade SSD controllers can support temperatures from -40°C to 85°C or even higher. Under these conditions, data must remain accurate and intact. Because NAND flash is inherently susceptible to errors, controllers must provide robust error correction capabilities to maintain data integrity across varying temperatures and over extended periods of operation. This is also one of the key differences between automotive storage and general consumer storage. Duann also noted that Quad-Level Cell (QLC) storage, with its lower cost per unit of capacity, could potentially help reduce storage costs in the future. However, it is still too early for QLC to be widely adopted in automotive applications, primarily because its write endurance and reliability have yet to fully meet automotive requirements. QLC will likely need to be validated and proven mature first in consumer applications such as smartphones before it can potentially expand into automotive use.

Cybersecurity | Data Storage and Updates Are Critical to System Trust

As vehicle functions become increasingly dependent on software and data, storage is no longer simply a place to store information. In-vehicle data must be stored, accessed, updated, and managed, meaning that storage solutions need to support not only capacity, but also data availability and consistency throughout the vehicle lifecycle. For automakers, if data cannot be stored reliably, maintenance diagnostics, functional updates, and system management may all be affected. This is also why NAND controllers and automotive-grade storage solutions are increasingly being evaluated earlier in the vehicle development process. AI is also creating new value for automotive storage. As large language models (LLMs) are introduced into vehicles, relying entirely on Dynamic Random Access Memory (DRAM) for data can create significant cost pressure. Nelson Duann noted that Key-Value Cache (KV Cache) can be offloaded to flash memory, reducing the need for more expensive DRAM. This is particularly meaningful for cost-sensitive entry-level and mid-range vehicles. As a result, the role of storage is expanding beyond simply storing data to influencing the cost structure of in-vehicle AI. As capacity and multitasking requirements increase, UFS is also expected to gradually replace eMMC. Duann observed that UFS controllers are moving from 6nm to 5nm process technology, while capacities are expanding to 128GB, 256GB, and 512GB. Beyond capacity growth, automotive storage is also seeking a new balance among process technology, reliability, supply chain continuity, and cost.

Mass-Production Competition Goes Beyond Specifications, with Long-Term Reliability Becoming Essential

Smart vehicles are changing the competitive dynamics of the automotive semiconductor market. In the past, automotive demand was often viewed in terms of individual chip specifications or product categories. As vehicle functionality expands, however, automakers are placing greater emphasis on whether a comprehensive set of capabilities can support mass-produced vehicle platforms over the long term. Control, power management, sensing, security, and connectivity remain fundamental, while data management and storage reliability are emerging as increasingly important requirements.

The strategies of traditional automotive chipmakers reflect their efforts to respond to the increasingly diverse requirements of automakers as vehicle functionality evolves. Silicon Motion offers another perspective: as in-vehicle data volumes grow, automotive storage and NAND controllers are becoming an increasingly important part of the semiconductor landscape for smart vehicles. While suppliers in this segment may not define the overall vehicle platform, their ability to provide reliable data storage, controllers, firmware, automotive-grade qualification, and long-term supply support will play an important role in ensuring stable operation after vehicles enter mass production.

Nelson Duann also noted that the challenges facing automotive storage extend beyond technology to supply chain realities. Fluctuations in memory supply and demand have become a factor that automakers must address early, while automotive products still require multi-year supply commitments. Suppliers that can maintain stability amid capacity upgrades, product transitions, and supply fluctuations will be better positioned to become long-term partners in the smart vehicle supply chain.

Common Automotive Storage Specifications

This also illustrates that while competition in the smart vehicle semiconductor market often draws attention to high-performance computing chips, AI platforms, and smart cockpit features, the challenges automakers face once vehicles enter mass production involve much more complex supply chain management. The memory industry is highly cyclical, with pricing, production capacity, and product generations influenced by demand from data centers, smartphones, PCs, and consumer electronics. Although the automotive market places a premium on long-term supply stability, its scale may not always be sufficient to secure priority during every supply shortage. This makes automotive storage suppliers an important link between automakers' long-term requirements and the supply dynamics of memory manufacturers.

For automakers, these fundamental components may not have traditionally been among the first considerations when planning smart vehicle features. However, as in-vehicle data volumes grow, software updates become more frequent, and AI applications are introduced into vehicles, storage selection will have an impact on product maintenance and feature evolution for years to come. Focusing solely on current capacity and cost may underestimate the risks associated with product discontinuation, supply fluctuations, and specification transitions. If storage is viewed as part of the vehicle's data infrastructure, automakers will need to evaluate controllers, firmware, automotive-grade qualification, long-term supply, and product transition capabilities earlier in the development process.

As automotive semiconductor demand expands, the shift involves more than simply bringing a wider range of components into vehicles. It also reflects a change in how automakers evaluate their supply chains. High-performance computing determines the range of applications that smart vehicles can support, while control and sensing technologies enable vehicles to gather information about their surroundings and operating conditions. Storage, meanwhile, is taking on a more fundamental role in data retention, access, and long-term maintenance. As data becomes more deeply integrated into vehicle functionality, automotive storage and NAND controllers may not always be at the forefront of industry discussions, but they are becoming areas that automakers need to evaluate earlier in the development of mass-production vehicles.