Automotive CIS chips, a "chip" is hard to find

In today's rapidly evolving intelligent driving technology, onboard cameras have become the core "eyes" for vehicles to perceive their environment, while CMOS image sensors (CIS) serve as the "retina" of these eyes.

As autonomous driving advances from Level 2 to Level 5, the number of cameras in each vehicle has surged from single digits to double digits, leading to an explosive growth in demand for high-resolution CIS chips. However, the supply of this critical component has fallen into a predicament of "hard-to-find chips."

The launch of BYD's "Eye of God" system, the popularization of Tesla's full-vision solution, and the global automotive industry's pursuit of advanced driver assistance features have pushed the supply-demand contradiction of 8-megapixel (8M) onboard CIS chips to the forefront.

Technological Upgrades and Explosive Demand

As a core component of onboard cameras, the performance of CIS chips directly determines a vehicle's ability to perceive its surroundings. In the current technological landscape, mainstream intelligent driving systems heavily rely on onboard cameras to capture vast amounts of image data, which are then processed through complex and sophisticated algorithms to achieve a series of critical functions such as precise lane recognition and rapid obstacle detection. Among these, high-resolution CIS chips (especially those with specifications of 8M and above) have become an indispensable configuration for Advanced Driver Assistance Systems (ADAS) and autonomous driving technology due to their ability to provide clearer and more detailed image information and longer effective recognition distances.

Taking BYD's meticulously crafted "Eye of God" system as an example, the 8M CIS chip it is equipped with boasts powerful performance advantages, enabling precise recognition of target objects within a range of up to 200 meters, providing solid and reliable visual support for intelligent driving. Similarly, Tesla's full-vision solution maximizes the role of multiple high-resolution cameras, constructing a precise and accurate three-dimensional environmental model to assist vehicles in achieving safe and efficient autonomous driving in complex and variable road conditions.

Despite signs of oversupply in certain categories of chips, such as microcontroller units (MCU) and power management integrated circuits (PMIC), from a macro perspective of the global automotive chip market by 2025, the shortage of CIS chips stands out as particularly pronounced, exhibiting typical characteristics of "structural shortages." A deeper analysis of its core causes can be summarized into three key factors:

First, the continuous rise in the penetration rate of autonomous driving is undoubtedly one of the key drivers behind the surge in demand for CIS chips. According to predictions from DIGITIMES Research, by 2027, the average number of cameras per vehicle will steadily increase from 9 in 2024 to 13, while for Level 5 vehicles pursuing an ultimate autonomous driving experience, the number of various sensors required may even exceed 30. Data released by market research firm Yole indicates that the global automotive CIS market size will soar from $2.3 billion in 2023 to $3.2 billion by 2029, with a compound annual growth rate of 5.7% during this period Behind the growth, it is not just a simple increase in the number of chips, but more importantly, the enhancement of added value brought about by the rapid iteration and upgrading of technology. Nowadays, a series of advanced functions such as High Dynamic Range (HDR) and LED Flicker Mitigation (LFM) are continuously integrated into CIS chips, which undoubtedly further raises the technical threshold in this field, making CIS chips increasingly crucial in the entire automotive industry chain.

Secondly, the immense pressure generated by technological iteration has become another important driving force for sustained demand growth. In actual driving scenarios, the widespread application of HDR and LFM technologies is particularly urgent. For example, during nighttime driving, vehicles need to accurately suppress the strong light interference from oncoming headlights to ensure that their cameras can clearly capture the road conditions ahead; in adverse weather conditions such as rain and fog, CIS chips must possess strong capabilities to penetrate water mist interference, providing reliable visual information for drivers. Clearly, traditional low-resolution sensors are already struggling to meet the increasingly stringent driving safety requirements in the face of these complex and variable lighting and road conditions.

At the same time, the push from regulatory policies has also become the third key driving force for demand growth. The EU NCAP 2025 new regulations explicitly require that new cars must be equipped with Automatic Emergency Braking (AEB) systems as standard, while in China, the "Guidelines for the Access Management of Intelligent Connected Vehicles" also stipulate that L3 level vehicles must be equipped with high-resolution vision modules. These binding regulatory provisions directly stimulate the strong demand from automotive manufacturers for high-resolution CIS chips from a policy perspective.

Looking at the overall development trend of the automotive industry, the intelligent transformation has undoubtedly become the unwavering core strategic layout direction of major automakers. At the influential 2025 CES exhibition, traditional automotive giants such as Volkswagen and General Motors showcased full-scene intelligent driving solutions based on 8M CIS chips, demonstrating their profound technical foundation and grand strategic blueprint in the field of intelligent driving. Meanwhile, emerging brands like XPeng are also not to be outdone, quickly stirring the market with their unique high-spec low-price strategy, accelerating their market penetration process.

Concentrated Supply Pattern, Limited Production Capacity

Currently, the global production capacity of 8M CIS chips is highly concentrated, almost monopolized by three industry giants: OmniVision Technologies (a subsidiary of Weir Group), Sony, and ON Semiconductor.

Among them, OmniVision Technologies leads the market with an impressive 43% shipment share due to its outstanding technical strength and market development capabilities. However, even so, its limited production capacity has long been occupied by the massive orders from numerous automakers such as BYD and Xiaomi, reaching near saturation.

While Sony and ON Semiconductor can supplement market supply to a certain extent, their product prices are generally higher than domestic solutions, which undoubtedly forces many automakers to weigh the pros and cons between cost and performance when procuring.

OmniVision Technology's ADAS Automotive Image Sensors (Partial Products)

What makes the situation more challenging is that the delivery cycles of these two giants can last up to 36 weeks, which undoubtedly adds to the woes of automakers urgently in need of chips. In contrast, domestic manufacturers such as Sitaro and Gekewei have successfully achieved mass production breakthroughs for 1M-8M CIS chips, and Sitaro has some products that have passed certain levels of automotive certification. However, overall, to ensure that the entire range of 1M-8M products meets higher-level certifications, a significant amount of testing and validation work is required, which is difficult to complete in a short time. Gekewei's progress in automotive certification is relatively slower and still needs time to advance the certification process to enter the original equipment market. This makes it challenging for them to quickly fill the significant market gap.

A deeper exploration of the root causes of supply chain vulnerabilities mainly reflects three key aspects: high certification barriers, severe dependence on foundries, and prominent cost dilemmas.

Automotive-grade CIS must undergo tests such as temperature cycling from -40°C to 125°C, vibration shock, and electromagnetic compatibility, with a certification cycle lasting 2-3 years. OmniVision's OX08A10 has passed ASIL-B functional safety certification and can support L4 level systems; however, domestic manufacturers are still in the sample testing phase, with some companies' 8M CIS only passing AEC-Q100 Level 3 certification, which cannot meet extreme working condition requirements.

In terms of cost, the mainstream process for automotive-grade CIS is 55nm, which lags behind the 45nm of mobile phone CIS. However, the stringent reliability requirements of automotive scenarios limit the speed of process upgrades. For example, the cost of 8M CIS is significantly higher than that of low-pixel products due to process complexity, and the yield may be lower. Automakers may respond to supply-demand imbalances through prepayment orders or by switching to lower-tier CIS.

A deeper issue lies in the regional fragmentation of the supply chain. The U.S. "CHIPS and Science Act" restricts the export of high-end equipment, forcing Chinese companies to shift towards mature process R&D; Europe, on the other hand, has invested €43 billion through the "Chip Act" to support the local semiconductor industry, aiming to increase its global semiconductor manufacturing market share from 10% to 20% by 2030. This geopolitical competition exacerbates the complexity of technological standards and capacity allocation, further driving up supply chain management costs.

Technological Leap

In the short term, the severe shortage of CIS chips in 2025 will continue, undoubtedly bringing pressure and challenges to the entire automotive industry. However, looking further ahead, with continuous technological iteration and innovation, as well as deep adjustments and optimizations in the supply chain, the future may reshape the entire industry landscape.

In terms of technological development direction, the R&D work for 12-megapixel CIS chips has already been launched at full speed, with a clear core goal aimed at further breaking through detection distance limits, striving to achieve ultra-long detection ranges of over 300 meters to perfectly meet the stringent requirements of L4 level autonomous driving. For example, the AR0820AT sensor launched by ON Semiconductor innovatively adopts back-illuminated (BSI) technology, which doubles the signal-to-noise ratio compared to traditional technologies, significantly improving the imaging quality of the chip under complex lighting conditions At the same time, cutting-edge technology fields such as photonic integration and quantum dot materials have also shown great potential and are expected to become key tools for breaking through the energy efficiency bottleneck of CIS chips. Currently, SmartSens and Jinghe Integrated are collaborating on key technologies such as CIS Stacked processes, and breakthroughs could enhance the competitiveness of domestic high-end CIS.

In terms of capacity release, TSMC's 2nm process is expected to promote the heterogeneous integration of CIS chips and AI accelerators, significantly improving the overall performance and computational efficiency of chips. However, due to the high reliance of automotive chips on the reliability of mature processes, this may somewhat delay the application speed of new processes. Taking Tesla's HW3.0 as an example, it still uses the relatively mature 14nm process, and even compatibility testing for 55nm CIS chips requires several months, which fully illustrates the caution and conservatism faced by automotive chips during the process upgrade.

From an industry ecosystem perspective, the cooperation model between automakers and chip manufacturers is being restructured. For example, XPeng plans to develop its own smart driving chips, while Li Auto is investing in silicon carbide (SiC) power modules and promoting their integration with electric drive systems. This trend of vertical integration will shift the supply chain from "production-driven" to "demand-driven," but it also raises higher requirements for the funding and technological reserves of small and medium-sized manufacturers.

Conclusion

The essence of the shortage of automotive CIS is a reflection of the mismatch between intelligent demand and supply chain resilience. In the future, whoever can achieve a "dual leap in pixels and capacity" first will occupy a commanding position in the visual revolution of autonomous driving.

From a global perspective, the intelligent transformation has entered the "deep water zone." As Hou Fushen, Vice Chairman and Secretary-General of the China Society of Automotive Engineers, stated, the automotive industry needs to "strengthen top-level design, coordinate technology, supply chain, and market layout," and the CIS shortage crisis provides the industry with an opportunity for reflection and restructuring. Through policy guidance, capital investment, and cross-border cooperation, China is expected to achieve a leap from "catching up" to "running alongside" in the automotive chip field, ultimately gaining the initiative in the global automotive intelligence competition.

Author of this article: Fang Yuan, Source: Semiconductor Industry Overview, Original title: "Automotive CIS Chips, a 'Chip' Hard to Find"

Risk Warning and Disclaimer

The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial conditions, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances Invest based on this, and you bear the responsibility