ZoZo Auto Research and Shuiqing Muhua Research Center jointly released the "2026 Autonomous Driving Sensor Chip Research Report".
In 2026, the autonomous driving sensor chip industry is ushering in a critical turning point. With the official implementation of L3 autonomous driving access policies, advanced intelligent driving has entered a stage of large-scale application. This report focuses on the evolution of sensor chip technology trends. It mainly elaborates on the technological direction of autonomous driving sensor chips under the three core changes: regulation-driven perception upgrades, the transition of L3 from pilot to large-scale deployment, and the hardware market entering a re-evaluation cycle. It systematically reviews the technological evolution, market competition, and industrial landscape of five major categories of sensor chips: in-vehicle cameras, LiDAR, millimeter-wave radar, ultrasonic radar, and UWB near-field perception, covering the product dynamics of major suppliers and the application status of OEMs.
The technological evolution of the five major categories of sensor chips mainly revolves around the following core points:
Camera CIS Chips: Evolving from "single imaging" to an intelligent perception terminal with "integrated sensing and computing";
LiDAR Chips: Entering a new era of the "thousands of lines era" and "6D full-color perception";
Millimeter-Wave Radar Chips: Transitioning from "edge processing" to "satellite architecture" and "high channel count";
Ultrasonic and UWB Chips: Upgrading from "auxiliary perception" to "full-scenario collaborative nodes".
Camera Sensor Chips: Resolution Enhancement and Intelligent Fusion
In 2026, the in-vehicle camera chip market presents three major trends:
01. Large-Scale Popularization of BSI and Stacked Structures
CIS is divided into Front-Side Illumination (FSI), Back-Side Illumination (BSI), and Stacked structures based on structural technology, with the stacked structure being the future mainstream direction. In-vehicle CIS has higher requirements for image resolution, dynamic range, low-light imaging capability, and multi-camera collaborative processing capability, and must meet automotive-grade certification and functional safety requirements. In 2025, the proportion of BSI + 3D stacked structures in global mass-produced CIS has exceeded 75%, and the unit pixel size has broken through 0.6μm.
Taking SmartSens SC860AT as an example: The SC860AT is built on SmartSens' brand-new CARSens®-XR Gen 2 technology platform. It adopts a single-pixel architecture design, equipped with advanced technologies such as Lofic HDR®2.0 and SFCPixel®. It supports the AB-Exposure™ dual-frame exposure control function and has performance advantages such as high frame rate, high sensitivity, and high dynamic range. It can adapt to the new image upgrade needs of applications such as in-vehicle front-view, side-view, rear-view, Occupant Monitoring System (OMS) cameras, and E-Mirror electronic rearview mirrors.
02. High Resolution: 8M Has Become the Standard, 12M Is Accelerating
In 2026, with the popularization of Advanced Driver Assistance Systems (ADAS), 8-megapixel sensors have become the standard for advanced intelligent driving, and 12-megapixel products have also begun to emerge.
Taking OmniVision's OX08D20 as an example: This 8-megapixel CMOS sensor is specifically designed for automotive exterior cameras in Advanced Driver Assistance Systems (ADAS) and Autonomous Driving (AD). By introducing an innovative capture solution developed in collaboration with Mobileye, it increases the frame rate to 60 frames per second and adopts an a-CSP package, reducing its size by 50% compared to similar exterior sensors.
03. Fusion of HDR and LFM Technologies - 140dB+ Becomes a Hard Threshold
In Advanced Driver Assistance Systems (ADAS), High Dynamic Range (HDR) of over 140dB and LED Flicker Mitigation (LFM) have become hard technical thresholds for in-vehicle CMOS Image Sensors (CIS). This fusion technology aims to solve imaging challenges under extreme lighting conditions, ensuring that vehicles can accurately recognize traffic signals and avoid visual misjudgments.
Taking SmartSens SC126AT as an example: The SC126AT is built on the CARSens®-XR Plus technology platform. It adopts a single-pixel architecture design and is equipped with advantageous technologies such as SFCPixel® and PixGain HDR®. It has multiple performance advantages including high sensitivity, high dynamic range, and low noise, and supports on-chip ISP functions.
Representative Products of the Three Technological Trends in Camera Sensor Chips
Source: Shuiqing Muhua, ZoZo Auto Research Compilation
LiDAR Sensor Chips: The Era of Thousands of Lines and 6D Full Color
In 2026, the LiDAR sensor chip market presents three major trends:
01. Exponential Leap in Line Count from Hundreds to Thousands
In April 2026, RoboSense released the world's first native 2160-line automotive-grade SPAD-SoC "Phoenix" chip; Hesai released the "Picasso" 6D full-color chip supporting up to 4320 lines; Huawei launched an 896-line dual-optical-path image-level LiDAR. The perception capability of in-vehicle LiDAR has officially stepped from the "point cloud level" to the "image level".
Taking RoboSense's dual flagship chips as an example:
As the world's first automotive-grade SPAD-SoC integrating a native 2160 lines on a single chip, the "Phoenix" chip achieves ultra-high resolution of over 4 million pixels and ultra-long-range detection capability of 600 meters, elevating the perception fineness and range of LiDAR to a completely new height. Currently, this chip has obtained automotive-grade certification and is planned to be mass-produced and installed in vehicles first within 2026, providing an ultra-HD perception cornerstone for advanced intelligent driving above L3 level.
The "Peacock" chip is the industry's first mass-producible full-solid-state large-area array SPAD-SoC with a resolution of 640×480. It features an ultra-wide field of view of 180°×135° and millimeter-level accuracy, enabling close-range, dead-angle-free ultra-high-density point cloud coverage at an unprecedented cost-performance ratio. This chip is planned for mass production in the third quarter of 2026 and will mainly target the broad markets of in-vehicle side and blind-spot LiDAR, robotics, and spatial intelligence, promoting the large-scale popularization of high-performance perception.
02. Digital Architecture Comprehensively Replaces Analog Architecture
From 2024 to 2025, the in-vehicle LiDAR industry underwent a generational change in core architecture. The digital architecture, represented by SPAD-SoC (Single-Photon Avalanche Diode - System on Chip), has comprehensively replaced the traditional analog architecture, becoming the standard solution for high-performance in-vehicle LiDAR.
The SPAD-SoC adopts an "integrated sensing and computing" architecture design, integrating the Single-Photon Avalanche Diode (SPAD) array and a complete digital signal processing system on the same chip. The SPAD-SoC can simultaneously complete the entire process of photon detection, time measurement, signal processing, and data output, enabling direct output from photon incidence to depth/image data.
Compared with traditional discrete solutions, the SPAD-SoC has advantages such as high integration, low power consumption, small size, and fast response speed. It is the core component of current high-line long-range LiDAR and Flash short-range LiDAR, and has the greatest impact on LiDAR performance. Therefore, it has become a technological high ground that LiDAR manufacturers are competing to seize.
Representative Products of LiDAR Sensor Chip SPAD-SoC
Source: Shuiqing Muhua & ZoZo Auto Research "2026 Autonomous Driving Sensor Chip Research Report"
03. Chip-Level 6D Full-Color Perception
In 2026, the chip-level 6D full-color perception marks the evolution of LiDAR from a "colorblind surveyor" to a "full-color observer", upgrading from simply outputting 3D spatial coordinates (XYZ) to the six-dimensional perception capability of synchronously outputting spatial coordinates and color information (RGB).
Achieving pixel-level native fusion of color sensing (RGB) and ToF ranging (XYZ) at the chip level directly generates color point clouds. This means that from the moment of its birth, every data point is naturally bound to precise spatial position and real color information—no stitching, no calibration, and no guessing required.
Taking Hesai's 6D full-color LiDAR chip "Picasso" as an example: Hesai has directly integrated the RGB color sensing function into the receiving chip of the LiDAR. The Picasso SPAD-SoC chip uses a single chip to simultaneously complete the "full fusion" of near-infrared laser and visible light perception. Each of its pixel units can both emit lasers to measure distance and capture light to identify color. This means that at the moment the laser hits an object and reflects back, the chip not only knows where this point is but also knows what color it is.
Millimeter-Wave Radar Sensor Chips: Satellite Architecture and High Channels
01. Adaptation Trend of Satellite Architecture
In 2026, the Satellite Architecture is becoming the most core technological transformation direction in the field of in-vehicle millimeter-wave radar sensor chips. This architecture subverts the traditional design concept of "edge intelligence" and drives the evolution of millimeter-wave radar from an "independent black box" to an "open perception front-end".
The essence of the satellite architecture is a "division of labor mode" for the intelligent driving perception system. The radar, like a "satellite", focuses on front-end perception (RF transceiving and raw data collection), and all complex signal processing, target detection, and tracking algorithms are shifted to the central domain controller.
There are two main solutions for satellite radar:
Dedicated RSP IC Solution: Uses a dedicated radar bridge chip, the Radar Bridge, which offers high energy efficiency and high flexibility, led by radar module suppliers;
RSP IP Integration into ADAS SoC Solution: Integrates the radar signal processor IP into the ADAS System on Chip (SoC), possessing the highest level of integration, led by automakers.
02. Accelerated Popularization of 8T8R Imaging Radar
2026 has become the first year of mass production for 8T8R imaging radar chips. The intensive release of single-chip 8T8R solutions such as NXP's TEF8388 and Infineon's CTRX8188F is driving 4D millimeter-wave radar from a high-end option to large-scale popularization. Relying on lower cost, higher integration, and better power consumption, the 8T8R architecture is becoming a key bridge connecting the mid-range 6T8R and the high-end 16T16R/24T24R. The single-chip 8T8R solution highly integrates the RF front-end and signal processing through CMOS technology, finding the optimal balance point among performance, cost, and mass producibility. It is expected to be rapidly popularized in mainstream models priced around 150,000 RMB in the next three years, becoming a "safety rigid demand" for advanced intelligent driving perception systems.
Taking Infineon's RASIC™ CTRX8188F as an example: This chip is an 8-transmit, 8-receive automotive radar transceiver, targeting imaging radar and advanced ADAS applications. Such high-channel radar chips are not only used in traditional radar modules but also target central computing architectures, capable of sending richer raw radar data to backend computing platforms.
The CTRX8188F chip has multiple industry-leading indicators in RF performance. Its transmit power reaches 14.5dBm, ensuring the radar's long-range detection capability; the phase noise is as low as -100 dBc/Hz, providing excellent signal-to-noise ratio performance; the noise figure is 10.2 dB, achieving industry-leading receiving sensitivity; it supports an ultra-large bandwidth of 4GHz, enabling centimeter-level distance resolution; and the modulation speed is up to 200 MHz/µs, meeting the detection needs of high-speed moving targets.
Ultrasonic and UWB Near-Field Perception Chips: From Auxiliary to Core
01. Main Technological Trends of Ultrasonic Radar Sensor Chips: Fusion Architecture
Integration of Edge AI Inference Capability: In 2026, mainstream ultrasonic radar modules have integrated microcontrollers with NPUs (Neural Processing Units) internally, which are sufficient to support real-time inference of lightweight deep learning models.
Cross-Sensor Modal Alignment and Fusion: The collaborative capability of ultrasonic radar chips is deepening from "inter-sensor collaboration" to "multi-modal fusion". In the high-end intelligent driving architecture of 2026, the close-range high-precision distance information provided by ultrasonic radar, the rich texture information from cameras, and the velocity vector information from millimeter-wave radar are fused through a cross-modal attention mechanism based on the Transformer architecture.
02. Main Technological Trends of In-Vehicle UWB Near-Field Perception Chips: Multi-Function Fusion
With the technological combination of millimeter-wave and UWB, UWB technology is attempting to cover the "full-scenario security protection" for vehicle owners, from vehicle searching and unlocking, to obstacle avoidance during ADAS driving, and then to parking, door opening, and locking. This is not only a superposition of functions but also a reconstruction of system efficiency and cost structure. UWB is evolving from a "smart key" into a "smart nerve" that connects the perception networks inside and outside the vehicle.
From "Communication Chip" to "Integrated Communication and Perception SoC"
The "three-in-one" integration of chip architecture: Current mainstream UWB chips have evolved from communication chips that merely support ranging into a trinity SoC solution of "positioning + radar + communication".
From "Single-Point Perception" to "Collaborative Perception"
Traditional UWB radar anchors work independently, facing the problem of near-range blind spots. By introducing the Sensing technology defined by the IEEE 802.15.4ab standard, Calterah has realized a multi-station radar mode—transmitting from R1 and receiving from R2 at the rear of the vehicle, and vice versa—turning UWB radars that fight alone into a "networked radar" that fights collaboratively.
Taking Calterah's UWB series products as an example: Calterah has launched two mature development kits: a development kit for in-cabin CPD (Child Presence Detection), and a parking assistance system development kit based on 4ab Sensing. They have handed over a complete set of "Lego components" including hardware reference designs, software SDKs, AI models, and toolchains to Tier 1 suppliers and OEMs.
The development kits integrate multi-function fusion solutions such as digital keys, kick-sensor tailgates, in-cabin child presence detection, and exterior parking assistance, and even sentry mode, becoming engineering solutions that can be quickly imported. By reusing a set of UWB anchors on the vehicle body, multiple intelligent scenarios of perception and interconnection can be leveraged.
Contact Information Mobile number is the same as WeChat number
Industry Research Department | Mr. Fu 15810027571
Mr. Zhao 18702148304
Data Service Department | Ms. Zhang 13716037793
Strategic Consulting Department | Ms. Han 15810133447
Promotion and Communication Department | Ms. Liao 13718845418
Mr. Du 13910162318