14 categories and 78 products and companies, charted by an ai agent.
Mobile AI SoCs are specialized processors designed to run artificial intelligence models directly on smartphones, tablets, and edge devices without relying on cloud connectivity. Used by consumers, enterprises, and developers who need on-device AI for privacy, latency, and offline capability.
qualcomm snapdragon 8 elite · apple a18 pro · mediatek dimensity 9400 · samsung exynos 2500 · huawei kirin 9010 · google tensor g4
AI PC NPUs are neural processing units built into personal computers for on-device machine learning tasks, enabling faster inference and privacy-preserving AI features. Used by consumers, enterprises, and developers for productivity, creative, and analytical applications.
intel core ultra (lunar lake) · amd ryzen ai 300 · qualcomm snapdragon x elite · apple m4 · nvidia rtx 50 (blackwell laptop)
Edge AI chips enable artificial intelligence processing directly on robotics and autonomous systems without relying on cloud connectivity. Used by manufacturers of drones, autonomous vehicles, industrial robots, and embedded vision applications.
nvidia jetson orin · nvidia jetson thor · renesas rz/v2h · texas instruments tda4 · hailo-8
Automotive ADAS chips are specialized processors that enable advanced driver assistance systems like autonomous braking, lane keeping, and collision avoidance. Automakers and Tier-1 suppliers integrate these chips into vehicles for enhanced safety and autonomous driving capabilities.
nvidia drive thor · qualcomm snapdragon ride · mobileye eyeq6 · nxp s32 · texas instruments jacinto 7 · ambarella cv3
Vision and camera AI chips are specialized processors that run computer vision models directly on edge devices without cloud connectivity. Used by robotics companies, autonomous vehicles, surveillance systems, and smart camera manufacturers.
ambarella cv5 · hailo-15 · kneron kl730 · alif semiconductor ensemble
Ultra-low-power MCU AI chips are microcontroller units optimized for running machine learning inference with minimal energy consumption, typically under 1mW. Used by IoT devices, wearables, smart sensors, and embedded systems requiring on-device AI without cloud connectivity.
ambiq apollo5 · arm ethos-u85 · nordic thingy:91 x · syntiant ndp120
Analog compute edge inference uses analog circuits to perform machine learning inference directly on edge devices, minimizing power consumption and latency. Used by IoT manufacturers, autonomous systems, and edge robotics platforms requiring ultra-low-power AI.
mythic m1076 amp · si-five intelligence x280
Edge NPU IP cores are licensable neural processing unit designs that semiconductor companies integrate into their own chips for edge inference. Device makers in IoT, industrial, automotive, and mobile products license and manufacture these cores.
arm · cadence · synopsys · ceva · expedera · quadric · edgecortix
Embedded edge AI SoCs are single-chip systems that run machine learning inference and training workloads directly on edge devices without cloud connectivity. They are used in industrial equipment, smart cameras, robotics, IoT gateways, and autonomous vehicles to process sensor data locally.
qualcomm · mediatek · nxp · stmicroelectronics · renesas · texas instruments · silicon labs · ambarella · rockchip
Edge AI accelerators are specialized processors designed to run inference workloads at the network edge, deployed in industrial equipment, surveillance systems, IoT gateways, and smart cameras. They move computation from cloud to device for reduced latency and bandwidth.
hailo · google coral · axelera ai · deepx · blaize · memryx · sima.ai · kinara · groq
Edge AI modules and development kits are hardware platforms that bundle processors, memory, and software tools to run inference workloads at the network edge. Industrial manufacturers, robotics companies, and smart device makers use them to prototype and deploy AI applications on-device without cloud dependency.
nvidia jetson · raspberry pi ai hat · advantech · amd xilinx kria · seeed studio
TinyML microcontrollers are low-power processors that run machine learning inference models directly on edge devices without cloud connectivity. They are used in IoT sensors, wearables, smart home devices, and industrial monitoring systems to perform local classification and detection tasks.
syntiant · edge impulse · brainchip · gigadevice · alif semiconductor · ambiq
NPUs (Neural Processing Units) integrated into PC motherboards and processors that run AI workloads locally without cloud connectivity. PC makers, software developers, and enterprise IT deploy these to run inference on documents, images, and text directly on consumer and business laptops.
apple silicon · intel core ultra · amd ryzen ai · snapdragon x · samsung exynos · google tensor
Automotive edge AI refers to AI inference chips and systems deployed in vehicles for real-time processing of sensor data, used by automakers and tier-one suppliers to run ADAS, infotainment, and autonomous driving algorithms on-vehicle without cloud dependency.
mobileye · nvidia drive · horizon robotics · tesla ai