< Manufacturers

Inference & Compute Manufacturers


Inference & Compute is the digital cortex of autonomy and electrification. This page catalogs manufacturers of automotive AI SoCs, edge compute modules, training & simulation platforms, and toolchains that power perception, planning, and control across vehicles and robots. Data-center training clusters and governance appear on partner nodes (DatacentersX, 137AI).

Company Headquarters Platform / Silicon Focus Primary Applications
AMD United States Versal AI • Ryzen Embedded Heterogeneous AI/CPU/FPGA ADAS, cockpit, domain control
Ambarella United States CVflow (CV3) Computer vision SoCs ADAS, surround, perception
Horizon Robotics China Journey Series Automotive AI SoC L2–L4 driver assistance
Huawei China Ascend (Atlas) • MDC Edge AI, driving compute ADS stack, intelligent cockpit
Intel United States Core/Atom Automotive • Flex CPU + accelerator mix Cockpit, ADAS offload
Mobileye Israel EyeQ • EyeQ Ultra ADAS/AV compute + stack Tier-1 supply to OEMs
NVIDIA United States DRIVE Orin • DRIVE Thor Automotive AI SoC Centralized vehicle compute
NXP Netherlands BlueBox • S32 family Automotive controllers + AI Domain/control, safety MCU
Qualcomm United States Snapdragon Ride / Ride Flex ADAS/AV + cockpit fusion Central compute, telematics
Renesas Japan R-Car H3/H4 • R-Car S Automotive SoC + safety MCU ADAS, gateway, body domain
Tenstorrent Canada Grayskull/Blackhole IP Licensable AI cores Automotive inference IP
Tesla United States FSD Computer • Dojo D1 In-vehicle inference + training Autonomy stack, training clusters

Inference & Edge Compute Modules (A–Z)

Automotive-grade boards and domain controllers integrating the SoCs above with memory, I/O, and sensor interfaces. These platforms accelerate OEM time-to-market for L2–L4 functions.

Vendor Headquarters Module / Platform Compute Base Target Use
NVIDIA DRIVE United States Drive AGX Orin • Drive Thor Orin / Thor SoC Central vehicle computer
Qualcomm Ride United States Ride / Ride Flex Platforms Snapdragon Ride ADAS/AV + cockpit fusion
Mobileye Israel EyeQ Ultra Development Platform EyeQ Ultra Hands-off/eyes-off AD
Renesas Japan R-Car Starter / Dev Platforms R-Car SoC Domain control, gateway
NXP BlueBox Netherlands BlueBox 3.0 NXP S32 + accelerators ADAS, autonomous dev
Ambarella United States CV3 Reference Platforms CVflow (CV3) Perception domain ECUs
Tesla United States FSD Computer (HW 3/4) Tesla in-house SoC Consumer AV, robotaxi

Training & Simulation Platforms (A–Z)

Vendors providing synthetic data generation, closed-loop simulation, and validation environments to develop and test autonomous stacks prior to deployment.

Company Headquarters Platform Focus Primary Users
NVIDIA United States Omniverse • DRIVE Sim Digital twins, AV simulation OEMs, Tier-1s, robotics
Applied Intuition United States Applied Suite Scenario, validation, analytics AV programs, defense, robotics
Cognata Israel Cognata Simulation Synthetic data, scenario gen ADAS/AV development
Foretellix Israel Foretify Safety, coverage-driven testing OEMs, AV verification
dSPACE Germany ASM • AURELION • HIL HIL test, sensor simulation Tier-1s, validation labs
MathWorks United States MATLAB/Simulink Model-based design, test OEMs, R&D, academia

Toolchains & Middleware (A–Z)

Foundational SDKs and middleware enabling high-performance inference, robotics orchestration, and safety-aligned deployment on automotive compute.

Vendor Headquarters Stack / SDK Capability Where It’s Used
NVIDIA United States CUDA • TensorRT • DriveWorks • DriveOS Acceleration, inference, AV SDK DRIVE, Jetson, datacenters
Qualcomm United States Ride SDK ADAS/AV middleware Snapdragon Ride platforms
Mobileye Israel RSS • Responsibility-Sensitive Safety Safety model, policy layer ADAS/AV deployments
Autoware Foundation Global Autoware • ROS 2-based stack Open-source AV stack Shuttles, research AVs
Apex.AI United States Apex.OS (ROS 2 safety) Safety-certified middleware Automotive, industrial robots
Open Robotics / ROS 2 Global ROS 2 Robotics framework Robots, AMRs, research