Rutronik introduces the Intel Core Ultra X7 for next-gen edge robotics

Rutronik now offers the Intel Core Ultra Series 3 processor family, highlighted by the Intel Core Ultra X7 Processor 358H, to accelerate the development of intelligent robotics and automation systems. Codenamed Panther Lake, the new architecture consolidates a CPU, GPU, and dedicated Neural Processing Unit (NPU) onto a single hardware platform. This integration enables autonomous mobile robots (AMRs) and collaborative robots to simultaneously process high-bandwidth sensor data and execute time-critical control tasks directly at the edge, eliminating the need for multiple specialized computing units.

High-performance AI and deterministic control
Delivering up to 180 TOPS of total AI performance, the new processors feature up to 16 CPU cores and an integrated GPU with 50 percent more cores than the previous Arrow Lake-H generation. This represents a 1.8-fold increase in AI processing capabilities, allowing complex machine vision and spatial computing models to run locally without cloud latency. To ensure deterministic operation for physical actuation, the platform incorporates Intel Time-Coordinated Computing (TCC) and Time-Sensitive Networking (TSN). This allows a single chip to map environments, detect obstacles, and drive precision motor controls synchronously, saving physical space and extending battery life through highly optimized performance-per-watt efficiency.

Industrial ecosystem and rapid deployment
To streamline system integration, the Intel Core Ultra X7 Processor 358H is supported by a broad ecosystem of embedded boards and box PCs from manufacturers including Kontron, Advantech, ASUS, DFI, and AAEON. Developers can leverage the Intel OpenVINO toolkit to efficiently optimize and deploy existing AI models across the processor's various compute engines. This scalable foundation simplifies the transition from prototyping to volume production for smart manufacturing facilities, automated quality control systems, and logistics automation networks.

Additional Context: Heterogeneous edge computing and thermal management
This section details technical specifications and architectural analysis not included in the original product announcement.

In autonomous mobile robotics (AMRs), traditional hardware architectures typically rely on a primary x86 industrial PC for kinematic control via fieldbus protocols, paired with a discrete GPU module to handle spatial AI algorithms like SLAM (Simultaneous Localization and Mapping). This dual-board approach severely drains the robot's battery and complicates thermal management. By utilizing a heterogeneous compute architecture, the Panther Lake platform intelligently distributes workloads internally: the NPU handles continuous, low-power inferencing for object detection, the integrated GPU processes dense 3D point-cloud data from LiDAR, and the CPU cores—supported by hardware-level TCC—manage the strict microsecond-level jitter requirements of the motor drives. This monolithic approach drastically reduces the overall Thermal Design Power (TDP) footprint, enabling engineers to design fully fanless, IP67-rated robotic controllers capable of operating continuously in harsh, dust-filled industrial environments without thermal throttling.

Edited by an industrial journalist, Lekshman Ramdas, with AI assistance.

www.rutronik.com