Power and Edge Computing Semiconductors for Industrial Infrastructure

ROHM Semiconductor is exhibiting its latest analog and power semiconductor architectures, including the Solist-AI edge computing platform, to address the hardware demands of artificial intelligence servers and electric mobility powertrains. This technology integration targets energy efficiency and signal processing speed within the automotive data ecosystem and high-density industrial server applications.

High-Density Power Architectures for AI Servers
At electronica 2026 (November 10 to 13 in Munich, Germany, Hall C3, Booth 520), the company is highlighting power delivery systems engineered specifically for AI infrastructure. As data centers scale up processing capabilities, they require advanced power management integrated circuits (PMICs) and wide-bandgap semiconductors to handle escalating thermal loads and power densities. These components facilitate efficient voltage conversion from the main power bus to the processors, minimizing energy dissipation and thermal output during continuous machine learning computing cycles.

Edge Computing and Automotive Powertrain Integration
The showcase incorporates the Solist-AI platform, an edge computing solution designed to execute machine learning algorithms directly at the local sensor level, which reduces bandwidth dependency and latency within complex digital supply chains. For the automotive sector, the technology portfolio extends to electronic powertrains (E-Powertrains), Advanced Driver Assistance Systems (ADAS), and solid-state lighting controls. These integrated circuits are structured to manage high electrical currents and switching frequencies in electric vehicles while maintaining the strict functional safety tolerances required by modern automotive control loops.

Industrial Terahertz Applications and Strategic Development
The hardware demonstrations also feature industrial applications for terahertz technology, which utilizes specific sub-millimeter electromagnetic frequencies for high-resolution, non-destructive material inspection and sensing. Commenting on the technical exchange driving these product developments, Wolfram Harnack, President of ROHM Semiconductor Europe, stated, "Beyond showcasing technology, we value the meaningful discussions that help shape future innovations and partnerships." This operational dialogue is supported by the presence of international engineering management, aiming to align future semiconductor fabrication with emerging industrial specifications.

Additional Context
This section details technical specifications and competitive benchmarking not included in the original news release.

In the market for AI server power delivery, semiconductor manufacturers are transitioning to 48-volt rack architectures to minimize distribution losses across the server backplane. ROHM’s power management solutions compete with advanced smart power stages from Infineon Technologies and Monolithic Power Systems (MPS), which establish accepted industry benchmarks for power density and thermal efficiency in high-current GPU clusters.

In the edge computing sector, the Solist-AI framework operates in the same ecosystem as NXP’s eIQ machine learning environment and STMicroelectronics’ STM32Cube.AI. These edge AI platforms are typically evaluated based on their inference speed per milliwatt and their ability to run quantized neural networks locally without relying on continuous cloud connectivity. Furthermore, for electric vehicle powertrains, ROHM is a primary supplier of Silicon Carbide (SiC) MOSFETs, competing directly with Wolfspeed, ON Semiconductor, and STMicroelectronics. Benchmark criteria for these automotive wide-bandgap components focus on specific on-resistance (RDS(on)) and switching loss metrics, which directly influence the thermal management requirements and maximum battery range of the vehicle.

Edited by Aishwarya Mambet, Induportals Editor, with AI assistance.

www.rohm.com