Thermal management specialist MHS announced on July 22 that it has completed the development of thermal management components for Mobilint's MLA400 and signed a mass production supply contract.
The MLA400 is a PCIe-based AI accelerator developed by Mobilint, built on its proprietary NPU 'ARIES.' It is designed to handle various inference tasks, including computer vision, AI video surveillance, and large language models (LLMs) and vision-language models (VLMs).
MHS is responsible for developing heat sinks and other thermal management components for both active and passive products of the MLA400. The two companies have verified heat dissipation performance and mechanical stability from the early stages of development. By taking on both development and mass production, MHS aims to minimize potential errors in component design changes and production processes.
AI semiconductors can experience throttling, which reduces processing speed, or shortened component lifespan when operating temperatures rise. Heat sinks and cooling structures play a crucial role in determining not only the chip's instantaneous performance but also long-term operational stability and system reliability.
Lim Jong-soo, CEO of MHS, stated, "We collaborated throughout the entire development and mass production process to ensure the MLA400 achieves stable performance and quality. We will continue to serve as a partner supporting Mobilint's product development and business expansion."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.