Semiconductor market news(Mar.10 to Mar.16)

1.Xiaomi CyberOne Robot is Moving Towards Mass Production

Recently, the Xiaomi Robotics Team announced that CyberOne is being phased into the production line in Yizhuang. It plans to publicize the mass production progress from March to April, open for on-site visits from April to May, and conduct PR promotion in the second half of the year. The Xiaomi Robotics Team has advanced to the 4th generation, with over twenty prototypes completed and all of them have been put into sample workstations for data collection. The 4th generation product is designed for standard operations, such as task scenarios like material handling and equipment inspection, and it can currently handle long - range tasks. Previously, Lei Jun disclosed that the unit cost of CyberOne is between 600,000 and 700,000 yuan.

2. Samsung Electronics is Developing the Next - Generation Packaging Material 'Glass Interposer'

The Device Solutions (DS) division of Samsung Electronics has initiated the development of the next - generation packaging material, the 'glass interposer'. The aim is not only to replace the costly silicon interposer but also to enhance chip performance. Reportedly, Samsung Electronics has recently received a joint proposal from the Australian materials supplier Chemtronics and the South Korean equipment manufacturer Philoptics for the development of the glass interposer. It is understood that Samsung Electronics is considering entrusting these companies to produce the glass interposer using Corning glass.

Meanwhile, Samsung Electro - Mechanics, a subsidiary of Samsung Electronics, is dedicated to the development of glass carriers (also known as glass substrates) and plans to commence mass production the year after next. These two concurrent R & D efforts are fostering internal competition, with the hope that this will boost semiconductor productivity and innovation.

?

3.Arm Launches the World's First Armv9 Edge AI Computing Platform, Leading a New Round of IoT Innovation

Driven by innovative technologies such as DeepSeek, the explosive growth of AI in the field of edge computing has shown initial signs this year. Against this backdrop, Arm recently introduced the world's first edge AI computing platform based on the Armv9 architecture for the Internet of Things (IoT) market. This computing platform can support the operation of end - side AI models with over one billion parameters, which undoubtedly presses the acceleration button for the performance improvement of IoT chips.

As Paul Williamson, Senior Vice President and General Manager of the IoT Business Unit at Arm, put it, "The innovation of AI is no longer limited to the cloud. With the increasing connectivity and intelligence of the world, from smart cities to industrial automation, processing AI workloads at the edge not only brings significant advantages but is also an essential necessity."

?

4.Teradyne Officially Announces the Acquisition of Quantifi Photonics, Strongly Positioning in the Photonic IC Testing Field

Recently, Teradyne announced that it has signed a definitive agreement to acquire Quantifi Photonics, a leading private company in the field of photonic IC testing. The acquisition is expected to be completed in the second quarter of 2025, subject to customary closing conditions and regulatory approvals.

This acquisition will enable Teradyne to provide scalable testing solutions for photonic integrated circuits (PICs). PIC technology, which utilizes wafer - based manufacturing, multi - chip integration, and advanced packaging with high - speed I/O interfaces, empowers the rapidly evolving high - performance computing market to support AI workloads.

5.NXP's New - Generation MCU Embeds New MRAM Storage

Recently, NXP Semiconductors launched its new - generation automotive S32K5 series MCUs, embedding a new type of storage, MRAM (Magnetoresistive Random - Access Memory). This is the automotive industry's first 16nm FinFET MCU with embedded MRAM.

The S32K5 series adopts an Arm core and has an operating speed of up to 800 MHz. It integrates an Ethernet switch core shared with NXP's S32N automotive processor. Additionally, it is equipped with a dedicated eIQ Neutron neural processing unit (NPU), enabling machine - learning algorithms to perform high - efficiency real - time processing of vehicle - edge sensor data.