Tokyo Seimitsu Co., Ltd. and Advantest Corporation announced today that they will jointly develop a die-level prober.
Semiconductors are expected to become increasingly sophisticated and complex in the future. Close collaboration within the semiconductor value chain is essential to quickly respond to ever-evolving market needs and provide customers with high-performance total test solutions.
Leveraging their respective expertise, the two companies will jointly develop a die-level prober that will provide the advanced probing capabilities essential for testing AI and high-performance computing (HPC) devices.
To achieve the high computing power required for AI model training, inference, and generation, advanced 2.5D/3D packaging technologies are being adopted for AI/HPC devices such as GPUs and CPUs. These advanced packages generate large amounts of heat due to the massive data processing, making it difficult to control the temperature during testing.
Also Read: ON Semiconductor and FORVIA HELLA Expand Next-Gen Power Partnership
This joint development project will strengthen next-generation probing and handling technologies that address these challenges and contribute to the growth of the AI/HPC market.
Comment from Advantest Group CEO Douglas Lefever: Advantest is actively collaborating with key companies in the semiconductor value chain to solve customer challenges in semiconductor testing, including the rapidly growing AI/HPC market. Through collaboration with Tokyo Seimitsu, Advantest will realize high-performance, comprehensive test solutions that meet customers’ future needs in die-level probes.
Comment from Ryuichi Kimura, President and CEO of Tokyo Seimitsu: “In response to the advances in AI/HPC technology that are bringing about changes in the world, we are proposing a variety of solutions centered on precision positioning technology. By collaborating with Advantest to promote the joint development of die-level probers, we will open up new possibilities for next-generation semiconductor testing, which is required in the AI/HPC era.”
SOURCE: PRTimes

