SEOUL, South Korea, Jan. 10, 2024  — SEMIFIVE, a leading design solution provider and pioneer of platform-based custom silicon solutions, has announced a collaboration with HyperAccel to develop a generative AI chip. The aim of this collaboration is to develop the product by applying Samsung Foundry’s 4nm process technology.

HyperAccel, a South Korea-based startup that was founded in January 2023, develops a Latency Processing Unit (LPU), an AI chip specialized for transformer-based Large Language Model (LLM). HyperAccel has developed model parallelism technology to efficiently distribute LLM to multiple LPUs and its own networking technology for data synchronization between LPUs. Recently, it has released an Orion server mounted with its own LPUs. The Orion server is dedicated for emerging generative AI workloads based on LLM using Field-Programmable Gate Array (FPGA) technology.

“HyperAccel is among a select few companies that possess true understanding required to build a successful custom AI chip and system in the generative AI landscape,” said Brandon Cho, CEO and co-founder of SEMIFIVE. “We are excited to work with HyperAccel and its exceptional engineering team to help them bring their innovation based on SEMIFIVE SoC Platform.”

SEMIFIVE, a company specializing in SoC platforms and ASIC design solutions, develops SoC design platforms expert for AI chips. To date, it has developed 3 SoC design platforms and has completed 8 custom silicon tapeouts using the platforms. SEMIFIVE plans to expand its roadmap to respond to customer demand for AI custom silicon. Most recently, SEMIFIVE announced that it has begun the mass production of an AI inference custom chip for servers designed using its 14nm AI SoC platform. Also, the mass production of NPU chip targeting HPC applications, using SEMIFIVE’s 5nm HPC SoC platform is scheduled in the first half of 2024.