SEMIFIVE, a leader in custom silicon solutions, has announced a collaboration with HyperAccel to develop and mass produce Bertha, a groundbreaking generative AI chip. This partnership marks a significant step forward in the AI hardware landscape, with mass production slated to begin in the first quarter of 2026.
Bertha is built upon a 4nm process technology and leverages HyperAccel’s innovative LLM Processing Unit (LPU), a specialized AI chip designed specifically for transformer-based Large Language Models (LLMs). This groundbreaking LPU is the world’s first semiconductor designed for LLM inference, promising low cost, reduced latency, and domain-specific features. By replacing traditional, high-cost and inefficient Graphics Processing Units (GPUs), Bertha is expected to deliver performance improvements of up to 2 times and a remarkable 19 times better price-to-performance ratio compared to typical supercomputers.
SEMIFIVE’s expertise lies in SoC platforms and ASIC design solutions, with a focus on developing expert SoC design platforms for AI chips. The company is actively expanding its roadmap to meet the growing demand for AI custom silicon, and this partnership with HyperAccel is a testament to this commitment.
“We are thrilled to partner with SEMIFIVE, a leading provider of SoC platforms and comprehensive ASIC design solutions, for the development of Bertha for mass production,” said Joo-Young Kim, CEO of HyperAccel. “By collaborating with SEMIFIVE, we are excited to offer customers AI semiconductors that provide more cost-effective and power-efficient LLM features than GPU platforms. This advancement will significantly reduce the operational expenses of data centers and expand our business scope to other industries that require LLMs.”
“HyperAccel boasts the most efficient and scalable LPU technology for LLMs. As the demand for LLM computation is skyrocketing, HyperAccel has the potential to become a new powerhouse in the global processor infrastructure,” said Brandon Cho, CEO and co-founder of SEMIFIVE. “We are extremely excited to be the production partner for HyperAccel’s groundbreaking Bertha chip, and we are thrilled to contribute to yet another innovative success story built on the SEMIFIVE platform.”
This partnership between SEMIFIVE and HyperAccel promises to redefine the landscape of AI hardware. Bertha’s innovative design and impressive capabilities are poised to unlock a new era of efficient and cost-effective AI computation, paving the way for wider adoption of advanced AI technologies across various industries.
SEMIFIVE is a pioneer in platform-based SoC design, collaborating with customers to bring innovative ideas to life through custom silicon. Their SoC platforms provide a robust foundation for new chip designs, leveraging configurable domain-specific architectures and pre-validated key IP pools. SEMIFIVE offers comprehensive end-to-end solutions for custom SoCs, enabling faster development and reduced costs for key applications such as data centers and AI-enabled IoT devices. Their strong partnership with Samsung Foundry and their commitment to a larger ecosystem solidifies their position as a one-stop solution for any SoC design needs.
