SEMIFIVE Commences Mass Production of HyperAccel's AI Inference Accelerator 'Bertha'

SEMIFIVE, a leading provider of custom AI semiconductor solutions, has commenced mass production of a data center AI inference accelerator for HyperAccel. This marks a significant milestone as the company expands its portfolio into the data center AI market.

The new accelerator, dubbed 'Bertha,' is an AI inference chip optimized for large language models (LLMs) and is produced using Samsung Foundry's 4nm process node. With a die size exceeding 500 mm², it represents a substantial challenge in terms of power management, heat dissipation, and manufacturing yield. SEMIFIVE has successfully delivered a complete turnkey solution, encompassing front-end design and verification, as well as packaging, testing, and volume manufacturing supply.

This achievement is a critical step in SEMIFIVE's growth strategy, following the successful mass production of other key projects. In the first half of 2026, the company secured KRW 42.3 billion in new mass-production orders, nearly doubling its full-year order intake of KRW 21.2 billion from the previous year. Quarterly order intake surged by 71% from KRW 15.6 billion in Q1 to KRW 26.7 billion in Q2, with overseas orders accounting for 45% of Q2 bookings.

"The AI landscape is shifting from training to inference, and demand for ASIC accelerators in data centers is on the rise," said SEMIFIVE's CEO. "This successful mass production validates our unmatched execution capabilities in advanced processes. By combining HyperAccel's innovative architecture with SEMIFIVE's comprehensive turnkey capabilities, we have seamlessly executed the entire journey from design to volume production for a large-area die of more than 500 mm²."

According to the release, this mass production contract is part of a broader strategy to diversify SEMIFIVE's portfolio across various application domains and technical complexities. The company has already secured mass-production programs for an AI ASIC chip for Hanwha Vision's security cameras and a high-performance computing AI chip for a Japanese customer. These achievements demonstrate SEMIFIVE's ability to build stable, recurring revenue beyond one-time NRE projects.

SEMIFIVE's growth is driven by a virtuous cycle where design wins translate into mass-production revenue. The company has experienced rapid financial growth, with a robust order intake pipeline. CEO's commitment to carrying the strong mass-production momentum into the remainder of the year underscores the company's confidence in its future prospects.

As the AI inference market continues to grow, SEMIFIVE's successful execution of this large-scale project positions the company well to capitalize on the expanding demand for advanced AI solutions in data centers.