FPT Software Joins Efforts to Develop First Large Language Model for Semiconductor Industry

SemiKong, the world’s first open-source large language model (LLM) tailored specifically for the semiconductor industry, has officially launched. This groundbreaking model aims to optimize semiconductor manufacturing processes and technology, with the goal of tapping into the 500-billion-dollar industry’s potential over the next five years.

This achievement results from the collaboration between members of the World AI Alliance, including AITOMATIC, Tokyo Electron (TEL), and FPT Software, a subsidiary of FPT Corporation. FPT Software has played a crucial role in building infrastructure resources and refining the semiconductor platform model for SemiKong, leveraging its strong technological capabilities and expertise.

SemiKong surpasses standard language models like GPT and Llama3 in accuracy, coherence, and understanding of semiconductor processes. Its advanced versions significantly enhance innovation and cost efficiency throughout the semiconductor value chain.

SemiKong can lower semiconductor production costs, ultimately benefiting consumers through reduced prices for smartphones, laptops, and smart home devices in the coming years.

Dr. Anthony Annunziata, Head of AI Open Innovation and the AI Alliance at IBM Research, noted: “SemiKong exemplifies the World AI Alliance's mission to accelerate the integration of innovative AI technologies into crucial industries like semiconductor manufacturing.”

Dr. Christopher Nguyen, CEO of AITOMATIC and Head of the SemiKong project, stated: "With our deep expertise in semiconductors, AI application solutions have the potential to address complex manufacturing challenges in ways that were previously unattainable."

Atsushi Suzuki, Director of DX Product Lifecycle Management at Tokyo Electron, commented: "SemiKong represents a significant advancement in applying AI to semiconductor manufacturing."

Daisuke Oku, Senior Expert at Tokyo Electron and advocate of the semiconductor industry model expressed confidence: "SemiKong marks the beginning of an exciting journey for open-source AI in the semiconductor industry."

With SemiKong, FPT Software aims to elevate service value and drive innovation for clients. It includes customizing AI solutions for specific industries and offering platform models for community use. Additionally, FPT Software plans to leverage SemiKong technology to create smart chip design assistants, enhancing capabilities for chip manufacturing clients.

Mr. Nguyen Xuan Phong, AI Director at FPT Software, stated, "FPT Software is excited to be part of this innovative initiative and is eager to see the potential outcomes from the synergy between semiconductor technology and AI."

SemiKong is available on HuggingFace and GitHub starting July 9, 2024. This platform allows semiconductor companies to create their expert language models using collective industry knowledge. A more advanced version of SemiKong is expected to be released in December 2024, and the initial process models are anticipated in September 2024.

Developers are committed to ongoing R&D efforts to create an ecosystem of AI tools that will usher the semiconductor industry into a new era and accelerate advancements in fields ranging from quantum computing to renewable energy technology.

For more information, visit SemiKong's official website at https://www.semikong.ai.

AI is indeed a key focus for FPT Corporation. By integrating AI across its products and solutions, FPT aims to drive innovation and enhance user experiences within its Made by FPT ecosystem. As part of this initiative, FPT Software is actively working on expanding its capabilities in AI through investments in human resources, R&D, and partnerships with leading organizations like AITOMATIC and Landing AI. Membership in the AI alliance initiated by IBM and Meta further strengthens its position. These efforts are aligned with FPT's ambitious goal to reach 5 billion USD by 2030 and solidify its status among the world's top billion-dollar IT companies.