On December 29, Aitomatic and its “AI Alliance” partners launched the world’s first open-source large language model (LLM) specifically designed for the semiconductor industry, SemiKong. The model aims to become an integral part of semiconductor design companies’ workflows, acting as a “veteran expert” to accelerate the development and market launch of new chips.
Aitomatic, the company behind SemiKong’s development, highlighted the urgent need for gathering expert knowledge in the semiconductor industry, as the retirement of many senior experts has intensified the knowledge gap, with several companies suffering from a severe lack of expertise. SemiKong, a large language model tailored to industry needs, is expected to provide a reliable way for new engineers to maintain competitiveness and quickly access specialized knowledge.
According to the company, SemiKong is a 70-billion-parameter model version trained on a vast amount of semiconductor industry knowledge, based on Meta’s Llama 3.1 platform. The large model was co-developed by Aitomatic and AI Alliance members, including tech giants such as Meta, AMD, and IBM, as well as academic institutions like Yale University and the University of Tokyo. Aitomatic’s DXA system serves as the core framework for deploying SemiKong.
DXA (Domain Expert Agent) integrates smaller language models with the core capabilities of SemiKong 70B. By training on customer companies’ technical libraries and projects provided by expert engineers, it effectively captures and builds the knowledge of senior engineers, ensuring that key expertise is preserved and expanded for future use, with adjustments made according to company needs. The trained DXA can further be integrated into the core SemiKong deployment to automatically execute development tasks or serve as an interactive interface between engineers and staff.
In its current 70-billion-parameter version, combined with the use of the smaller DXA model based on SemiKong, the large language model’s application value in the semiconductor field has far exceeded that of general-purpose AI models.
Aitomatic claims that SemiKong outperforms multiple closed-source language models in generating semiconductor-specific content and understanding complex processes. The model has provided tangible benefits, including reducing time-to-market for new chip designs by 20-30% and improving first-pass yield by 15-25%. These tools have also enhanced the onboarding process for new engineers, accelerating their learning curve by 40-50%. For example, the DXA supporting SemiKong reduced the time required for etching recipes from several hours to just a few minutes.
In summary, by introducing SemiKong and DXA, a comprehensive framework has been provided for semiconductor researchers to preserve knowledge, improve productivity, and drive innovation. These advancements have the potential to reshape semiconductor manufacturing, offering scalable and cost-effective solutions to address the complexities of the field. Integrating AI tools like SemiKong is crucial for a more efficient and resilient semiconductor industry.
Related:
- DeepSeek Prompt Formula: 5 Steps to Unlock AI Potential
- The Secrets Behind DeepSeek R1 Training Revealed
- How DeepSeek Enhances Electronics Engineering Processes?
- Top AI Model Containers and How to Use Them Effectively
Disclaimer:
- This channel does not make any representations or warranties regarding the availability, accuracy, timeliness, effectiveness, or completeness of any information posted. It hereby disclaims any liability or consequences arising from the use of the information.
- This channel is non-commercial and non-profit. The re-posted content does not signify endorsement of its views or responsibility for its authenticity. It does not intend to constitute any other guidance. This channel is not liable for any inaccuracies or errors in the re-posted or published information, directly or indirectly.
- Some data, materials, text, images, etc., used in this channel are sourced from the internet, and all reposts are duly credited to their sources. If you discover any work that infringes on your intellectual property rights or personal legal interests, please contact us, and we will promptly modify or remove it.
