If you are an avid reader of SammyGuru, you might be aware of how Samsung is working hard to bounce back in the semiconductor industry. The Korean firm has adopted several strategies to improve the yield of its cutting-edge manufacturing process. Now, it has reportedly adopted Meta’s latest large language model, Llama 4, across the entire production line. This should help the company streamline operations and identify problems more efficiently.
Samsung DS Division implements Meta’s Llama 4 models
Earlier this month, Meta introduced two models in the Llama 4 series: Llama 4 Scout and Llama 4 Maverick, which understand text, images, voices, and videos simultaneously. According to a report from Korean outlet Sedaily, Samsung’s DS Division (in charge of semiconductors) has recently integrated Meta’s Llama 4 into its employee assistance programs this month. The report further says it has adopted both the basic model, Maverick, and the lightweight model, Scout.
This will now allow employees to use Llama 4 for several tasks within the semiconductor production process, from simple document handling to semiconductor design and manufacturing. So far, Samsung has been mainly using its own internal LLM model on its production line due to the risk of data leakage of the semiconductor design and manufacturing process.
However, its own LLM model is not perfect, leading to criticisms regarding the slow progress in performance. It looks like the division has adopted external generative AI to boost its competitiveness in semiconductor design and manufacturing. Since Llama 4 models do not rely on an external cloud and are not connected to external networks, they are safe from the threat of hacking.
The report also says that, besides Llama 4, Samsung is planning to use generative AI models in its semiconductor work process from several big tech companies. “We will continue to support high-performance open source AI models to improve the convenience of executives and employees,” a Samsung official told the publication.
An earlier report suggested that Samsung implemented Palantir’s AI-powered data analytics, which can analyze large amounts of data using AI. This will allow the company to improve yield in the semiconductor manufacturing process. So, it looks like the Korean firm wants to leverage AI to improve various aspects of its semiconductor operations.
