NGK and Laboro.AI Succeed in Developing Large-Scale Simulation Software Using Generative AI
NGK and Laboro.AI have established a method to develop large-scale simulation software using generative AI. It reduces the developer's coding burden to one-third and will be applied to NGK's crystal form prediction software.
📋 Article Processing Timeline
- 📰 Published: April 10, 2026 at 00:30
- 🔍 Collected: April 11, 2026 at 00:30 (24h 0m after Published)
- 🤖 AI Analyzed: April 20, 2026 at 00:13 (215h 42m after Collected)
NGK Insulators, Ltd. (President: Shigeru Kobayashi, Head Office: Nagoya City, Aichi Prefecture, hereinafter NGK) and Laboro.AI Inc. (Representative Director CEO: Tetsuo Shiibashi, Representative Director COO & CTO: Hiromasa Fujiwara, Head Office: Chuo-ku, Tokyo, hereinafter Laboro.AI) have jointly established a development method for large-scale scientific and technological calculation simulation software utilizing generative AI. NGK will utilize this achievement in the development of crystal form prediction software for its 'Organic Compound Crystal Search Service'.
## About the Project
This project was implemented with the aim of advancing the development of simulation technologies crucial in the manufacturing industry more accurately and efficiently. It combines NGK's theoretical and experimental knowledge regarding infrared irradiation (*1) and crystal precipitation (*2) with Laboro.AI's know-how in AI technology implementation.
In the organic compound crystal search service provided by NGK, the goal is to create different crystal structures even from the same compound components by varying the precipitation conditions. Differences in crystal structure affect solubility, stability, and moldability, making them crucial considerations, especially in the pharmaceutical field. On the other hand, to obtain various crystal forms, it is necessary to consider combinations of numerous parameters such as temperature, concentration, and infrared irradiation conditions. Since exhaustive verification through experiments alone is difficult, narrowing down conditions using preliminary predictions via simulation is desired.
To address these challenges, NGK and Laboro.AI have been working on developing simulation software that expresses physicochemical laws related to infrared irradiation and crystal precipitation as mathematical formulas to perform large-scale numerical calculations. Traditionally, such software development has centered on manual implementation and verification, posing challenges like the risk of programming errors and prolonged development periods.
In this verification, we evaluated whether large-scale numerical calculation programs could be correctly generated by documenting formulas, conditions, and constraints in a format understandable by humans and inputting them into generative AI. As a result, we confirmed that by appropriately providing 'development rules to follow' and 'model base codes', generative AI can generate accurate programs without arbitrary simplifications or misinterpretations. Furthermore, we gained the prospect that a general-purpose coding agent (*3) can handle most of the program implementation, significantly reducing the human burden to about one-third of the conventional level. This allows developers to focus more on essential technical considerations, such as examining theories and formulas and interpreting results.
[Configuration of the simulation software developed and verified this time]
By applying the generative AI program generation method verified this time to the development of crystal form prediction software, NGK will quickly expand functions according to customer needs, and service products
## About the Project
This project was implemented with the aim of advancing the development of simulation technologies crucial in the manufacturing industry more accurately and efficiently. It combines NGK's theoretical and experimental knowledge regarding infrared irradiation (*1) and crystal precipitation (*2) with Laboro.AI's know-how in AI technology implementation.
In the organic compound crystal search service provided by NGK, the goal is to create different crystal structures even from the same compound components by varying the precipitation conditions. Differences in crystal structure affect solubility, stability, and moldability, making them crucial considerations, especially in the pharmaceutical field. On the other hand, to obtain various crystal forms, it is necessary to consider combinations of numerous parameters such as temperature, concentration, and infrared irradiation conditions. Since exhaustive verification through experiments alone is difficult, narrowing down conditions using preliminary predictions via simulation is desired.
To address these challenges, NGK and Laboro.AI have been working on developing simulation software that expresses physicochemical laws related to infrared irradiation and crystal precipitation as mathematical formulas to perform large-scale numerical calculations. Traditionally, such software development has centered on manual implementation and verification, posing challenges like the risk of programming errors and prolonged development periods.
In this verification, we evaluated whether large-scale numerical calculation programs could be correctly generated by documenting formulas, conditions, and constraints in a format understandable by humans and inputting them into generative AI. As a result, we confirmed that by appropriately providing 'development rules to follow' and 'model base codes', generative AI can generate accurate programs without arbitrary simplifications or misinterpretations. Furthermore, we gained the prospect that a general-purpose coding agent (*3) can handle most of the program implementation, significantly reducing the human burden to about one-third of the conventional level. This allows developers to focus more on essential technical considerations, such as examining theories and formulas and interpreting results.
[Configuration of the simulation software developed and verified this time]
By applying the generative AI program generation method verified this time to the development of crystal form prediction software, NGK will quickly expand functions according to customer needs, and service products