Hitachi Develops Platform Technology to Select Chiplet Semiconductor Package Designs

Hitachi, Ltd. has developed platform technology for chiplet semiconductor package design that automates the narrowing of candidates in the early design stages and uses fast physical analysis to make design selection more efficient. Hitachi said there is demand for semiconductor development tailored to applications including edge AI, autonomous driving and industrial IoT.

Hitachi said there is demand for semiconductor development tailored to applications including edge AI, autonomous driving and industrial IoT. The technology targets the early-stage selection of chiplet package designs for those applications. No figures were provided for improvements in analysis time or design efficiency, so how much the technology helps with selection in actual design workflows will be key to assessing its practical value.

The focus of the technology is not chiplet manufacturing or packaging equipment itself, but design support that narrows candidate options early in the package design process and makes selection more efficient through physical analysis. Hitachi's announcement describes an approach that combines automation with fast analysis, but the information provided does not establish comparisons with conventional methods, reductions in analysis time or the design conditions covered. The novelty that can be identified at this stage is therefore a function within the design process; it cannot be taken as evidence of improved performance or lower development costs. A technology covered by this newspaper, Hitachi develops technology for safe operation of physical AI, focused on technology supporting AI operations in social infrastructure. This announcement, by contrast, concerns the package design stage for developing application-specific semiconductors. The two announcements do not directly link the technologies, but operational technology for running AI in the field and design support for the semiconductors that process it can be understood as challenges at different levels. Hitachi's physical AI technology for equipment maintenance is also an example of a field application using operational data, but publicly available information does not confirm any technical connection to this announcement. Selecting a semiconductor configuration suited to its intended application is likely to be an important step in chiplet design. If a system for narrowing candidates early in the process proves effective, it could influence how designs are selected for detailed work in later stages. However, the information announced does not establish which companies' design workflows the technology is intended to enter, which processes it automates or how it connects with external design environments and manufacturing processes. Design-support technology should also not be equated with semiconductor supply capacity or product performance. The next points to check are the physical properties and input conditions analyzed, comparisons of time and accuracy against existing methods, the package configurations to which the technology can be applied, and whether it has been tested in actual projects. Such information would help assess how far candidate screening replaces designers' judgment or instead assists by suggesting options. For now, it is appropriate to describe the technology as a platform for streamlining the early design stages of application-specific semiconductors.

The technology targets package design for chiplet semiconductors, which combine multiple chips. Hitachi's announcement headline identifies automation of early-stage candidate screening and fast physical analysis as elements of the technology. The excerpt of the announcement provided does not specify the physical properties analyzed, the number of candidates, analysis time, selection accuracy, application examples or availability date.