Samsung's HBM4 Production Expansion Targets Next-Year Computational Access

Samsung plans to more than double HBM4 output next year, driving cheaper AI hardware for knowledge services

Samsung will more than double its HBM4 and HBM4E DRAM output next year compared to current levels. This expansion will see glass carrier volume rise 2.5-fold from 20,000 to 50,000 sheets monthly, while wafer production grows nearly 40% to 250,000 units monthly. Sixth-generation HBM4 (1c) and seventh-generation HBM4E chips will form the majority of next year's output, constituting 80% of Samsung's total HBM production.

The scale increase hinges on glass carrier volume growth—where each sheet supports multiple wafers—projected to translate to at least twofold growth in HBM4 and HBM4E output. This expansion directly enables more affordable AI hardware, expanding computational access for knowledge services and technical applications.

For abundance, cheaper AI infrastructure means more people can access advanced computing without prohibitive costs. This supports the growing need for computational resources in education, healthcare, and small-scale AI tools that drive practical knowledge applications.

What to watch: Glass carrier volume growth depends on process reuse and yield variations. The wafer production projection uses average monthly input metrics, not final output. Samsung's HBM4 output expansion is expected to be fully operational by next year but remains dependent on semiconductor industry yield stability.

Source: Hacker News