The source @Tech_Reve recently tweeted that all of Qualcomm’s flagship chips, including the Snapdragon 8 Gen 4 released this year, are completely manufactured using TSMC’s 3nm process. However, the Snapdragon 8 to be launched next year will Dragon 8 Gen 5 will adopt a multi-fab solution.
Reports suggest that Samsung produced the Qualcomm Snapdragon 8 Gen 1 chips, but due to overheating and inefficiency, Qualcomm switched to TSMC, which has since become its exclusive foundry partner.
This pattern will change next year, with regular Snapdragon 8 Gen 5 chips continuing to be manufactured by TSMC’s 3nm process N3E, while the Snapdragon 8 Gen 5 chips for Samsung’s flagship Galaxy S26 series use Samsung’s SF2P process.
Note: Compared with SF3, the SF2 process can improve power consumption efficiency by 25% at the same frequency and complexity, improve performance by 12% at the same power consumption and complexity, and improve performance by 12% at the same performance and complexity. Reduce the area by 5%.
To make the SF2 process more competitive, Samsung will also provide a series of advanced IP portfolio for the process, including LPDDR5x, HBM3P, PCIe Gen6 and 112G SerDes.
SF2P is based on the SF2 process and is optimized for high-performance computing (HPC). It is expected to have greater improvements in performance.
Previous news said that Qualcomm Snapdragon 8 Gen 5 will use the “Pegasus” core, a “2+6” cluster design and the Slice GPU architecture.
