This paper successfully demonstrates a logiccompatible,high performance and high reliability,automotive-grade 2.5V embedded NVM processextending over several generations. A high-density flashmacro is used to debug process complexities which arisefrom the add-on modules. The modular approach isadopted for integrating self-aligned, floating-gate-basedsplit-gate SuperFlash® ESF3 cell into 40nm CMOS logicprocess. Key features of the product-like Macro aredual power supply with input voltage fluctuations, wideoperating temperature range from -40ºC to 150ºC, fastbyte/word program under 10μs and sector/chip eraseunder 10ms. The macro random read access time is only8ns under worst case conditions. Key process monitorsare characterization and yield of the Macro. Endurancewas extended to 200k cycles and satisfy automotivegrade requirement with wide read margin. Post-cyclingdata retention performs very well up to 150ºC. Wafersort yield is in high double digits, with consistent waferto-wafer and within-wafer uniformity, showing goodprocess control. The technology is suitable for highspeedautomotive MCU, as well as IoT, smart card, andindustrial MCU applications.