Additive manufacturing (AM) of alloys has attracted much attention in recent years for making geometrically complex engineering parts owing to its unique benefits, such as high flexibility and low waste. The in-service performance of AM parts is dependent on the microstructures and grain boundary networks formed during AM, which are often significantly different from their wrought counterparts. Characteristics such as grain size and morphology, texture, and the detailed grain boundary network are known to control various mechanical and corrosion properties. Advanced understanding on how AM parameters affect the formation of these microstructural characteristics is hence critical for optimising processing parameters to unlock superior properties. In this study, the difficult-to-weld nickel-based superalloy Inconel 738 was fabricated via electron-beam powder bed fusion (EPBF) following linear and random scanning strategies. Random scanning resulted in finer, less elongated, and crystallographically more random grains compared to the linear strategy. However, both scanning strategies achieve unique high grain structure stability up to 1250
℃
due to the presence of carbides pinning the grain boundaries. Despite significant difference in texture and morphology, majority of grains terminated on {100} habit planes in both linear and random built samples. The results show potential for controlling grain boundary networks during EPBF by tuning scan strategies.