This article aims to use a corpus-based analysis to examine loan translation (LT) as an Arabicization technique in Written Standard Arabic (WSA) in the field of information technology (IT), and to investigate the factors that impact the formation of LTs. The data were collected from four distinct types of corpora: four Arabic online newspapers, the Microsoft Language Portal, two user manuals, and the Arabic Corpus of King Abdul-Aziz City for Science and Technology (KACST). Four LT strategies were identified: morphosyntactic pattern alteration, literal translation, loan blending, and metaphorical extension. Literal translation occurred in three degrees of calquing: full LTs, partial LTs, and non-LTs. Metaphors based on similarity in function play an important role in facilitating the understanding and acceptance of IT terms in WSA. The study also found that paradigmatic variation prevails in LTs, with synonymous variations being more common than nonsynonymous ones, and a total absence of syntagmatic variation. The KACST analysis showed that these variants differ a lot in terms of frequency of occurrence. The creation and variation of LTs are influenced by linguistic factors such as lexical need, semantic transparency, and lexical borrowing, as well as extralinguistic factors such as nationalism and the lack of coordination among Arabic language academies. The study concludes with some implications for overcoming the absence of coordination between these academies and for improving the quality of IT-related LTs between English and WSA.