<p>The rapid transition to online higher education in Indonesia triggered a systemic shock to the learning process, characterized by a complex interplay of cognitive, technological, and pedagogical stressors. While Cognitive Load Theory (CLT), the Technology Acceptance Model (TAM), and Self-Regulated Learning (SRL) are often studied in isolation, their integrated impact on academic performance within a Community of Inquiry (CoI) remains underexplored. This mixed-methods study investigates this gap by analyzing longitudinal test scores and qualitative feedback from Academic English students at a private university in Surabaya across the 2019–2020 and 2020–2021 semesters. Quantitative results indicate a significant performance collapse during the initial shift (Cohen’s <i>d</i> = 1.66), followed by a marked recovery as the ecosystem matured (Cohen’s <i>d</i> = 0.88). Thematic analysis identifies a “negative cascade,” where technological friction triggered cognitive exhaustion and subsequent social detachment. Conversely, findings reveal that a robust Community of Inquiry acts as a “resilient hub,” where teaching and social presence provide the dynamic buffering necessary to stabilize cognitive load and foster student autonomy. These results offer a theoretically grounded roadmap for designing linguistically responsive and resilient digital environments, particularly for institutions navigating the complexities of post-pandemic digital maturation in resource-constrained settings.</p>