<p>The COVID-19 pandemic was a dynamic and often confusing period for clinical and biomedical research. As severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spread globally, knowledge accumulated rapidly through publications that were frequently based on preliminary or sometimes conflicting evidence, yet these papers played a critical role in shaping evolving medical, research and societal responses. Early in the pandemic, diabetes emerged as one of the strongest predictors of severe COVID-19 outcomes and mortality, placing it at the centre of early risk-stratification and therapeutic frameworks and prompting urgent efforts to understand the biological basis of these associations. As the pandemic progressed, reports of new-onset diabetes following COVID-19 infection raised the possibility of a bidirectional relationship between SARS-CoV-2 infection and diabetes. In this review, we provide a post-pandemic reappraisal of the clinical and experimental literature examining the intersection of COVID-19 and diabetes. We summarise proposed pathophysiological mechanisms, including the effects of SARS-CoV-2 infection in the pancreas and on peripheral insulin-sensitive tissues. We review key meta-analyses assessing the association between COVID-19 and incident type 1 and type 2 diabetes and highlight strengths and weaknesses of the epidemiologic studies underpinning these findings. We highlight the highest-quality evidence from prospective cohorts, as well as relevant clinical trials and registry-based studies that emerged from this collective experience. We discuss emerging relationships between long COVID and diabetes and the effect of vaccination on diabetes risk following SARS-CoV-2 infection. Finally, we identify critical knowledge gaps and outline priorities for ongoing and future studies needed to resolve remaining uncertainties.</p> Graphical Abstract <p></p>