Introduction: Cognitive deficits are typically seen in schizophrenia and in the prodrome, and are a major predictor of functional outcomes in patients. In Africa, few studies have investigated neurocognition in psychosis, which presents a gap in our understanding of the heterogeneity of the illness. In this study, we assessed neurocognition among the largest sample of psychosis-risk participants recruited in the continent to date. Methods: The study was conducted in Kenya, and involved 295 psychiatric medication-naïve participants at clinical high-risk (CHR) for psychosis and healthy controls, aged 15–25 yrs. Psychosis-risk status was determined separately using the Structured Interview of Psychosis-Risk Syndromes (i.e. CHR) and by self-report with the Washington Early Recognition Center Affectivity and Psychosis Screen. Eleven tests were administered using the University of Pennsylvania Computerized Neurocognitive Battery. Test performance across groups were investigated, as well as demographic and clinical effects. Results: Fewer participants were designated as being at psychosis-risk with structured interview (n = 47; CHR) than with self-report (n = 155). A MANOVA of cognitive test performance was significant only when groups were ascertained based on self-report (p = 0.03), with decreased performance in the risk group on verbal intelligence (p = 0.003; d = 0.39), emotion recognition (p = 0.003; d = 0.36), sensorimotor processing (p = 0.01; d = 0.31) and verbal memory (p = 0.035; d = 0.21). Only verbal intelligence was significantly worse in the CHR group compared to controls (p = 0.036; d = 0.45). There were no significant age and gender relationships. Conclusion: Deficits across multiple cognitive domains are present in Kenyan psychosis-risk youth, most significantly in verbal intelligence. The pattern of cognitive deficits and an absence of gender effects may represent ethnicity-specific phenotypes of the psychosis-risk state. Longitudinal studies of neurocognition in Kenyan patients who convert to psychosis may enhance risk prediction in this population, and facilitate targeted interventions.