Psychological safety, operationalized as "the collective perception of team members that the team is a safe place for interpersonal risk taking," constitutes one of the most well-documented predictors of team performance in organizational scholarship. However, despite a burgeoning body of work since the groundbreaking contribution of Edmondson, substantial voids exist in relation to the multi-national, multi-industry empirical investigation into the processes underlying psychological safety-performance relationships and the conditions that strongly facilitate or hinder the emergence of psychological safety in particular. The current project seeks to fill such gaps through a concurrent mixed-methods approach combining quantitative survey research with qualitative data collection involving 1,380 teams from 94 companies in six different countries, a ten-quarter longitudinal field study of 412 teams, and interviews with 44 team leaders and team members. Measured by the Edmondson Psychological Safety Scale together with established scales capturing innovation, task execution, error reporting, knowledge exchange, turnover, and customer satisfaction, the findings reveal that psychological safety is a significant independent predictor of all six performance indicators after controlling for team size, industry, tenure, and leader-member exchange style. The transformational form of leadership practice is found to be the most significant organizational facilitator of psychological safety, whereas blame culture and hierarchies of silence are considered as the key inhibitors of psychological safety. Mediation effect of psychological safety in the connection between transformational leadership practice and team innovation is established via structural equation modeling. In the current research, an architecture framework consisting of three tiers is suggested as an approach to institutionalize psychological safety, which involves Human-Centred Performance Architecture.