The purpose of this research was to study the spreading of political hate speech by the far right through Twitter. A mixed methodology was employed, combining both quantitative and qualitative tools, within the framework of digital ethnography. Five characteristic cases of campaigns linked to political hate speech were chosen, meeting the four criteria set: Latin American scope with representativeness in terms of breadth and impact, political motivation, more than 100,000 tweets, and massive use of fake accounts. The analysis was performed with T-Hoarder, Gephi and MAXQDA. The conclusions drawn are that these campaigns do not occur spontaneously. Rather, a destabilizing political intention lies behind them, sponsored by organizations with considerable ability to disseminate messages and with extensive funds. The massive presence of false accounts, the repetition of certain spelling errors in identical form and the striking increase in the number of accounts just before campaigns are evidence of the automation of these processes. The constant use of aggressive and disparaging terms associated with hatred triggers extreme polarization and a climate of tension, threatening the building and consolidation of democracy itself. Apart from punitive measures, there is a need to implement educational proposals.