Social media platforms have altered how social interactions take place online. This new era of user practices, micro-communication cultures, bots, and an increasing algorithmic shaping of sociability, opens up new research endeavours to understand how systemic racism articulates on social media platforms. Research points to the need of studying race, racism and other forms of systemic oppression as the result of user practices and technological mediation (Brook, 2009; Daniels, 2013; Massanari, 2015; McIlwain, 2016; Nakamura & Chow-White, 2012; Noble & Tynes, 2016; Sharma, 2013). However, access to data is becoming gradually scarce. This article unravels the methodological challenges involved in studying platform-articulated racism in a context of platform shutdowns of their Application Programming Interfaces (APIs), and in a moment when opaque apps like WhatsApp and WeChat are used as social platforms. Platforms, even though they are private entities, resemble public institutions in that they play a fundamental role in organising public discourse and people’s lives. Although platforms often allege that users have the possibility to opt out, the way social media is entangled in our everyday lives makes the prospect to leave the service only an option for a privileged few. Thus, platforms enactment and reproduction of racism is a matter of public concern rather than a market problem to be solved. Racism, therefore, is built into spaces (social media platforms) that go beyond our more traditional institutions (for example, the state, the school, the media). We argue that the obstacles facing empirical work on social media contribute to the reproduction and enactment of systemic racism. This article departs from an analysis of empirical studies of platform-articulated racism from 2013 to 2018 that have used social media data. Findings shows that this research face a range of interconnected and complex challenges. This includes: epistemological challenges (due to techniques of concealing, covert propaganda, cloaking, lack of authorship, etc.), lack of contextual knowledge (i.e. how can we understand what we observe on one social media platform without a larger context), lack of access to data (API limitations and opaque apps), and ethical challenges. Building on the presented findings, the article discusses overall limitations of the field, possibilities of overcoming these as well as future problems posed by increasing opacity and social media companies’ questionable arrangements to collaborate and support research.