A considerable body of work has invoked learning in seeking to explain observed patterns of EU policy change. This paper scrutinizes the relevance of learning for understanding policy outcomes. We apply a consolidated framework based on factual, experiential and constructivist learning across the individual and organizational levels to examine the unlikely policy outcome of dedicating 20% of the EU 2014-2020 budget to climate action. Learning did play some role in the policy outcome, in that the belief that climate policy integration (CPI) was an appropriate instrument to address climate change was the result of constructivist learning over the preceding decade. However, this learning was restricted to a handful of policy entrepreneurs in the Cabinet/DG Clima, who largely "pushed" the policy through based on preexisting convictions. Conversely, beyond some experiential learning, there is little evidence that learning was a significant feature of the policy process amongst actors in other European institutions.