Diplomats came back from the summer break on 1 September expecting to nudge Ukraine accession forward by one technical step, and Hungary stopped them inside an hour. At COELA, the Council working party that handles enlargement, Budapest again refused to endorse the screening results for clusters two and three, as Ukrainian outlets reported. The identical attempt collapsed on 22 July. Nothing in the substance changed over the summer. Only the calendar moved.
Screening approval reads like paperwork, and it is. It is also the gate. Member states must first accept the Commission’s assessment of how closely a candidate’s laws already track EU law. Only then can an intergovernmental conference open the cluster and start real negotiation on the chapters inside it. Every one of those steps needs unanimity, so a single capital keeps the gate shut simply by declining to speak.
Hungary did offer a route, and most governments like it less than the blockage itself. Budapest signalled it would clear cluster three for Moldova alone, splitting two candidates the Union has deliberately moved in tandem since 2022. Ireland, which holds the Council presidency, leads the majority refusing that split. Accounts of the meeting make the reasoning plain. Decoupling would hand Budapest a permanent lever over Kyiv while letting the Union advertise progress somewhere on the map.
Officials in Kyiv describe Hungary’s asking price as protection for the Hungarian minority in Zakarpattia, a demand that belongs to the fundamentals cluster rather than to either of the two files now frozen. Ukraine has legislated twice since 2023 on minority language and education under Commission supervision. The distance between the stated grievance and the blocked paperwork tells its own story about how the veto actually functions.
The pattern repeats. Hungary held clusters two through six at technical level for the better part of a year, then released cluster six, and negotiators opened External Relations for Ukraine on 14 July. That opening proved the machinery still works whenever Budapest allows it. It also showed how little the machinery achieves when Budapest does not, because the concession arrived after months of delay and cost the Union leverage it could have spent elsewhere.
Set that against the Western Balkans, where unanimity currently harms nobody. Montenegro expects to close its remaining chapters this year and aims at membership in 2028. Albania has entered the concluding phase of its own talks and targets the end of 2027. Both advanced because no member state chose to stand in the way. Enlargement policy works perfectly well. It stops working the moment one government decides otherwise, and the treaties supply no remedy.
Two dates matter next. The Commission publishes its annual enlargement package in the autumn, which will grade Ukraine and Moldova on reforms neither can convert into opened clusters. Leaders then meet in December, where several capitals want a serious discussion about shifting intermediate procedural steps to qualified majority voting. That change would itself require unanimity, which is the circular trap sitting at the centre of the file.
Nothing forces a resolution before then. The COELA meeting on 2 September left Ukraine and Moldova off the agenda entirely, so the preparations that resumed on 1 September now have no obvious landing point. Kyiv keeps filing reform legislation into a process that cannot register it, and Chisinau waits for a decision about its own future that turns on a dispute it has no part in. For a Union that has spent three years calling enlargement its central geopolitical project, this is an expensive stalemate to keep paying for.
