On 9 June 2026, a coalition of 37 organisations — spanning consumer group BEUC, digital-rights network EDRi, the Internet Society, commercial broadcasters' association ACT, the Motion Picture Association and Videogames Europe — published a joint statement urging the European Parliament and Council to make two changes to the proposed Digital Networks Act (DNA): delete draft Articles 191–193 on IP interconnection, and keep the Open Internet Regulation (OIR) as a standalone legal instrument rather than folding it into the new framework. It is a rare alignment: consumer advocates, civil society, Hollywood studios and game publishers do not usually sign the same letter. When they do, co-legislators should ask what is at stake.
The answer matters acutely for France, where regulator ARCEP has spent ten years building one of Europe's most rigorous open-internet enforcement practices on top of the OIR's text — and where that practice now rests on provisions the DNA proposes to recast.
What the recast actually deletes
The Commission adopted the DNA proposal on 21 January 2026, merging the European Electronic Communications Code, the BEREC Regulation, the Radio Spectrum Policy Programme and core parts of the OIR into a single regulation. On net neutrality, the Commission insists the principles are maintained, with a new mechanism to "clarify Open Internet rules for innovative services."
The coalition's objection is not to clarification but to demolition of the interpretative scaffolding. As EDRi documents, the recast deletes 18 of the OIR's 19 recitals — the interpretative material that has anchored a decade of BEREC guidelines and Court of Justice case law, from the Telenor zero-rating rulings (2020) onward. Analysis by epicenter.works, reported by Tech Policy Press, finds the DNA repeals the OIR's operative net-neutrality articles and only partially reintegrates them. Every deleted recital is an invitation to relitigate settled questions: what counts as reasonable traffic management, when zero-rating discriminates, how specialised services may be offered. Legal certainty is not a recital's decoration; it is the product.
The steelman: Brussels is not wrong about the problem
The Commission's underlying diagnosis deserves a fair hearing. Europe's telecoms market is fragmented across 27 national regimes, capital expenditure lags fibre and 5G targets, and the Commission projects the DNA could add roughly €400 billion to EU GDP by 2035 through a genuine single market for connectivity. Consolidating four instruments into one directly applicable regulation is defensible simplification. And operators' complaint that a handful of content giants generate most network traffic is not invented: ARCEP's own data-interconnection barometer found that around 53% of inbound traffic to France's main ISPs comes from just five companies — Netflix, Akamai, Meta, Google and Amazon. A legislator could reasonably conclude the interconnection ecosystem deserves a forum.
Articles 191–193: a remedy without a diagnosis
But a forum is precisely what Articles 191–193 are not. They establish an "ecosystem cooperation" mechanism and a "voluntary conciliation" procedure for IP-interconnection disputes — voluntary in name, yet structured so that a content or cloud provider that declines to negotiate payment terms with an ISP can be pulled into a formalised process. The joint statement warns this risks "institutionalising private commercial negotiations as a formal arbitration process," creating network fees by the back door after the Commission's explicit fair-share levy was abandoned.
The evidentiary record points the other way. BEREC's repeated assessments — cited by the coalition — find the IP-interconnection market functioning competitively with no indication of market failure; EDRi notes that more than 99% of peering agreements are settlement-free, concluded without money changing hands. Interconnection is the internet's quiet success story: a market so frictionless it has never needed a regulator's waiting room. Proportionate regulation intervenes where markets fail, not where incumbents wish their counterparties had less leverage. Telecoms operators, through Connect Europe, are already demanding the mechanism be hardened into mandatory arbitration — confirming the coalition's fear that Articles 191–193 are a foothold, not a ceiling.
The view from Paris
France illustrates what is being put at risk. ARCEP enforces the OIR through annual State of the Internet reports, a data-interconnection barometer, and continuous supervision of traffic management. Its 2025 report, marking the OIR's tenth anniversary, calls the regulation "an effective and resilient instrument that has stood the test of time" — and demonstrates its adaptability, concluding after joint work with BEREC that 5G network slicing can be deployed compatibly with the open-internet rules. That finding undercuts the DNA's central justification: the existing framework already accommodates innovative services without rewriting the statute. A regime that polices zero-rating, absorbs network slicing and keeps 99% of interconnection settlement-free is not a regime in need of rescue.
For ARCEP, the recast means a decade of doctrine — built on recitals that will no longer exist — becomes contestable, while conciliation under Articles 191–193 would sit awkwardly beside its evidence-based barometer approach, which has managed interconnection disputes (including the famous Free–Google congestion episodes that prompted the barometer's creation) through transparency rather than tribunals.
The proportionate fix
Parliament's ITRE committee, under rapporteur Michał Kobosko, and the Council working party that has been examining the text since March 2026 have a clean option: take the coalition's advice. Delete Articles 191–193 until someone produces evidence of a market failure they would cure. Keep the OIR — recitals intact — as a separate instrument, or transpose it verbatim. The DNA's genuinely pro-investment elements, from the single passport to spectrum reform, lose nothing. Europe's open internet was not an accident; it was a statute. Co-legislators should not unsettle it to solve a problem no regulator can find.