The debate surrounding the proposed Digital Networks Act (“DNA”) has revived a familiar concern. A number of stakeholders – in particular over-the-top (“OTT”), content provider and Internet activists, argue that Articles 191 and 192 may introduce, through the back door, a form of “fair share”: namely, a route through which network operators could obtain economic compensation from content and application providers (“CAPs”), cloud providers, content delivery networks (“CDNs”) or other actors in the wider digital ecosystem.

That concern is understandable in light of the policy controversy (mainly lead by former EU commissioner Thierry Breton) that has surrounded network-contribution proposals in Europe. It is nevertheless misplaced as a matter of legal interpretation. It rests on an insufficiently precise reading of the architecture, wording and function of Articles 191 and 192 – also with respect to DNA’s interconnection rules, which must be taken into account to understand the entire picture. 

The DNA certainly acknowledges the growing interdependence between networks, cloud, edge computing, CDNs and online services. It also recognises that effective cooperation can improve traffic management, resilience, environmental performance and end-to-end service quality. Yet recognition of an expanded connectivity ecosystem does not amount to the creation of a general right to payment for traffic, nor to a regulatory mechanism for reallocating the costs of network investment among different business models.

Articles 191 and 192 DNA must instead be read for what they are: a narrowly delimited framework for voluntary cooperation and non-binding conciliation in relation to agreements and activities that fall outside the regime governing the interconnection of public electronic communications networks.

The decisive textual limit

The crucial starting point is the final part of Article 191(1). Its wording confines the provision to agreements and activities other than those relating to interconnection between public electronic communications networks:

“Such guidelines shall cover matters outside of the scope of obligations under this Regulation which have an effect on the provision of electronic communications services or information society services”, 

as clarified also by recital 404: 

“Matters addressed in those guidelines, including issues relating to interconnection between publicly available electronic communications networks and networks owned by undertakings other than publicly available electronic communications networks, should not be subject to dispute resolution mechanism, which should be limited to the enforcement of obligations laid down in this Regulation and which should not extend to arrangements of a cooperative or voluntary nature”.

This is not an incidental clarification. It is the provision’s central jurisdictional boundary.

The distinction matters because public-network interconnection is already governed by a separate body of law within the DNA. That body of rules concerns the legal rights and obligations of undertakings operating public electronic communications networks, including their ability – and, in appropriate circumstances, their duty – to negotiate interconnection in order to secure end-to-end connectivity and interoperability.

Accordingly, Article 191 cannot sensibly be read as a supplementary, alternative or overlapping dispute-resolution route for public-network interconnection. The provision expressly excludes that very subject matter.

This point should put to rest the proposition that Article 191 may covertly enable a fair-share remedy. A payment obligation resembling a network-contribution scheme would require a legal basis sufficiently clear to determine at least:

• The categories of undertakings subject to the obligation.
• The economic trigger giving rise to payment.
• The beneficiary or beneficiaries of the payment.
• The calculation methodology.
• The relationship between traffic, costs, investment and remuneration.
• The role and powers of the competent authority.
• The procedural and substantive safeguards available to the parties.

None of this is contained in Article 191. More fundamentally, its express carve-out prevents it from becoming a vehicle for regulatory disputes about public-network interconnection – the area in which a compulsory access or compensation claim might otherwise be imagined.

Two distinct and non-communicating silos

The DNA should be understood as establishing two legally distinct silos.

The distinction is not merely semantic. It follows from the underlying regulatory logic.

Public electronic communications networks occupy a particular position in sectoral regulation because they provide the infrastructure through which users can communicate across networks. Their interconnection is linked to the public interest in universal, end-to-end connectivity and interoperable communications. That is why the legal framework may recognise reciprocal rights and duties to negotiate, and why an NRA may, where the conditions are met, intervene to resolve disputes or impose proportionate obligations.

Article 191 operates in a different legal space. It is designed to encourage cooperation among actors whose technical and commercial relationships increasingly affect the performance of the broader connectivity ecosystem, but whose arrangements do not thereby become public-network interconnection subject to the associated regulatory machinery. In that respect, Article 191 reflects the DNA’s broader policy ambition to facilitate innovation, cooperation and sustainable practices without automatically extending the full apparatus of telecoms access regulation to every cloud, CDN, platform or infrastructure relationship.

Article 192 has the same scope

Article 192 about “voluntary conciliation” must be interpreted together with Article 191. It does not create an autonomous and broader jurisdiction over interconnection disputes. It provides a conciliation facility for the ecosystem-cooperation matters covered by Article 191.

Its scope is therefore derivative: what Article 191 excludes, Article 192 cannot reintroduce.

This is legally important for two reasons.

First, the conciliation route is voluntary. It is not a compulsory regulatory procedure through which one party may force another to litigate an alleged entitlement to payment.

Second, the outcome is non-binding. Conciliation may help parties identify technical, operational or commercial solutions: it does not produce an enforceable regulatory determination equivalent to an NRA decision in an access or interconnection dispute.

A voluntary, non-binding procedure cannot plausibly be transformed into a fair-share mechanism simply because it brings together operators with different commercial interests. A fair-share system would require the opposite: a mandatory substantive obligation, enforceable against specifically defined entities, based on objective criteria and subject to an authority capable of issuing binding decisions.

Article 192 provides none of those elements. It is a facilitation mechanism, not a transfer-of-value mechanism.

The proposition that Articles 191 and 192 might overlap with public-network interconnection disputes must therefore be rejected under the current draft. The relevant provisions are not two doors leading into the same room. They are separate rooms, serving different regulatory purposes, with no legal corridor between them.

Are these provisions needed?

A further question is whether Articles 191 and 192 are necessary or useful in the present market context. In my opinion, they firstly should be seen as the residual trace of a political debate that ultimately did not produce a fair-share mechanism in the DNA. In that sense, they reflect a compromise: the DNA proposal does not establish a legal obligation for CAPs or other digital actors to finance telecoms networks, but it preserves a framework in which the issues underlying that debate – traffic growth, investment needs, cost allocation, technical cooperation and the functioning of the wider connectivity ecosystem – can be examined in a structured manner. 

Article 191 should therefore not be dismissed as wholly superfluous. Although it may not have been indispensable, it can promote greater transparency and, under BEREC’s aegis, support a more evidence-based and relatively objective understanding of the technical and commercial questions that have fuelled the fair-share debate. OTTs, cloud providers and CAPs should not regard that exercise as a threat: properly applied, it may help distinguish verifiable interconnection-related facts from broad claims for business-model compensation. 

By contrast, article 192 is more difficult to defend. Creating a formal conciliation procedure – yet one that is voluntary in outcome and non-binding – around such a politically charged subject risks generating strategic disputes rather than practical solutions. It may invite parties to use the procedure as a platform for advancing objectives that the legislature deliberately chose not to embed in substantive law. For that reason, while Article 191 may have a modest but defensible transparency function, Article 192 appears of limited practical utility and may reasonably be considered a candidate for deletion during the legislative process. 

The cloud, CDN and OTT question

A residual interpretative issue may arise in relation to the legal classification of cloud, CDN or OTT infrastructures. Could such infrastructures be characterised as “public electronic communications networks” and thereby fall within the public-network interconnection regime rather than the ecosystem-cooperation framework?

The question exists because the concept of a public network may be read narrowly or broadly.

Under a narrower approach, a public electronic communications network is one whose central purpose is to provide connectivity directly to end-users or to make public electronic communications services available to them. Under a broader functional approach, one could argue that an infrastructure should be considered public where it carries traffic that eventually reaches end-users through a chain of subsequent interconnections.

The broader approach should not prevail merely because an infrastructure participates in the technical delivery of Internet traffic. The fact that a cloud platform, CDN or OTT provider operates servers, caches, backbone capacity, private links or content-delivery nodes does not, in itself, mean that its infrastructure is a public electronic communications network.

The DNA’s recitals, as well as the new legal definition for interconnection set by art. 2(29) DNA (omitting the term “public”), expressly acknowledge that interconnection may occur between public networks and private networks, including networks owned or operated by CAPs and content-delivery infrastructures. They also recognise that caches, content-delivery nodes and related techniques can improve traffic exchange and end-to-end performance. That acknowledgment is highly significant: it confirms that technical interconnection does not erase the distinction between public and private networks. 

The relevant question is not whether an infrastructure handles or facilitates connectivity. Virtually every important element of the modern Internet does so. The question is whether the infrastructure’s principal purpose is the provision of public electronic communications or digital services falling within the relevant regulatory perimeter.

In general, the primary function of a cloud infrastructure, CDN or OTT infrastructure is not to provide electronic communications services to the public – whether access services or public transport services – but to enable the provision of Internet-based, hosting, computing, storage, application, content-distribution or information-society services. The Commission’s explanatory memorandum expressly states that the DNA does not seek to regulate cloud services, notwithstanding the increasing convergence between electronic communications networks and cloud or edge computing. 

A provider may potentially fall within a general-authorisation regime where its specific activities meet the applicable statutory conditions. But general authorisation and classification as a public electronic communications network are not identical questions. The former does not automatically entail the latter; nor does it transform a cloud, CDN or OTT business model into a public-network operator for the purposes of mandatory interconnection rights and duties.

For reasons of legal certainty, however, the co-legislators could usefully add a recital clarifying that cloud, CDN and OTT infrastructures do not qualify as public electronic communications networks solely because they carry, cache, optimise or transmit traffic that ultimately reaches end-users through public networks. Such clarification would not change the system. It would merely preserve its intended boundaries.

Regulatory interconnection is not business-model compensation: the Meta / DT case, and others

The fair-share debate also risks misunderstanding the purpose of regulatory interconnection itself.

Where public-network interconnection is subject to regulation, the regulatory objective is to facilitate connectivity and interoperability. It is not to compensate one undertaking for the totality of its historical, current or prospective investment in its own business model.

An NRA’s role in an interconnection dispute is ordinarily to ensure that interconnection can take place on fair, reasonable, transparent, non-discriminatory and, where legally required, cost-oriented terms. This may entail recognition of efficiently incurred costs directly associated with the interconnection arrangement. It does not entail a general power to redistribute the costs of deploying networks, operating platforms, financing content, building data centres or sustaining distinct commercial models.

Swiss practice offered a useful illustration. Under the Swiss telecommunications framework, parties are expected first to negotiate interconnection. If they fail to reach agreement, ComCom may decide on interconnection terms and prices, applying a cost-oriented approach based on the relevant legal framework. In the Swisscom/Init7 peering proceedings, ComCom’s 2024 decision required zero-settlement peering, on the basis that the chargeable interconnection costs consisted of the relevant router-port and cable costs, which were effectively incurred by both parties. 

The lesson is straightforward. Regulatory intervention is directed at enabling interconnection and recovering the costs genuinely attributable to it. It is not a mechanism by which a network operator may invoice an online service or CAP for a share of the operator’s network investment simply because the service generates significant traffic.

The dispute between Meta and Deutsche Telekom is frequently invoked in policy discussions as though it demonstrated the emergence of a regulatory fair-share principle. That characterisation is incorrect.

The litigation concerned a commercial IP-transport arrangement between the parties. The Cologne Regional Court upheld Deutsche Telekom’s civil claim because the relevant contract had not been effectively terminated and Meta had continued to use the services. The court therefore found that payment was due under the contractual relationship, including on the basis of an implied contractual arrangement arising from continued use. 

The case was not a regulatory interconnection dispute in which a telecoms authority imposed a statutory obligation on a CAP to contribute to network-investment costs. Nor did it establish a general legal rule according to which high-traffic online services must pay access-network operators for delivering their content to users.

The distinction is not technical legalism: it is dispositive. Contractual remuneration for dedicated capacity, IP transport or bespoke interconnection arrangements is a normal feature of commercial dealings. It does not become fair share merely because one counterparty is a large platform, or because traffic volumes are high.

Mandatory conciliation would not solve this

Some proposed amendments, including amendments associated with the European Parliament’s IMCO debate, reportedly seek to make the Article 192 conciliation procedure mandatory. Even if such a change were adopted, it would not create a legally workable fair-share system.

It would simply create an awkward procedural hybrid: parties could be compelled to participate in a dispute-resolution process concerning matters that Article 191 deliberately places outside public-network interconnection, while the resulting conciliation outcome would remain non-binding.

Mandatory participation in non-binding conciliation is not the same as a mandatory payment obligation. It is also not a substitute for a coherent regulatory regime capable of defining a contribution base, identifying payers and recipients, establishing a calculation methodology, ensuring proportionality, avoiding discrimination, and protecting the integrity of the open Internet and the internal market.

If the policy objective were genuinely to introduce a fair-share model, a few strategically inserted words would not suffice. The co-legislators would need to redesign the system at a much more fundamental level. They would need to legislate expressly for a new category of regulated relationship, establish the conditions under which a contribution might be required, determine how it would be calculated, identify a competent authority, specify the available remedies and reconcile the mechanism with competition law, net-neutrality rules, the freedom to conduct a business and the existing framework for public-network interconnection.

That would not be an amendment to Articles 191 and 192. It would be a different regulatory project.

Conclusion

The present DNA draft does not contain hidden fair share. Articles 191 and 192 do not regulate public-network interconnection, do not provide a route for binding regulatory awards, and do not establish any substantive basis for compelling CAPs, cloud providers, CDNs or OTT services to finance telecoms-network investment.

Their purpose is narrower and more constructive: to facilitate voluntary ecosystem cooperation and non-binding conciliation in an environment where technical interdependence is growing but not every relationship should be absorbed into the traditional framework of telecoms access and interconnection regulation.

The real legal challenge is therefore not to discover a fair-share mechanism where the text does not provide one. It is to preserve the conceptual distinction on which the DNA is built: public-network interconnection remains subject to its own carefully structured regulatory regime, while the wider ecosystem is encouraged – rather than compelled – to cooperate.

That distinction is sound. It is also essential.