I have great respect for Benoit Felten, the author of the recent study on the Digital Networks Act (DNA) for CCIAA and its alleged “clear path” to network fees. The report makes a number of important points about legal uncertainty and regulatory scope. That said, there is a core legal argument I do not share, and I think it is worth unpacking it carefully.

The study claims that the extension of the general authorisation regime in Article 9 DNA to a broader range of actors – including CDNs and content/application providers (CAPs) – would in itself expose them to increased regulatory burden and, in particular, to dispute resolution mechanisms amounting to network fees. In my view, this is not legally correct. While the point about increased regulatory burden is grounded (although it is not clear which exact obligations would derive from the general authorisation regime, and why), the argument about the risk of “network fees” has no legal background.

A premise is important: under both the EECC and the DNA, only public networks are subject to a dispute settlement regime in case of controversy. This is because only public networks are subject to obligations to interconnect, while other networks (such as private networks) are not. It follows that CDNs and CAPs may fear a payment obligation in an interconnection dispute (the so-called network fee) only if they are formally qualified as public networks.

This said, it is worth noting that under both EECC and DNA, general authorisation is a regulatory status or condition for operating, not the criterion that determines whether a given actor is to be regarded as a public electronic communications network. To put it simply: being subject to general authorisation does not magically turn you into a “public network”.

So far, CDNs, cloud providers or CAPs have never been qualified as public networks in the way Proximus or Deutsche Telekom have. Fact is, the legal literature only concern cases between telcos operators, which were de plano public networks. The only relevant diverging case was the 2012 Cogent/Orange dispute, a competition case between a carrier and an incumbent (a dispute on termination settled with French antitrust rules, thus not an IP interconnection regulatory controversy).

And therefore, what is the real problem here? I believe that the narrative about network fee, in connection with art. 9 DNA, is purely instrumental. What CDNs, cloud providers and CAPs want to avoid is not, in abstract, any form of authorisation, but rather a telecom‑style regulatory jurisdiction being extended to them, whether or not they are formally treated as public networks. The jurisdiction by the local telecom regulator may involve various consequences (the above mentioned regulatory burden, in respect to which Benoit is right): registration duties, reporting obligations, transparency, more consumer protection, to say a few. Most CDNs and CAPs prefer to avoid such a scenario and are actively looking for ways to escape it. Framing the issue primarily as a “network fee” risk, and attributing it to Article 9 DNA as such, serves to this point.

I deliberately do not take a position here on whether the broader material scope of Article 9 is normatively good or bad policy, that is a separate debate. My point is narrower and strictly legal: you cannot derive “network fees” from the authorisation regime alone. There must be an additional step – the qualification of CAPs or CDNs as public networks – and that step is logically and normatively distinct from Article 9’s extension.

For this reason, I find the report’s causal chain problematic: “Article 9 extends authorisation → CAPs/CDNs are subject to dispute resolution → network fees follow.” In legal terms, Article 9 is at most instrumental: it sets out who must notify and comply with a series of horizontal obligations. Whether CAPs, CDNs or cloud providers are actually subject to IP interconnection dispute settlement mechanisms depends on their material classification within the DNA architecture, not on the mere fact that they are authorised.

Because of the above, a more valuable discussion would be about the conditions under which CDN, cloud or CAP infrastructures can be regarded as ‘public networks’ under both the EECC and the DNA. Prima facie, I do not see the risk, because such infrastructures do not themselves provide connectivity to the public; however, to be constructive, I recognise that the EU legislator should treat this point seriously and solve potential legal uncertainty.

The same caution applies to the discussion of the DNA’s conciliation mechanism in Articles 191–193. I acknowledge that, if the current voluntary mechanism were transformed into a mandatory one by future legislative changes, there could be legitimate concerns about how it might be used. But even in that scenario, the mere existence of a conciliation procedure does not equate to a network fee regime. The material scope of the mechanism, and the conditions under which it can be invoked, remain decisive.

In particular, the notion that “conciliation between providers of electronic communications networks and other undertakings” automatically implies payment obligations from CAPs/CDNs to telecom operators is, again, a non sequitur. It would require a further normative move: either an explicit pricing obligation, or an interpretation that stretches the purpose of conciliation beyond what the relevant recitals and safeguards (including Recital 404) currently allow. That is not present in the current text, and attributing this outcome directly to Articles 191–193 risks conflating potential future policy choices with the actual legal design of the DNA as proposed.

To be clear: I do not deny that the DNA raises real questions about scope, legal certainty and the future relationship between telecom regulation and the broader digital ecosystem. Nor do I underestimate how easily complex mechanisms can be repurposed over time through practice and precedent. However, if we want to have a constructive debate, we should be precise about where the legal risks actually lie. Network fees do not “flow” automatically from the extension of general authorisation, nor from the mere existence of a conciliation framework. They would require specific, additional legal steps – above all, the re‑qualification of CAPs and CDNs as public network operators and the introduction of concrete payment obligations.

Precisely because I value the author’s work and the broader discussion it has sparked, I think it is important to distinguish carefully between regulatory jurisdiction, authorisation status, and pricing obligations. Otherwise, we risk turning “network fees” into a catch‑all label that obscures the real legal and institutional questions at stake, especially for those actors whose primary concern is avoiding being treated as telecom operators by another name.