Concerns about so-called “Fair Share” proposals have been central to debates in anticipation of the EU’s Digital Networks Act (DNA).
These proposals, requiring large content and application providers (CAPs) to contribute financially to network costs, have resurfaced repeatedly over the past years, often framed as addressing traffic-driven market failures.
The Internet Society has consistently opposed these proposals, and extensively detailed why the Fair Share narrative is flawed and risks fragmenting the global Internet.
With the European Commission’s release of its first draft of the DNA, these long-standing concerns are back in focus. In this blog post, we’ll take a critical look at the draft through this lens, examining what the DNA says, the assumptions it reflects, and what it may enable over time.
Understanding Fair Share—Three Concerns
Before turning to the DNA itself, it is worth restating three concerns that frame our critique:
- First, there is no problem that Fair Share rules would solve. There is no evidence of free riding by CAPs, no systemic market failure in IP interconnection, and no breakdown in cooperation between network operators and content providers. On the contrary, the Internet ecosystem is characterised by extensive voluntary collaboration, with aligned incentives on both sides to exchange traffic as efficiently as possible.
- Second, Fair Share payments conflict with the Internet’s architecture because they are a regulatory construct rather than a real need. These proposals seek to force CAPs into paid arrangements through regulatory mechanisms such as dispute settlement, reflecting a problematic principle we refer to as cross-network responsibility: the idea that a service provider should be accountable for costs or service quality across networks they do not operate or control. Such a principle is a radical departure—in practical terms an absolute rupture—from how the Internet works, where each network is responsible for its own operations, investment, and costs.
- Third, Fair Share proposals are incompatible with net neutrality rules. Net neutrality prevents network operators from using traffic management as leverage to pressure CAPs into paid arrangements, for example, by congesting transit links. Fair Share would turn this on its head by granting operators a regulatory right to payments based on traffic source or volume, effectively authorizing discrimination between traffic flows.
A Cynic’s Review of the DNA
At first glance, the draft appears to step back from Fair Share. It imposes no explicit obligations on CAPs to pay for network costs, and net neutrality principles are preserved via the embedded Open Internet Regulation (OIR).
However, through the lens of our three concerns, the draft is less reassuring. It legislates solutions to problems that do not exist, creates regulatory machinery for future intervention, and embeds net neutrality within a framework that risks weakening it.
Concern 1: Legislating for a Problem That Does Not Exist (Articles 191-193)
Article 191 requires BEREC to publish guidelines facilitating “cooperation on technical and commercial matters” between network operators and undertakings in closely related sectors. Article 192 establishes a “Facility for Voluntary Conciliation,” under which national regulatory authorities (NRAs) convene structured meetings, BEREC issues opinions, and written accounts document the parties’ positions.
The implicit premise for these provisions is that cooperation between network operators and CAPs is somehow lacking or insufficient, a premise that is difficult to square with reality.
BEREC investigations have consistently found IP interconnection markets function competitively, with no evidence of market failure justifying intervention, and the case for Fair Share has been tested through numerous EU consultations and consistently rejected. Instead, voluntary collaboration is already the dominant mode of operation. Internet exchange points enable vast traffic volumes through cooperative arrangements. CDNs are widely used, settlement-free peering remains widespread, and CAPs invest heavily in compression techniques to reduce network load. In other words, incentives are already aligned.
Against this background, legislating a framework for “ecosystem cooperation” appears unnecessary at best. If there is no problem to be addressed, why build a framework to solve it?
Concern 2: Fair Share Payments as a Regulatory Construct (Articles 192 and 193)
Formally, conciliation under Article 192 is voluntary and produces no binding outcomes. It does not empower regulators to impose terms, mandate agreements, or require payments.
However, its practical effect could be less benign. Conciliation institutionalises negotiation within an EU-recognised framework, generates records, and creates visibility over refusals to agree. Nothing prevents a party from initiating conciliation requests strategically by filing repeated requests, documenting every refusal, and building a paper trail that portrays normal commercial disagreement as systemic failure to cooperate.
Furthermore, Article 193 mandates a Commission review of “ecosystem cooperation” three years after application, taking into account a BEREC report assessing how the framework has functioned. By that point, the conciliation process will have generated documentation: requests, opinions, written accounts, records of positions taken and refused. If agreements are not reached, even where no agreement is necessary for markets to function, this material can be cited as evidence that voluntary cooperation has “failed.”
In other words, the mechanism risks becoming a solution in search of a problem, designed to justify further interventions.
Concern 3: Collision with Net Neutrality Rules
The DNA incorporates the Open Internet Regulation (OIR) into its broader framework, which is a completely unnecessary step. The OIR already exists as a separate regulation with direct effect, robust enforcement, and clear case law. And even if incorporating the OIR into the DNA would not change the net neutrality rules per se, it does change their context. Net neutrality is no longer a stand-alone instrument to protect end-users, but now read alongside objectives such as resilience, performance, and ecosystem cooperation.
The DNA’s sustainability provisions are a case in point. As noted by telecom expert Rudolf van der Berg, commenting on an earlier leaked draft, concepts such as “end-to-end traffic efficiency” frame digital sustainability as a justification for regulatory involvement in traffic management and IP interconnection. Once such an intervention is accepted as legitimate, net neutrality is no longer assessed solely against its original purpose of preventing discrimination but is instead placed against competing objectives such as efficiency and sustainability.
In other words, while this shift does not weaken net neutrality on paper, it undermines its foundations in practice and makes future reinterpretation easier. This is why keeping the OIR as a separate, standalone regulation matters: it preserves interpretive autonomy and prevents the contextual drift that embedding enables.
A Cynic’s Conclusion
The Commission’s draft is only one of many that will be debated. Parliament and Council will produce their own versions, and the final regulation will emerge from trilogue. The concerns here should be read with this in mind: what matters is not only what the text does today, but what it makes possible.
Two risks deserve extra attention. The first is that specific mechanisms are strengthened. The provisions on ecosystem cooperation, voluntary conciliation, and BEREC guidelines are the kind of “soft” mechanisms that intensive lobbying can harden into de facto fair-share rules. A voluntary framework of cooperation can become mandatory. Non-binding conciliation can become a duty. Reviews can be accelerated or their criteria tweaked.
The second risk is subtler but no less significant: the draft carries forward flawed assumptions and embeds them in EU law. The very existence of an “ecosystem cooperation” framework implies that cooperation is lacking.
For instance, the recitals’ references to “fair, reasonable, and proportionate use of the resources of the other party,” and for traffic to “not lead to disproportionate or economically unsustainable investment needs for network providers” echo language used by Fair Share proponents to imply that current arrangements may be unfair or disproportionate.
Once encoded in legislation, these become the starting point for future interpretation and intervention, even though the underlying claims have been previously examined and rejected.
Three Changes for Future Drafts
Given the concerns outlined above—the absence of a problem to solve, the construction of regulatory machinery for future intervention, and the collision with net neutrality—we believe three changes to the current draft are necessary. Each addresses one of the concerns outlined above, and none would undermine the regulation’s stated objectives.
First, the recitals should clearly recognise that ecosystem cooperation already functions well, with no demonstrated market failure requiring intervention, as evidenced by BEREC’s findings and established market practice.
Second, Articles 191-193 should be removed. Where no market failure exists, conciliation and subsequent reviews add no value and serve primarily as a tool for leverage rather than resolution. Removing these articles eliminates the risk that they are strengthened during the legislative process into something far more consequential.
Third, the OIR should remain separate, preserving its clarity and interpretive autonomy. It is sufficient to reference its provisions, rather than incorporating them into the DNA.
In summary, while the DNA draft avoids explicit “Fair Share” obligations, it embeds assumptions and mechanisms that could enable future interventions, weaken net neutrality, and misrepresent cooperation in the ecosystem.
Policymakers and other stakeholders should remain vigilant throughout the legislative process, ensuring that future drafts preserve voluntary collaboration, protect the Internet’s open architecture, and maintain the OIR as a separate, standalone regulation.
Image © Guillaume Périgois via Unsplash