Arcep's public consultation on how to reassign France's mobile spectrum closed on 23 September 2026 at 18:00 Paris time. It opened on 23 July. The regulator has said it will publish the responses, excluding material covered by business secrecy. The consultation does not settle who gets which frequencies. It settles the rules and the order of events, and those choices will shape how much Orange, SFR, Bouygues Telecom and Free Mobile are willing to invest over the next decade.
What Arcep actually proposed
According to Arcep's July 2026 announcement, every mobile authorisation in mainland France expires between 2030 and 2035. The exception is the 3.5 GHz band, where licences carry a five-year extension to 2040, subject to the holder's agreement. Arcep proposes a two-stage approach. The first stage would be conditional reassignments that bring the expiry dates of five authorisations to a common date. The second would be a competitive selection procedure. Bbox-Mag's account of the consultation lists the five bands as 800 MHz, 900 MHz, 1800 MHz, 2.1 GHz and 2.6 GHz. It describes the competitive stage as comparable to the earlier 4G and 5G allocations.
The question the consultation puts to the sector is whether the timetable for reassigning frequencies already in use should be separate from the one for allocating new bands. Arcep's own wording calls for a calendar of reassignments "décorrélé" from that for new bands. The distinction is not academic. Arcep's earlier consultation cycle ran from October 2025 and raised new candidates: the 3410–3490 MHz, 1.4 GHz and 6 GHz bands, plus direct-to-device satellite services and 6G.
The strongest case for tight regulatory control
The argument for a hard, competitive process deserves a fair statement. Spectrum is a scarce public resource. Four incumbents hold all of it, and a renewal that quietly favours them could lock out challengers and undo the competitive gains France's four-operator market delivered. Competitive selection also forces holders to reveal what the spectrum is worth. The 2020 3.5 GHz award shows the mechanism working. According to Arcep's page on that award, each operator secured a 50 MHz block for €350 million, and the auction of additional 10 MHz blocks closed at €126 million per block against a €70 million floor. Renewals without competition would forgo that price signal, and Arcep is right to keep the auction available.
Why decoupling is the pro-innovation choice
The risk on the other side is that uncertainty is itself a cost. Operators must commit capital to sites, fibre backhaul and radio equipment years before their licences run out. If renewal is bundled with the allocation of brand-new bands, a delay or dispute over one can freeze decisions on the other. Separating the calendars means operators know the terms for the bands they already run without waiting on 6 GHz or satellite questions. Arcep says as much: it is starting now so that conditions can be set well ahead of the first expiries, giving the sector predictability.
The common-date mechanism is where the design needs scrutiny. Aligning five staggered expiries in the 800 MHz to 2.6 GHz range simplifies the second-stage procedure. It also concentrates risk. If all five bands come up for selection together, an incumbent could face the loss of its entire low- and mid-band base at once. That would create pressure to bid at almost any price, and a price paid to the state is money not spent on networks. Arcep should therefore state clearly, before the second stage, what the conditions for the first-stage reassignments will be, and how far the bidding can reach into the spectrum that carries operators' existing customers.
Obligations: coverage, not a new tax
According to Bbox-Mag, the framework would let existing coverage commitments be "renewed or even reinforced" and could add new ones on coverage and territorial connectivity. Rural and indoor coverage is a legitimate public aim, and licence conditions are a standard way to secure it. The test should be proportionality: obligations tied to measurable outcomes, such as a stated share of the population or of roads covered at a defined quality, rather than open-ended promises that raise the cost of holding spectrum. Every euro of obligation is priced into what bidders will pay, so obligations and fees should be judged as one package. Arcep's earlier statement that reorganising the 3.5 GHz band will cost operators dealt with the cost of moving spectrum. It should not become a template for stacking costs at the same moment the licences renew.
What to watch when responses are published
Three things will show whether the consultation moves toward a workable outcome.
- Whether operators accept the 2040 extension at 3.5 GHz. It requires the holder's agreement, so it will show how they weigh a longer horizon against the conditions attached.
- Whether challengers and incumbents diverge on the common date. Free Mobile and the three larger operators may not want the same thing from a joint selection.
- Whether local authorities push for coverage duties beyond what operators say is economic. That trade-off is the hardest one in the file.
The consultation page states that the responses will be published, apart from trade-secret material. That transparency is worth defending. It lets outsiders test whether a competitive process is being designed for competition or for the convenience of the four incumbents.
Arcep has not announced a final decision date. Its choice to split renewals from new allocations is sound. The proof will be in the terms: a renewal path that is predictable enough to justify investment, and an auction that stays real enough to keep the market open.