The Independent Communications Authority of South Africa (ICASA) gazetted final regulations on 22 May 2026 opening two new bands for wireless broadband — and, unusually, wrote the country's biggest mobile operators out of the picture entirely. The lower 6GHz band (5.925–6.425GHz, roughly 500MHz) becomes fully licence-exempt, available on a shared basis to Wi-Fi deployments, wireless internet service providers (WISPs) and community networks. A second band, 3.8–4.2GHz, opens on a licensed-but-discounted basis for 5G Standalone deployment — but is structured, in the regulations' own words, as a "non-market-based, non-competitive" allocation aimed at "non-dominant players, SMMEs and community network operators." MTN, Vodacom, Telkom and Cell C are not excluded by name; they are excluded by design (TechCentral).
The Case for Locking Out the Incumbents
ICASA's logic deserves a fair hearing before it gets a critique. South Africa's mobile market has consolidated hard around MTN and Vodacom, who between them control the overwhelming majority of licensed spectrum secured in the 2022 IMT auction and have shown, through years of litigation over spectrum-pooling arrangements, that they will use regulatory and legal leverage to protect that position. If ICASA had simply auctioned the new 3.8–4.2GHz band on commercial terms, the same two balance sheets would likely have won it, adding another layer of scarcity to a market where WISPs and community operators already struggle to get contiguous, usable spectrum at any price. Reserving this specific allocation for smaller players is not a blanket assault on the incumbents — they retain their existing 2022 holdings — it is a targeted correction aimed at a segment of the market auctions have never served well: rural connectivity, campus and industrial private networks, and price-competitive fixed wireless. ICASA chairperson Cllr Thabisa Faye framed the underlying dynamic spectrum access framework in March 2025 as a way to lower barriers to entry for SMMEs while making more efficient use of a genuinely scarce resource (ICASA).
The technical groundwork also looks serious rather than rushed. The process ran for over a year — draft regulations were published in March 2025, followed by public consultation, coexistence simulation work, and field trials in KwaZulu-Natal in January 2026, where WISPs running 5G Standalone in the 3.8–4.2GHz band hit download speeds up to 200Mbit/s over ranges beyond 4km without disrupting incumbent satellite earth stations (TechCentral). Access to both bands will run through a "unified spectrum switch," a CSIR-built geolocation database that dynamically assigns operating parameters to stop secondary users from interfering with fixed-link and satellite incumbents — the same architecture South Africa already validated through its earlier TV white-space regime. Paul Colmer of the Wireless Access Providers' Association called it "the most important thing Icasa has ever done in its entire history," and while that is a stakeholder talking his book, WAPA's broader claim that the combined bands could meaningfully expand affordable broadband access is at least grounded in a regulator that ran real trials rather than modelling in a vacuum.
Where the Design Overreaches
The steelman only goes so far. ICASA didn't just tilt the playing field with technical caps like the 3-contiguous-cell limit and asymmetric urban/rural channel allocations (two 10MHz channels in cities, four in rural areas) — it wrote a permanent, non-market exclusion into the regulation's own text. That is a meaningfully different and more fragile instrument than a technology-neutral cap that happens to favour smaller entrants. A cap on contiguous cells calibrated to prevent any single WISP from re-creating MTN-scale dominance is defensible market engineering. A rule that categorically bars four licensed national operators from ever applying is closer to industrial policy by fiat, and it invites exactly the kind of legal challenge South African spectrum policy has been mired in for years — Vodacom has already shown a willingness to sue ICASA over far narrower disputes about how spectrum access was allocated.
The 3-cell cap also cuts against the goal it's meant to serve. A successful WISP that builds out coverage, proves demand and wants to scale past three contiguous cells hits a hard ceiling that has nothing to do with interference management and everything to do with keeping it small. That's a strange way to reward the operators this regulation was designed to help — it protects them from MTN and Vodacom today at the cost of capping their own growth tomorrow. And the entire framework leans on a single, not-yet-designated unified spectrum switch operator; until ICASA names who runs that database and on what commercial terms, every WISP's access to either band is contingent on an entity that doesn't officially exist yet.
The contrast with how South Africa is handling satellite broadband is instructive. Amazon's Kuiper-derived Leo service is entering the market in 2027 not by seeking its own licence but by partnering with Herotel, an established local ISP — a model regulators clearly find comfortable, while Starlink remains locked out over the 30% local-ownership requirement it has refused to meet (TechCabal). The spectrum rules follow the same instinct: channel access through domestic intermediaries rather than open it to whoever can deploy fastest. That instinct is proportionate when it corrects a genuine market failure, as the 6GHz license-exempt band mostly does by simply following the global standard-power Wi-Fi trend already adopted by the US and EU. It becomes protectionism dressed as inclusion when, as in the 3.8–4.2GHz band, the ceiling on success is built into the regulation itself.