On 28 August 2026, South Africa's communications regulator, the Independent Communications Authority of South Africa (ICASA), extended the deadline for written submissions on two sets of drafts. The first is its draft radio frequency spectrum assignment plans (RFSAPs) for the mobile broadband (IMT) bands. The second is its draft exclusive geographic areas for millimetre-wave (mmWave) spectrum. The notice ran in Government Gazette 55268 as Notice 4120 and set a new deadline of 30 October 2026. A later erratum, Notice 4139 in Gazette 55343, corrected how Notice 4120 cites the July drafts. It re-cites them as Notices 4040 to 4043. An extension is dull procedure, but it is the part of this overhaul where the public can still change the outcome.
What is on the table
The drafts belong to a wider overhaul. ICASA's 27 July 2026 statement describes an IMT Roadmap 2026 as a five-to-ten-year framework that includes IMT-2030, the ITU's name for 6G. It aims to raise broadband spectrum available for assignment from 850 MHz to at least 18,588 MHz by 2030, which ICASA says would lift high-demand spectrum by about 215%. The package also includes feasibility studies for the identified bands. The draft exclusive geographic areas cover the 24.25-27.5 GHz, 37.5-43.5 GHz, 45.5-47 GHz and 47.2-48.2 GHz bands.
Law firm Webber Wentzel's explainer gives the baseline. It reports that 1,155 MHz is identified for IMT, with 794 MHz already assigned and a further 1,145 MHz of mid-band spectrum under feasibility review. It also lists draft assignment plans for the 1710-2200 MHz, 3300-3800 MHz, 4800-4990 MHz and 6425-7125 MHz bands, plus mmWave. One caveat on sources: Webber Wentzel calls the roadmap finalised, while ICASA's own release describes it as published for consultation. I rely on ICASA's wording.
The case for moving fast
The strongest argument against delay is that spectrum is a wasting asset. South Africa's long wait for high-demand spectrum cost the country years of network quality and affordability. Regulators that dither push operators into expensive densification, and consumers pay for it. ICASA's own release ties the work to the National Development Plan 2030 and the SA Connect broadband policy. Every extra month of consultation can look like a month of delay for the people who most need cheaper, faster data.
That argument deserves respect, and it is why the extension should stay short. The deadline moved to 30 October, not into 2027, and the final plans will only become the basis for later Invitations to Apply under the Radio Frequency Spectrum Regulations, 2015. Moving the deadline does not hold up any licence.
Why the extra time is still worth it
The drafts are technically dense, and the choices in them are hard to undo. Three stand out.
- Exclusive geographic areas for mmWave. Carving mmWave spectrum into geographic exclusivity can suit dense urban use and private networks, and it can also suit fixed wireless in places where fibre is slow to arrive. It can strand spectrum if the areas are drawn too large or too coarse. Operators, enterprise users and rural broadband providers all need time to model what the areas would mean in practice.
- Incumbent migration. Webber Wentzel reports that incumbents in the 1880-1900 MHz sub-band would have to vacate within 24 months of the 1710-2200 MHz plan taking effect. Whether 24 months is realistic depends on equipment lifecycles and network rollout plans, which only the affected licensees can document.
- The scale of the leap. Going from 850 MHz to more than 18,000 MHz in four years is ambitious. Most of the added spectrum would be mmWave, where propagation is short-range and the business case differs sharply from mid-band. A plan that counts megahertz made available is not the same as a plan that counts megahertz put to use.
This last point is where a pro-innovation reading should be careful. Headline megahertz targets are easy to publish and hard to monetise. The better test is whether spectrum is assigned in ways that get used: pro-competitive rules, coverage obligations that are realistic, and a licence-exempt layer for experimentation. The overhaul's reported plans for licence-exempt "innovation spectrum" in 3800-4200 MHz and 5925-6425 MHz point in that direction, and they deserve public support. Wi-Fi 6E-class use of the upper 6 GHz range has generated large unlicensed economic value elsewhere, and unlicensed bands lower the cost of entry for small firms.
What good submissions look like
A longer window is only useful if participants fill it with evidence. Regulators decide better when submissions contain data, not slogans. The most useful responses from operators, vendors, municipalities, rural network builders and civil society would do four things.
- Quantify rollout costs and device availability by band, so ICASA can sequence assignments by what will be used first.
- Test the geographic-area boundaries against real population, traffic and fibre-coverage data.
- Show what the migration windows demand in equipment replacement and service continuity.
- Say plainly where sharing and licence-exempt use could deliver more than exclusive assignment.
ICASA, for its part, should publish how it weighed the comments it receives. Transparency about trade-offs is the cheapest way to build legitimacy for a plan this large, and it lowers litigation risk for later auctions or Invitations to Apply.
Bottom line
The extension is a modest, defensible choice. It sits against a broader and more consequential question: whether a regulator can set ambitious spectrum targets while staying disciplined about what gets assigned, to whom, and with what obligations. Proportionate regulation here means moving at the pace of evidence. It does not mean moving slowly. South Africans benefit most when spectrum reaches networks quickly and stays in use. The 30 October deadline is the last cheap moment to make sure the plan is built on data from the people who must deploy it.