The Great Gig in the Sky

There are many details that still need to be ironed out regarding e.g. spectrum fees and registration of satellites. But the one important issue that has not been addressed is the requirement in the Electronic Communications Act (ECA) of a 30%¹ equity ownership by persons from historically disadvantaged groups shareholding in any telecommunications licence. As pointed out by SpaceX in their submissions to ICASA during the August hearings, any new spectrum regulations require a clear approach on how this intersects with the 30% requirement for licence holders².

It is against this background of regulatory hurdles for satellite entry in South Africa, that one must consider the recent distribution agreement between Amazon Leo and Herotel. This deal has received a lot of media attention during the past few weeks. Herotel started out as a Fixed Wireless Access (FWA) provider of broadband services but has since morphed into a fully-fledged Fibre Network Operator (FNO). It has grown into one of SA’s largest Fibre to the Home (FTTH) providers, with a large proportion of customers in secondary towns and peri-urban areas³. Herotel was recently acquired by Maziv, who also owns Vumatel and Dark Fibre Africa (DFA)⁴. Herotel has strong growth ambitions for its business in SA, with a stated aim to reach an “additional 750,000 homes in underserved areas, aiming to expand its total footprint to more than 1.1 million homes and reach approximately 6 million people”⁵.

It is well understood in the telecommunications market in South Africa, that the next frontier in terms of broadband access is serving people outside of the traditional “leafy suburbs” where fibre roll-out was initially focussed, into township and rural areas. However, fibre roll-out is expensive. In cities and high-density areas, there are the trenching costs, municipal wayleaves which may delay operations, etc. But if the numbers are there in terms of demand from households and businesses, then the fibre investment makes sense. From an economic point of view, it is also important to understand that fibre roll out is characterised by a strong “first mover” advantage. That is why you do not usually find a 4th and 5th FNO in most residential suburbs (or even in business areas). The first mover can tie up a significant part of the addressable market, meaning successive FNOs get access to a smaller and smaller part of the market.

But for SA to bridge the digital divide and democratise internet access (policy goals that are repeated in numerous policy documents⁶), it matters that people in townships and rural areas get access to the internet. And given rapid technological advancement and the rise of Artificial Intelligence (AI), quality broadband access becomes increasingly important. While some fixed wireless access (FWA) packages can deliver high speed internet access, fibre is generally viewed as the “end game”.

But if one were to rely on fibre to connect the whole of SA, you will wait a long time since the business case will not make sense in rural and even some peri-urban areas. In countries like France, rural fibre deployment is subsidised by the government⁷. Research by Analysys Mason on European rural broadband found FTTH becomes dramatically more expensive in the least dense 5–10% of areas, up to five times the average cost. In contrast the costs of satellite connectivity remain essentially flat regardless of how remote the customer is⁸.

It is against this background that one can understand the excitement about the distribution deal between Herotel and Amazon Leo. Herotel has access to peri-urban and rural customers and has ambitious growth plans. A LEO satellite may be just the technology required to fulfil these growth plans.

But what is a low earth orbit (LEO) satellite and is Amazon Leo the best partner for this service in SA?

LEO satellites orbit between 160 and 2,000 kilometres, circling the Earth every 88 to 127 minutes. The advantage of their low altitude is that it reduces latency significantly, compared to Geostationary Earth Orbit (GEO)⁹ and Medium Earth Orbit (MEO) satellites, and therefore can provide speeds comparable to fibre networks. This performance makes LEO satellites suitable for home and enterprise broadband solutions wherever a good fixed or mobile connection is inconvenient, unreliable or unavailable. LEO satellites are increasingly being deployed for high-speed connectivity in remote and underserved areas. The one drawback is that their lower altitude requires hundreds of satellites for global coverage.¹⁰

Given the requirement of hundreds of LEOs for high-speed coverage, does Amazon Leo fit this bill?

The first thing to note is that Amazon LEO has not launched commercially in any jurisdiction to date. Secondly, as of July 2026, Amazon Leo had 396 satellites in orbit, out of a planned constellation of 3,236 satellites¹¹. In contrast, Starlink, has roughly 10,900 satellites in its active orbital catalogue (as of mid-August 2026). Starlink therefore has 27 satellites for every one of the Amazon LEO satellites. While Amazon has plans to launch more satellites, it will be a catch-up game for the next number of years as Starlink has been building its constellation since 2019, versus Amazon Leo’s deployment which started much later.¹²

Would Starlink therefore have been a better partner to Herotel, or more to the point as a new entrant providing rural broadband access and helping SA to bridge the digital divide?

On the numbers, and given the technological requirements of LEOs in terms of density of coverage, it would certainly seem that the Starlink option would have been more advantageous for SA in terms of global coverage. While Starlink has indicated its intentions to enter SA as a LEO satellite provider, it has run into regulatory hurdles that at the moment, seem too large to overcome. While Starlink is on record that it is willing to give 30% of its equity in the form of an “equity equivalence programme”, it is not willing to give away 30% shareholding (the current Broad Based Black Economic (B-BBBE) requirement for licence holders in the telecommunications sector in SA). Starlink has spent billions to develop the proprietary technology to develop its satellites, and giving away 30% of this intellectual property rights in the form of equity shares is simply not attractive to this globally successful firm.¹³

The populist voices¹⁴ celebrating the exclusion of Starlink on the basis that the company does not want to adhere to B-BBBE policies, probably misses the more nuanced point that Starlink is willing to invest the required 30% in the form of an “equity equivalent” programme, which will of course benefit many South Africans. But until this regulatory impasse can be overcome, Starlink will continue to operate in our neighbouring countries, but South Africans will not have access to its services.

The issue is that one must see the Herotel/ Amazon Leo deal in this light. We effectively have the “second best” option in terms of the number of satellites and coverage because of regulatory hurdles. In terms of competition economics, Amazon Leo will now get a “first mover” advantage, but apparently not on the merits. The Amazon Leo/ Herotel transaction is a brilliant response to the regulatory hurdles set by the regulator (ICASA) and B-BBBE policies. And while the intention of the new draft Spectrum Licence regulations may be good, it does not address the equity issue that is keeping Starlink out of the SA market.

Since the Amazon LEO deal with Herotel is a distribution agreement, it does not need a telecommunications licence and there is no associated B-BBBE licence requirement (since Herotel holds the necessary licences). We therefore have the “second best” satellite alternative, which will apparently start providing satellite internet access in SA from 2027 (via the Evry brand) with no known track record in this field. But importantly, it does not seem that there is any requirement on Amazon Leo to invest anything in SA apart from this distribution agreement. Amazon seems to have ambitions for growth in SA through its global acquisition of Globalstar, which also has a business (albeit small) in SA.¹⁵

But SA is missing out on the larger player with more satellites, Starlink, that was willing to also invest 30% of its total SA value in equity equivalent programmes. One should therefore consider the excitement about the Amazon Leo/ Herotel deal in the context that we may indeed be getting an inferior solution that may negatively impact on consumers who may be in dire need of quality internet access in outlying areas.

For the moment, Amazon Leo has won the race to provide satellite access in SA, but not on the merits at the infrastructure level. It has won the race because the main rival, Starlink, is excluded by regulatory hurdles. This is a severe market failure, as we now have the second-best alternative, with a first mover advantage and no rival in the race. This may impact quality and affordability in the medium term, until a new entrant is allowed.