Ghana is awarding 5G spectrum to the operators it once expected to wait for a single shared network. On 6 October, MTN Ghana received a notice of award for new 700 MHz and 3 GHz spectrum, with licence fees of about $200 million. Telecel Ghana was reported to have won spectrum in September. Two years ago, the plan was for one wholesale operator to build Ghana’s 5G, with mobile operators leasing capacity from it.

Ghana is one example of a wider shift. The single wholesale network, where one government-backed operator has exclusive rights to build national infrastructure for retailers to lease, was meant to solve a basic problem: how to expand networks without making several operators build the same infrastructure. A 2019 World Bank-led report estimated that universal broadband in Africa by 2030 would need around $100 billion  in investment. Ghana and Rwanda built versions of the model, and South Africa planned one. All three have since changed course, in different ways.

What is replacing it is a different kind of sharing. Tower companies, fibre providers and operators share separate layers of infrastructure through commercial agreements, and operators keep more control over their networks and spectrum.

This edition of #TechTalkThursday looks at why the single wholesale model stalled, what is replacing it, and what still needs working out.

What the Single Wholesale Network Promised

The pitch was simple: build once, let everyone lease it, and stop operators from duplicating towers, fibre and other network assets. In theory, that could lower costs, speed up coverage and reach areas where several competing networks would not pay. In practice, each country ran into a different problem.

Ghana: the problem was speed

Next Gen Infraco (NGIC) received a ten-year exclusive licence in May 2024 to build shared 5G infrastructure. By March 2026 it had 49 operational 5G sites against an original plan of 4,400, and in January 2025 no operator had leased any capacity from it.

On 15 July 2026, the National Communications Authority (NCA) removed the exclusivity clause. The next day, it opened applications for 5G spectrum in the 700 MHz, 2.3 GHz and 3 GHz bands. Four companies applied, and in September MTN Ghana, Telecel Ghana and Goal Telecom qualified for the next stage. Telecel was reported on 17 September to have been awarded three 2.3 GHz lots.

Then came MTN Ghana’s notice of award on 6 October: two 20 MHz lots in the 700 MHz band and 150 MHz in the 3 GHz band, on 15-year licences. The fees are roughly $100.9 million for the 700 MHz spectrum and $101.1 million for the 3 GHz spectrum. MTN is still completing formalities and payment before the licences are issued, and says it will use the spectrum to expand 5G and bring high-speed broadband to homes.

Ghana has gone from expecting operators to depend on one shared 5G network to letting them buy spectrum and build  to their own commercial priorities.

South Africa: the problem was spectrum

South Africa’s plans for a wholesale open access network (WOAN)  set aside a block of high-demand spectrum for a provider that was never licensed. The 2022 auction sold 700 MHz, 800 MHz, 2.6 GHz and 3.5 GHz spectrum to existing operators, and the government shelved  the WOAN. The spectrum reserved for it stayed unassigned while the market waited. A policy meant to open access ended up delaying access to valuable spectrum.

Rwanda: the network worked, but the market did not

Rwanda is the case that keeps this from being a simple story of wholesale networks failing. Korea Telecom Rwanda Networks (KTRN), a joint venture between Korea Telecom and the government, received a 25-year licence in 2013 and reached about 97% of the population with 4G.

Coverage did not turn into usage. By the start of 2023, only about 2% of Rwandans were using the network. Operators could not build their own 4G networks, wholesale prices stayed high, and retailers had little room to differentiate. After the market opened in 2023, MTN and Airtel launched their own 4G networks, and GSMA reports that 4G penetration rose from 2% to 30%  over the following two years. KTRN still operates and still sells wholesale 4G to MVNOs, but it no longer holds a monopoly.

The lesson from Rwanda is that a wholesale network can deliver coverage without creating the competition needed to drive adoption, pricing and service innovation.

The Problem Was Not Sharing

The three cases share a weakness. An exclusive wholesaler has little pressure to build quickly or price aggressively, and the operators buying from it have less control over coverage timelines, network quality and how they differentiate their services. The companies selling to customers are normally the ones with the strongest reason to invest, and the single wholesale model separates network investment from retail competition.

None of this argues against sharing itself. It argues against making sharing depend on one state-designed monopoly. Emmanuel Manasseh of the ITU explains the economic case:

“Duplication occurs when multiple operators build networks targeting the same customers in the same locations, but they can share infrastructure. So if you already have fibre in a place, it is better to share it rather than having two fibres for two different operators for the same services. For example, if you have a mobile network tower, instead of building another one, you share access to the existing tower. Sharing the infrastructure reduces operational costs, reduces capital expenditure, and also helps to identify the gaps so that they can be filled together rather than duplicating what has already been done by others.”

-Emmanuel Manasseh, Regional Director Africa, International Telecommunication Union (ITU)

What changes is who controls the sharing, who invests, and how much commercial flexibility operators keep.

Three Layers of Sharing

Sharing is now developing across three layers, each at a different stage.

Layer 1: Passive infrastructure

Towers, civil works, power and land are the most established form of sharing. A tower costs roughly the same to build whether it serves one operator or three, which makes duplicate towers hard to justify in sparse, low-ARPU areas. Tower companies let operators swap some upfront capital spending for lease payments.

The money is moving, particularly in the Democratic Republic of Congo. On 4 August 2026, Eastcastle Infrastructure announced a $32.8 million loan to build more than 700 new towers, taking its DRC portfolio from just over 1,000 sites to 1,800. Helios Towers plans to invest about $110 million in the country in 2026, on top of the roughly 2,700 towers it already manages there. Orange and Vodacom’s joint venture, Esengo Towers, plans 1,000 solar-powered rural towers, though in October 2025 it was still waiting for its operating licence and expected to start building in 2026.

Behind these projects is a 2026 to 2035 national strategy to connect nearly 68 million rural residents through tower companies. It is at pilot stage, and public money funds much of it. The infrastructure is shared, but ownership and investment sit with several commercial companies instead of one national entity.

Layer 2: Active networks

Here operators share radio equipment, antennas and base stations while keeping separate core networks and commercial operations. Because active sharing affects capacity and service quality, it is more sensitive than sharing towers, and it is spreading first into secondary cities and rural corridors where parallel networks often do not pay.

MTN and Airtel Africa signed sharing agreements  for Uganda and Nigeria in March 2025 and have since explored similar arrangements in Congo-Brazzaville, Rwanda and Zambia. They say they will keep competing freely in each market. In August 2025, Airtel Africa and Vodacom agreed to share network infrastructure in Mozambique, Tanzania and the DRC, subject to regulatory approval in each country. Operators do not need to own every physical asset to compete. They can share parts of the network and still compete on coverage, pricing, products and customer experience.

Layer 3: Spectrum

Instead of reserving valuable spectrum for a single wholesale entity, regulators are increasingly awarding it directly to operators and letting them decide how to deploy it. Ghana is the latest example. Its first competitive 5G spectrum process has already produced awards for MTN and Telecel, and the NCA expects to finish assigning spectrum before the end of 2026. Rwanda moved the same way in 2023 when it opened its 4G market to competing networks.

Operators that hold spectrum can deploy it according to demand, and can sign voluntary spectrum-sharing or national roaming agreements where sharing makes more sense.

Operators say the attitude has changed:

“Even as competitors, it has become a business imperative for us to collaborate in the provision of critical infrastructure required to build resilient network with strong capacity to support the emerging digital technologies as well as the growing need for data-enabled products and services.”

-Sunil Taldar, Chief Executive Officer, Airtel Africa

Taldar made the comment when Airtel Africa and Vodacom announced their infrastructure sharing agreement. For operators facing high capital costs and rising data demand, sharing is increasingly a commercial decision.

What Still Needs Working Out

The commercial case for sharing is getting clearer, but the day-to-day rules are not. Who gets priority on a shared radio network when it is congested? How are wholesale access charges set? What happens when the network goes down, and who settles a dispute when a partnership breaks? Kenya is developing draft sharing rules that cover some of this, and many other regulators are still building the tools to oversee these arrangements.

There is also the question of what operators do with the money they save. Rethink Research has argued that Vodacom’s sharing deals look more like cost cutting than a push for more investment. The deals are still young, and coverage, network quality and capital spending figures over the next few years will show which it is.

From One Network to Shared Risk

The single wholesale model put the investment risk on one entity, and Ghana, South Africa and Rwanda each showed what happens when that entity cannot deliver. The emerging model spreads the risk. Tower companies finance and operate sites, fibre providers build shared transport, operators share active assets while keeping their own commercial networks, and governments step in where the commercial case is too weak for private capital.

Lacina Koné of Smart Africa puts that last point in practical terms:

“When we talk about investment in infrastructure and the duplications, they are right. The economy is driven by the private sector. The private sector has a say first. At the end of the day, it is a business. It is not a charity. So it is a private sector-first investment before the government can chip in. You may see that the private sector will not deploy infrastructure in areas which are not lucrative, but the government has an obligation not to leave anyone behind.”

-Lacina Koné, CEO, Smart Africa

 

Africa is still building shared infrastructure. What has changed is that sharing no longer needs one network, one owner and one national wholesale model. The test is whether this approach delivers what the wholesale model struggled to: faster investment, wider coverage, stronger competition and more people using the networks being built.