South Africa’s race to get connected

Jeremy: South Africa needs new sources of growth and productive investment, and one sector is attracting both: digital infrastructure. Now, government says South Africa already has 55 data centres with more than 50 billion rand in further digital infrastructure investment expected over the next three years or so.

But this is about more than just data centres. It's about fibre, mobile networks, subsea cables, satellites, cloud infrastructure, and the power that keeps the system running. Demand is rising fast, driven by cloud computing, artificial intelligence, and an economy increasingly dependent on reliable connectivity.

Now, the opportunity is substantial, but so too are the constraints. So can South Africa supply the power, the capital, and regulatory certainty that this investment actually needs? Can new infrastructure reach beyond the country's wealthiest urban centres? And how do investors separate durable demand from the risk of overbuilding?

Hello, I'm Jeremy Maggs, and this is No Ordinary Wednesday, Investec's fortnightly podcast on the forces shaping business, markets, and economies. 

Joining me are Louise Pillay, equity analyst covering telecoms and media, and David Abbey from Leveraged Finance, both at Investec. And today we are going to ask whether digital infrastructure can become a genuine engine of South African growth.

So Louise and David, a very warm welcome to both of you to No Ordinary Wednesday.

Jeremy: David, 10 years ago, most people would probably have thought about mobile towers and fibre as digital infrastructure. Today you're going to tell me it's a very different picture. So what has changed and where do you think investment opportunities or activity are at its strongest?

David: Jeremy, speaking of mobile, 10 years ago, the iPhone 7, almost to the day, was released, and that release had double the processing capacity and double the storage of the previous iPhone 6 that was released a year before that. And that signifies one key theme: data consumption. Companies found themselves maturing from simple data storage to efficient processing and monetising of data.

As data consumption has grown, so have the likes of Amazon Web Services, Microsoft and Google, which have moved from a few centralised regions into a global network of data centres and cloud zones.

Additionally, traditional CPU-based data processing has evolved and given way to specialised AI networks that allow GPUs to handle generative AI workloads and the development of bespoke large language models.

All of this culminates in the broadening of digital infrastructure from traditional towers, which in themselves have increased their capacity through 3G, 4G, 5G and next-generation technologies, as well as fibre networks that have also had increased competition from new technologies such as 5G over-the-air networks.  

This broadens the scope of where technology is moving and what will enable technological progress around the world.

Jeremy: And all of this in the space of 10 years. David, I must tell you, I can't even remember what the iPhone 7 looks like. So, it's astonishing and interesting that we're talking about it on this day. 

All right, Louise, South Africa's coverage numbers are strong, we know that. But affordability and rural access are still problems, they still lag.

Do you think we are confusing coverage with real digital inclusion?

Louise: Thanks Jeremy for the great question. Network coverage is where the signal reaches you, whereas inclusion is whether you can use that data affordably, consistently and appropriately to make your life better. So, in a more useful way. 

If you look at the stats ICASA has put out, 4G coverage reaches about 99.5% of South Africa. According to Stats SA, household penetration for internet access anywhere stands at about 80%. Now, let's take that back to the home, where we see that internet in the form of fibre, fixed wireless access is sitting at about 17% household penetration.

So there is quite a gap in terms of rural and urban rollout of network coverage. Where could the barriers be? Smartphone adoption, affordable devices and access to a reliable electricity source, especially in rural areas, can be significant barriers to entry for digital services.

Jeremy: All right, David, let me throw it back to you.

Investment, as you referenced earlier, in data centres in South Africa as well as cloud and AI infrastructure is accelerating; that we know. You spoke about Microsoft and Amazon committing billions. What do you think then is driving that demand for scale?

David: Simple answer: digitisation and increased computing power.

With the emergence of AI and other advances in technology, different sectors within the market have digitised, trust frameworks have been developed, data protection laws have been promulgated, and cybersecurity protocols have become mandatory layers of national infrastructure rather than just a nice-to-have add-on.

If we bring it to the everyday world, cars are a lot more digitised today than they ever have been. With the emergence of e-commerce, if you look at grocery shopping and ordering takeaways, all of that is done online today, whereas if we go back to 10 years ago, not much of that was done on a cellphone.

All of that has driven the demand for digital infrastructure to support these technological advancements and to support the economy of companies that are growing into this digital environment, and that is what is driving this technological advancement.

Jeremy: So Louise, let's push it on a little bit then. As demand for AI grows, obviously, so will the demand for data centres.

But is there a risk of overbuilding?

[00:06:21] Louise: I've covered the digital infrastructure, or let's say telecommunications, industry for a number of years. When I first started out, it was standard to have a six- to seven-player market structure. In recent times, we've seen that the best market structure in mobile is two to three players.

In terms of the fibre space, we've also seen this overbuild. Fifteen years ago, driving around Joburg, you would see a trench being dug one month, closed the next and reopened the following month. So you've seen this overbuild across different technologies. Coming back to data centres specifically, there will be some, let's say, additional need for certain cities. 

For example, a city like Johannesburg will need to service or have more AI servicing capacity in its data centres compared to the smaller cities in South Africa. So there will be an assessment that will need to be done. What is key is that each data centre provider will need to have an anchor tenant.

They would need to have adequate planning and committed capacity so that they can deliver a positive return to their investors. That is key in the medium to long term.

Jeremy: So David, fair to say with electricity supply improving, and that's a given right now, do you think that grid access is now the bigger constraint for data centres?

David: Data centres need three key ingredients, among others, to survive: power, cooling and connectivity. Power is a key part of this because data centres are measured by uptime. And now without power, you lose uptime, and the companies that those data centres support cannot afford to be offline at all.

If you think about radio and if you think about banking, we cannot afford to be offline at any point in time, otherwise our customers are affected down the line. 

So the availability of more electricity on the grid is a definite plus to data centres. Allied to that, data centre providers are looking at a mix of power, not just grid power, but also renewable power that comes from renewable energy and new energy sources that allow them to supplement grid energy to limit that bottleneck that the grid power brings.

Jeremy: I'm so glad you raised the issue of cooling, and let me come back to you, Louise, on that. Let's add water and social licence to this equation. Do you think power and water are becoming regulatory and reputational risks as much as the bigger questions around engineering? I would imagine, yes.

Louise: Yes, absolutely.

In terms of power usage, the electricity used by a 100-megawatt data centre could power 100,000 households. To put it in simple terms, every 10 megawatts of power used by a data centre could power 10,000 households. So you've seen the exorbitant electricity usage, but there are ways to manage it if data centre providers plan adequately and work with municipalities through public-private partnerships to assess worst-case scenarios.

One scenario is a drought, as we're heading into El Niño. The second there is a grid failure or a return to stage six load shedding. There are ways to try to mitigate those risks. Right now, the government of South Africa is trying to work quite closely with the sector.

We've seen efficient mechanisms on public-private partnerships. So I hope, especially, the major data centres look to work with the community and with government at municipal level.

Jeremy: All right, both of you hold your thoughts. We're going to continue this conversation in just a moment with a look at how digital infrastructure is financed, a critical part of this conversation, and whether regulation is in fact keeping pace and where satellite technology is going to fit into all of this.

But first, a quick reminder to follow Investec Focus Radio SA wherever you get your podcasts or on YouTube. If you value informed conversations on the forces shaping economies, markets, and businesses, please subscribe to the channel and also take a moment to rate us. So let's continue with the conversation.

Jeremy: David, back to you. Let's follow the money, shall we? Always a good thing to do. When you assess projects like data centres, what turns an attractive tech story into a financeable infrastructure asset?

David: I'd like to talk about the attractiveness of the tech story first. Similar to all infrastructure, digital infrastructure enables sectors to thrive, and developing the infrastructure, certainly in Africa, allows Africa to be competitive on a global basis.

The world is going digital, and technology is enabling that growth, and it's incredibly important that Africa doesn't lag behind. Secondly, what justifies funding these digital projects is essentially the business case behind them. What prevents everyone from putting up a data centre around town is the need for a specific business case: what is that data centre being developed for, and is there demand for what it provides?

These are incredibly expensive infrastructure projects, which limits the ability of just anyone to put one up. Reviewing the business case means understanding who the offtaker is, where the demand is coming from and whether there is a business case or a social case. That is what we fund against, the cash flows from the infrastructure being built.

Jeremy: And of course, Louise, overarching all of this is regulation, right? So do you think new deployment rules can meaningfully speed up fibre rollout? And where do you think the biggest regulatory bottleneck is right now?

Louise: So that's very topical at the moment because ICASA has recently released draft regulations on rapid deployment.

I attended some of the public hearings that ICASA hosted, and the key takeaway is that some FNOs, fibre network operators, are waiting 6 to 12 months for municipal wayleaves, right? So these are approvals to trench around a public road or so, and they have to wait for municipal approval before they commence work, and that significantly delays it.

So what the industry is advocating for is for that to be regulated within 30 days. So imagine you apply for fibre, especially in a peri-urban area, and you have your fibre within, let's say, 2 months, and that would be a great achievement for us in South Africa. However, our different laws and regulatory bodies limit ICASA's authority because some of that will actually sit within the municipal guidelines, and every municipality has different guidelines.

You're dealing with different people, dealing with different stakeholders. So for me that's the biggest bottleneck. It's trenching that fibre and getting it into households in a timely manner.

Jeremy: So let's move from digging trenches and let's cast our eyes heavenward if we can. I want to talk about satellite connectivity.

Does satellite compete with fibre and mobile, Louise, or does it simply fill the gaps that can't be reached economically?

Louise: It does compete, but there is a world where both can coexist. Satellite reaches the most remote areas, where you cannot have multiple base stations or trench fibre in the Kruger National Park, for example.

You cannot do this deep into a mine or mining area. So you'll need satellite to try and help with the network access in those types of regions. So you definitely need both complementary technologies in the medium to long term. If you look at South Africa's national road network, we have about 750,000 of road network, but only 250,000 of fibre.

For operators, trenching fibre has a longer payback period, so the returns profile is a bit more drawn out. So satellite would bridge the coverage gap. It'll also help narrow the cost curve for the overall mobile connectivity.

Jeremy: So David, if that's the case, does satellite change the economics of rural connectivity, or do you think that fibre is always going to remain essential?

David: I agree with Louise, they are complementary. If you look at the value of fibre, it's the density of the area in which you lay the fibre. A fibre producer wants to put fibre in an area that is densely populated.

So a fibre provider will not go into a remote area because the take-up there will be incredibly low, and the value and business case wouldn't stack up, and you wouldn't get the funding that you need to go trench in those areas. And that is the gap that satellite bridges or covers to ensure that we can achieve almost complete connectivity within a specific region.

Jeremy: So Louise, let's talk about the important role now of policymakers. How do they balance technological access, investment, and consistent rules for operators, all of whom are in competition?

Louise: Let's start with the outcome we want. We want more people connected, reliably and consistently. So in order for us to regulate that, if you are serving the same customer, the same type of regulations would need to apply to the different technologies, whether mobile, satellite or fibre.

Having regulatory parity across all the different technology streams is imperative going forward. There's also a risk of over-regulating, and coming back home, African telecommunications regulators have generally burdened the sector, in my view, with a fair amount of regulation compared to maybe our global peers.

Jeremy: All right, last question to both of you. David, firstly to you, who then funds the next phase of South Africa's broadband build-out? We spoke a little earlier about public-private partnerships. Is it infrastructure sharing and institutional capital? Do they become more important? What does the landscape look like?

David: The sector is fortunate to attract lots of capital from a range of sources. Public-private partnerships will probably relate more to infrastructure, so land and everything else that's required to actually establish these networks. Institutional capital is already active in the space, and so are traditional lenders.

At Investec, as a leader in this space, not only in capital provision but also as an adviser and structurer, we are certainly open for business to fund these digital projects that help advance technologies across Africa and beyond.

Jeremy: And Louise, finally then, what policy, market, or competitive development do you think would most change your view of the sector, either positively or negatively?

Louise: On the positive side, some of the regulations that ICASA has proposed. It shouldn't take two to three years for draft regulations to be gazetted. That timeframe needs to be shortened. We spoke about this earlier. We cannot wait 6 to 12 months for fibre to be trenched to households.

So we need shorter deployment times. If regulators can set specific timeframes, they need to do so. Let's be honest, a lot of the operators are ready to invest capital, but what limits them is regulatory uncertainty.

Jeremy: And I wonder if we have this conversation in 10 years’ time, David, what device you'll be talking about then. 

Let's finish the conversation. Louise Pillay, David Abbey, thank you both for joining me on this edition of No Ordinary Wednesday. Just a quick reminder, a new episode of the programme drops every fortnight.

To make sure that you don't miss a show, follow Investec Focus Radio SA wherever you get your podcasts or subscribe on YouTube. And if you value the channel, please take a moment to rate us. It helps more listeners find the programme. So until next time, goodbye from me, Jeremy Maggs, and the entire Focus Radio team.

Disclaimer: The views expressed are those of the contributors at the time of publication and do not necessarily represent the views of the firm. They should not be taken as advice or recommendations. Investec Limited and subsidiaries: authorised financial services providers, registered credit providers, and long-term insurer.

Focus and its related content is for informational purposes only. The opinions featured on the site are not to be considered as the opinions of Investec and do not constitute financial or other advice. The information presented is subject to completion, revision, verification and amendment.

Although information has been obtained from sources believed to be reliable, Investec Securities Proprietary Limited (1972/008905/07) or its affiliates and/or subsidiaries (collectively “ISL”) does not warrant its completeness or accuracy. Opinions and estimates represent ISL’s view at the time of going to press and are subject to change without notice. Past performance is not indicative of future returns. The information contained herein is for information purposes only and readers should not rely on such information as advice in relation to a specific issue without taking financial, banking, investment or other professional advice. ISL and/or its employees and/or other Investec Companies may hold a position in securities or financial instruments mentioned herein. The information contained in this document alone does not constitute an offer or solicitation of investment, financial or banking services by ISL. ISL accepts no liability for any loss or damage of whatsoever nature including, but not limited to, loss of profits, goodwill or any type of financial or other pecuniary or direct or indirect or consequential loss howsoever arising whether in negligence or for breach of contract or other duty as a result of use of the reliance on information contained in this document, whether authorised or not. This document may not be reproduced in whole or in part or copies circulated without the prior written consent of ISL.

Full Investec Bank Limited disclaimer