Google and Meta have been spending a small fortune to lay undersea cables bringing more broadband to Africa. Journalists Andrew Blum and Carey Baraka join us this week to talk about their years-in-the-making article for Rest of World asking if those companies will control what Africans can and can’t access online.
Andrew Blum is an American journalist who wrote a 2012 book about the Internet’s physical infrastructures, Tubes: A Journey to the Center of the Internet. Carey Baraka is a Kenyan journalist who has covered a wide variety of local topics for Guernica and The Guardian, and last wrote a piece for Rest of World about McKinsey’s attempts to build smart cities in Kenya.
Transcript
Ethan Zuckerman:
Hey, everybody. This is Ethan Zuckerman, welcome back to Reimagining the Internet. We’ve got two guests with us today. We’ve got Andrew Blum, author and journalist. He’s the author of 2012’s Tubes: A Journey to the Center of the Internet. And more recently, The Weather Machine: A Journey Inside the Forecast. We also have with us, Carey Baraka, who’s a journalist living in Nairobi, Kenya, but from Kisumu. He’s ridden for Guernica, for the Guardian. He previously wrote about Kenya’s attempt to work with McKinsey to build smart cities for Rest of World. Andrew and Carey wrote a terrific piece called Sea Change, for Rest of World, in May of this year. It’s a piece about African connectivity, and I wanted to bring the pair of them on to talk about connectivity and its complications. Hi, gentlemen.
Andrew Blum:
Hello. How’s it going?
Ethan Zuckerman:
Bringing connectivity to Sub-Saharan Africa feels like it should be a feel good story. You both make the case that it’s a little bit more complicated to that. Who’s bringing connectivity to Sub-Saharan Africa now, and why should we be concerned about that? Andrew, do you want to go first?
Andrew Blum:
The folks bringing connectivity to Africa these days, via undersea cables, are Google and Facebook, now Meta. And what’s remarkable about that is that they are building these huge infrastructure systems that wrap around the continent, that are not technically replacing the existing undersea cables, but are functionally replacing the existing undersea cables. Partly because they have so much bandwidth that they make the old cables obsolete by their shear presence, and the effect. Clearly, more bandwidth is always a good thing, but what’s remarkable is to understand that this infrastructure will last for a generation. And we need to understand what the consequences of that may be when these cables are wholly owned, or at least primarily owned by Google and Meta.
Ethan Zuckerman:
Carey, how does this differ from how Kenya’s been previously connected to the internet? Who was building the previous generation of cables, and why is it concerning to the two of you that Google and Meta are now getting into this business?
Carey Baraka:
For a country like Kenya, it was multiple parties. You have private players with private cables, but then you also had certain cables being built by the government itself. So there’s one that was built just solely by the government of Kenya. And then one that was built by the government of Kenya collaboration with the government of South Africa. So it ran from, to touch Mombasa and then head to South Africa. In all honesty, the concern is less with countries like Kenya and South Africa, which are relatively well served by multiple players. But more with countries where, when Google and Facebook and Meta enter, then they will be the only players. Because if it’s just one cable serving the country, then it means that that country’s internet control is seated almost exclusively to the company running that cable.
Ethan Zuckerman:
Let’s talk about that from a legal point of view. Cables, generally speaking, have maintained a pretty high degree of neutrality. Thus far in the internet, it hasn’t mattered all that much whether you’re using a cable that’s owned by Tata or owned by a US corporation, or owned by any of these major global telecoms, the bits generally come through. Let’s back up a little bit and talk about Free Basics, because I think not all of our listeners are going to be familiar with this. Free Basics was a Facebook project designed to provide free mobile phone data access in a number of developing countries.
The idea was it was a reduced data plan, it wasn’t going to let you go everywhere on the internet. It was designed to let you go to Wikipedia, local job searching sites, a couple of other selected sites, and of course Facebook. And the view of this for many critics was that this was Facebook introducing its own non-neutral internet, where it was really the only interpersonal communication service. And it was a way of leveraging its way into markets, taking advantage of the economics of the situation to capture audiences.
It feels like a reasonable concern, but also something of a leap to assume that this happens with Trans-Oceanic cables. Andrew, have we seen cases in which cables have become non-neutral in this fashion?
Andrew Blum:
Well, a couple ways to think about that. I think it’s just worth stepping back and noting that one of the primary goals for Meta and Google, and for that matter Microsoft and Amazon, is to create cohesive global networks of their own. And we used to talk about those networks as primarily logical networks riding on many other parties’ physical infrastructure. But this key shift is that now they’re building their own physical infrastructure. So it used to be that we could say, oh, fine, they have a logical network, the internet’s a network of networks. That network now spans the earth. But there’s always this layering where content providers are running on the physical infrastructure of telecom providers. And there’s some sort of stratification of that. And because of that stratification, there’s some sort of give and take among those parties.
But now we have a vertical integration where content providers are owning their own physical infrastructure. It’s sort of like, well, are they transmitting other people’s content? Are they transmitting their own content? Can they be controlling their own content? Are they controlling other people’s content? There’s a collapse of the differences there, all of which point to the same thing, which is that the internet is primarily made up now of Google and Meta. It’s like, well, are telecoms controlling what goes over their cables? Not necessarily, although some government history of government spying and things like that might counter that a little bit. But we have new territory where you have no longer an internet of a jumbled traffic altogether, but a global internet that is astonishingly dominated by a handful of American companies.
Ethan Zuckerman:
It’s interesting, because this possibility of vertical integration has existed at other points in time. There wasn’t much preventing US telecoms, European telecoms from going in and creating services on top of their data networks. Generally speaking, US companies like Verizon just did a terrible job of doing so, and they ended up partnering with other internet services out there. Carey in the Kenyan context, we have the incredible example of M-Pesa, which has become another key infrastructure owned in no small part by a global telecommunications company. Can you talk at all about how concerns over international control, foreign control, and essential infrastructure have come into play in the Kenyon economy?
Carey Baraka:
The example you give of M-Pesa, which is partly owned by, yes, a foreign telecom giant. But then M-Pesa also owned a lot by the government of Kenya. Which means that it’s not as easy for M-Pesa to be “subsumed” by so-called foreign interest. So the very fact that the government of Kenya has stakes in M-Pesa and it’s inspired company Safaricom, sort of insulates the Kenyan consumer from a lot of things that would happen if it was entirely foreign owned. But then the problem, say what Google and Meta trying to do with internet infrastructure, as Andrew has pointed out, is that they are increasingly trying to own a bigger chunk of the internet.
The problem is less than say a foreign company or American based company is owning the cable that is serving a particular territory. And more, what happens if that company is the only company that owns that cable? M-Pesa is a good example of, yes, it’s owned a lot by, I think British telecom. But then the government of Kenya also has a huge stake, and an inspiring company, it’s Safaricom. Because Safaricom was for a long time a government [inaudible]. And it sort of still is. So it’s like, okay, there’s British telecom, but also there’s the government of Kenya, so control is never entirely seated.
Ethan Zuckerman:
Right. M-Pesa comes to the scene in 2007 as a partnership between Safaricom and Vodafone in the UK. And it’s probably the most successful consumer facing mobile money service on the continent right now. There are a decent number of complaints in Kenya that rival mobile money services that might charge less, have difficulty getting traction because you have a service baked into a very popular network. I was raising it in part because it sounds like that’s the fear that, Carey, you and Andrew are raising in this piece. This idea that connectivity gives the possibility of baking services into it. And obviously with Free Basics, that’s in fact what we saw occur. Andrew, when we were talking earlier, you mentioned that in some cases we’ve seen what almost feel like vanity projects focusing on the feel good effects of Sub-Saharan African connectivity. It sounds like a bête noire of yours is Project Loon. Can you talk about that and how it might give some context for Google’s efforts on the African continent?
Andrew Blum:
Yeah. Well, Project Loon was Google’s X project to supply internet access from balloons floating in the stratosphere, and required a very complicated, elaborate system of following air currents in order to stay above the places where they could be most useful for providing that access. And the Loon did prove its usefulness in different situations, both in Africa, in South America, particularly for disaster response. But it was always striking to me that it was very hyped, very complicated. Certainly very cutting edge, but solving a problem that we had already solved. And the challenge of, which was connecting people to the internet. The more straightforward way to solve that problem would be to dig trenches and lay fiber, which is what’s happened in most of the world.
But clearly that was too big a challenge, for different reasons, for Google. And then in a different context for Meta. I guess what always stuck in my [craw] was this idea that we would spend a lot of attention looking at complex ways to solve a relatively straightforward problem. And rather than focus on the hard work of building infrastructure more generally. I was quite struck by the replacement of Loon with big undersea cables, which are the way that we have successfully connected the entire world, and which are perhaps even cheaper than solving a slightly different problem. But it’s like, okay, well now it’s time to get down to serious business. Now we’re not doing these balloon toys, now we’re actually putting major infrastructure in the water. And what shift does that indicate in seriousness and thinking in the view of the potential market that the continent provides for Google and Meta, rather than a recipient of this somewhat elaborate and ostensibly charitable scheme to provide internet access?
Ethan Zuckerman:
Carey, in a funny way, it seems like Elon Musk may be moving into the same space with Starlink, with this new satellite project. Are you seeing enthusiasm from that within the east African tech community? Or, are you seeing more demand for what Andrew’s talking about, which is the traditional trenches and fiber connected to these undersea cables?
Carey Baraka:
Something like Starlink, going by the prices they’re setting in Nigeria, in Mozambique, they’re not really solving any problem that Andrew is pointing out has not been solved already. Because the problem of, say internet access in Africa, A, first of all, people having access to this internet. But then B, people being able to afford this internet. And C, people being able to access it at a workable speed. So starting to offer into, yes, upgrade speeds exponentially. But just, honestly, the prices they’re giving people in Nigeria and Mozambique means that it’s never going to be workable for the so-called rural communities that they claim to want to connect to the internet. Ain’t no way someone earning $10 a day is going to spend $1,000 on installation and then a few hundred dollars each month on subscribing to the service.
Just like Loon and [inaudible], it seems a lot like it’s like… I talked to someone in the east African tech community, say that it seems like it’s basically a toy project. It’s a pet project. It’s exactly Andrew’s pointing out, it’s not serious work. It’s like, oh, it’s fun to have balloons in the sky, or to have to use radar to beam internet, or to use satellite. But to a lot of people who have done the work of building up internet infrastructure, the work that counts is still in the cables.
Ethan Zuckerman:
Something like Starlink is deservedly getting some credit for providing emergency access in Ukraine. You can imagine it being very, very valuable for a group, whether it’s an NGO or a school in rural areas that aren’t otherwise served. But it’s not a viable solution to wiring hole cities because of the cost per unit, and so on and so forth,
Carey Baraka:
Starlink, in particular, would be very useful in, as you say, Ukraine where there’s a war. So hypothetically, if there was a war in hypothetically northern Mozambique. And again, hypothetically, if one of the components that Tesla needs for its cars is gotten in, again hypothetically, northern Mozambique. And then hypothetically, if Mozambique the first country to get access to Starlink in Africa, just hypothetically, maybe all those are connected.
Ethan Zuckerman:
It’s an interesting hypothetical situation that you put forward there. And I think we’ll probably let people look up the mineral industry of Mozambique on their own on Wikipedia here. But it is an interesting thought that in some cases these projects essentially serve as almost diplomatic efforts for these big companies which have massively multinational interests. And certainly a government like the government of Mozambique may well be very concerned with connectivity, it may be something to offer to citizens. Where you and Andrew seem to be going in this piece is this idea that we’re moving from these one-off, almost toy solutions, into these serious solutions.
Having capacious undersea cables, and then connecting them in meaningful ways via fiber, and getting large chunks of people online. Even if people aren’t doing fiber to the home, suddenly having the ability to have four and 5G back-ended to these very big pipes is potentially an enormous change here. How does Kenya do this right? If the concern is that a Meta or a Google could decide that these cables that they’re laying won’t be neutral, could decide that they’re going to prioritize some traffic over others, how does Kenyan law protect around this? Is there a regulatory authority that can demand neutrality on this cable? Is that supported in law? And if it’s supported in law, is there the technical capacity to monitor and find out what’s going on?
Andrew Blum:
Sorry, Carey, before you… Can I just clarify, put a finer point on one thing? We can worry about whether or not Facebook and Google will treat their traffic equally, but they’re basically saying that they are building these cables for their own traffic. We have to listen to what they’re saying as well. And not only that, but they’re building these cables because they see a huge increase in the number of users coming and they want to serve those users, AKA their customers. So they’re being very clear that these cables are for their own use and for their own business use. And that they fully intend to prioritize their own traffic on these cables, that is why they’re building them. I just think the hypothetical there is there isn’t, we just need to say that this is what they’re actually doing.
Ethan Zuckerman:
And Andrew, do you think that means that literally Google will parse out traffic? So that traffic going to Google affiliated sites will go through their cable, and traffic to other open websites will have to go via another cable? Are there going to be routers and internet exchange points that split the packets up at that point?
Andrew Blum:
I think explicitly, yes. And I think it’s because they are bringing traffic onto their own network, their own global network as early as possible in order to move it as efficiently as possible. So at the earliest possible moment, Google will put traffic going to Google servers on Google’s own global network, and shift that other traffic to whoever, whatever telecom they’re trading with. I mean, that’s very deliberate. They say they’re doing that in order to provide better service, which they do. And all hats off to the quality of a Google reading and how well Google Docs works and Google Classroom, et cetera, et cetera. But it’s by no means a speculative, this is precisely the reason they’re building theses cables. And they say as much. The part that is left unsaid, of course, is that if you want to listen to a track that isn’t on YouTube, then it isn’t going to work as well. And that’s the consolidation of Google’s control of the internet writ large, which they talk about, which is the entire goal. That’s all there is to it.
Carey Baraka:
You raise a point there, it’s a much bigger version of M-Pesa. Where M-Pesa is, yes, revolutionary, and people can send money to whoever. But it’s mostly that people can send money to whoever is on M-Pesa. And whoever is on Safaricom. And for years now, other network providers in Kenya have been filing cases in court, filing cases like regularity authority, essentially accusing Safaricom of being a monopoly. I think that’s exactly what happened. It’s happened that a number of internet service providers have existed in Kenya because they were able to get integration to M-Pesa. Where Safaricom just seemed to point blank refuse to let people transact money across networks.
I can definitely see a situation where it becomes the same with Google. But the legalities of it, if you check back to Safaricom, it was only very recently that some of these companies managed to achieve to get the high code to mandate Safaricom to integrate the mobile money networks. But that’s in a country where, well, there isn’t a lot of heavy legal needs protecting companies and consumers. There is more than a lot of other countries in Africa. So if it was, say in DRC or in Ethiopia, or in this other country, it’s very skeptical. Part of me wants to say that the same thing that happened M-Pesa might happen.
Let me just add the other problem of the fact that a lot of countries in Africa, in the telecom infrastructure still in the hands of the state monopoly. So in a lot of countries, it’s actually the state where it’s just one company controlling everything. There’s a part of me that is curious to see what would happen in this country is if Google or Meta went to the country. Would the state monopoly fight back? There’s that bit of some of these countries, it’s still the state monopoly which controls the internet, which makes it so hypothetically easier during elections or whatever for internet to be shut down because it’s just one person controlling the entire internet. So they’re like, okay, you’re not going to access the internet.
Yeah. And then of course there’s the framework of… Yeah, countries like Kenya, South Africa, Nigeria have very strong regulatory systems. Of course, Nigeria just went into entire war with Twitter. Granted, because of the President’s Twitter habits. But then other [inaudible] like while Kenya has certain laws to protect the consumer, historical with Safaricom, it wasn’t very good at doing that because Safaricom was partly owned by the government, and the government was invested in M-Pesa’s success. So if we were just to push another hypothetical and wonder, Kenya has a minister for ICT. And he used to be the country director. Either the country director or regional director for Google. So it’s like, if it so happened that Google became a monopoly, and if the people in control of [inaudible], say the people, ex-Google people, would they be as invested in fighting that control?
Andrew Blum:
And what’s the likelihood that they return to Google at some point after their government service is also-
Carey Baraka:
Sure, sure.
Andrew Blum:
… extremely high. Yeah.
Ethan Zuckerman:
But Andrew, understanding now a little bit better what the concern is, and it sounds like the concern in many ways is that these aren’t internet cables, these are Facebook cables and Google cables to carry their traffic. They’re being celebrated as internet cables. Is this mostly a case of misbranding? Is this a case of, if Google and Meta are building their own cables, it’s going to undermine the market for open internet cables? Is it a sense that we simply shouldn’t allow single company cables to do Trans-Oceanic systems? What’s your proposed solution here?
Andrew Blum:
Well, I think it’s important to acknowledge that Google and Meta are sharing the cable a little bit. They are swapping fibers both between each other, and they are also working with some of the terrestrial network providers. Liquid is the biggest of them. And Carey, I’m sorry, I’m blanking on the other one. [inaudible] is another Pan-African terrestrial cable provider that is also going to get bandwidth on these cables. And Google and Meta are very quick to say, “No, we are sharing this,” but I think it is a kind of… Well, there’s two things to recognize. It is a kind of sliver. It’s a generous slice, but it by no means creates any meaningful scarcity on what Google and Meta can use going forward for the next decades.
Andrew Blum:
And also, these cables are not in and of themselves business propositions for Google and Meta, they don’t need to make money on these cables. They make money on the eyeballs that the cables connect. And so we end up with, essentially, a new regime after 150 years of telecom, where the infrastructure itself is the lost leader for the advertising that it can support. I suppose you could say that’s similar to television in some way, maybe not entirely new, but it’s a different model. It’s not a point to point model. And so we’re in uncharted territory, except for the fact that we are confronted right in front of our nose with these incredibly dominant, very small number of companies.
And I wonder as well how long Google and Meta will last as intact companies. I increasingly think that the moment is coming when they’ll be broken up. And I had a very interesting exchange with a network engineer at one of these companies who said that he’s doing his work. One of the things that motivates him to do his work is to make sure that when the inevitable comes and this network is broken up, it has a strong security foundation for whoever owns it next. Because that is what the past tells us, that these big telecom networks are built and then they change hands. And so he’s trying to make sure that they are in good enough shape. He can put them in good shape for whatever the future may bring.
Ethan Zuckerman:
Carey, you were suggesting in some ways that maybe the answer in this is part government ownership. One of the reasons that it sounds like you are less worried about M-Pesa and that dominance than you are about Google and Meta getting involved with this, is the notion that the government of Kenya has a substantial stake, and therefore has an interest in terms of national service and consumer rights in making sure these systems function and such. Is that a solution here? Is this solution that there needs to be a partnership with government of Kenya? Or is there a regulatory solution that says, sure, Google, you can have a cable, but we need to make sure that some parts of the fiber are open and neutral, even if you’re using other parts of the fiber purely as your own back haul?
Carey Baraka:
I don’t think I would say that since I like the solution. I think from promotions with the people in the tech scene is that for them, first of all, the more cables, the better. It’s like, the more options there are. Yes, Meta and Google are bringing cables, but also there’s another [inaudible] cable called the PEACE Cable being fronted by a Chinese company. Actually, they’re referred to as something called the digital silk road, or something. The underlying condition by a lot of these people is that as long as there’s as many cables as possible, then the market is insulated from they are being one dominant player. And sometimes such a cable is operated by a central government, but then there’s other cables operated by central players.
So it becomes, what is happening in DRC where Meta are landing in Rwanda, or on their Atlantic? And then they are landing in [inaudible] Mombasa. And they are building a terrestrial cable network running through the DRC, running through the rain forest, running through the Congo River. Where there’s previously been nothing, almost nothing. And the question then becomes, for me the problem becomes that they are the only people who operate this cable. That it’s the Liquid telecom and Meta partnership. Whereas if there were multiple interests or networks, that this DRC company owns part of the network, so the foreign company owns another part of the network. The central government owns another part of the network. So then a lot of fears are mitigated.
Ethan Zuckerman:
Right. The way that the US handled this at a certain point was in breaking up monopolies. They created competitive local exchange. And so the theory was, yes, there’s one cable there, we understand that it’s probably too expensive to lay a second cable. But you have a legal right to purchase bandwidth on that cable and run an alternative operation. So that’s one regulatory strategy that could get there. Carey, part of what you’re suggesting is that in some ways what you really just want is diversity. That maybe it’s not so bad if Meta or Google have cables that they own. And that, as Andrew was putting out, are funded via different economics.
But what’s really dangerous is if it’s the only cable. For instance, the only transcontinental cable going through DRC, which is going to be a very expensive cable to run. And possibly even worse, if it’s the only cable and no one has rights to carriage on top of it. So we’re getting pretty deep into the regulatory weeds. But as you’ve both very helpfully pointed out in the story, this really comes down to how individuals on the continent get bandwidth, what it costs, and what they can do with it.
Carey Baraka:
Let me just add one thing. A part of the case, other telecom providers are file against Safaricom with M-Pesa, is the idea of M-Pesa agents. M-Pesa, I don’t know how familiar you’re with it, but M-Pesa works is their agents all over the country and you can go withdraw or deposit money. Sort of like an ATM, basically. And all these agents are paid. And so a lot of this other telecom providers, the authority is saying that we can’t hire independent agents by ourselves, it’s just too many agents. Why not we also use M-Pesa agents? Which sounds a lot like what was happening with the internet exchanges. And that fight hasn’t been won yet.
Ethan Zuckerman:
Right. Right. If we can’t get M-Pesa agent neutrality, can we in fact get cable neutrality? Andrew and Carey, thank you so much. Andrew Blum and Carey Baraka are authors of a great piece called Sea Change, for Rest of World. It’s a real eye-opener about these really complex politics of who is creating the next generation of connectivity for Sub-Saharan Africa, other parts of the developing world, and the implications that this has associated with it. Andrew and Carey, thanks so much for being with us.
Andrew Blum:
Thanks for having us.
Carey Baraka:
Thanks for having us, it’s been a pleasure.