Google and Meta are now investing fortunes into building massive subsea cables, bringing internet infrastructure to millions, write ANDREW BLUM and CAREY BARAKA.
One sunny Thursday morning in September 2021, three men gathered at the Multimedia University of Kenya, in Nairobi, for a small ceremony. The usually pristine grass of the campus was brown and dry. Jacaranda petals, fallen off the trees but still vibrantly purple, littered the ground, and a troop of baboons rummaged through garbage cans for leftover food. Inside the university’s ICT Museum, which showcases developments in communications technology, the men turned their attention to an object at the center of the room.
Above a table crowded with old oscilloscopes, printers, and telephones, hung a lone box, white apart from two black solar panels fanning out from its top. Gray antennae protruded beneath, trussed like the legs of an oil rig. This jerry-built device was among the last remaining specimens of Loon, what had been Google’s “moonshot” project to connect rural Africa and other locations to the internet, using balloons floating in the stratosphere.
In 2015, when asked which current project he was most excited about, then–Alphabet CEO Larry Page mentioned Loon. “There’s very many places you go in the world where you still don’t have a cell signal. And I think Loon actually could change that,” he said. Page dreamt of giant white balloons semi-autonomously navigating along atmospheric currents for thousands of miles, beaming connectivity down to remote areas or disaster zones. But by January 2021, after nearly a decade of work and hundreds of millions of dollars spent, the project’s lead announced that its “journey [was] coming to an end.” Loon was dead, owing to the costs being too high to “build a long-term, sustainable business.” Instead of connecting the unconnected, Loon would join the artifacts at the ICT museum. (Rest of World was founded by Sophie Schmidt. Ms. Schmidt is related to Eric Schmidt, former CEO of Google.)
For more than a decade, US tech giants have had designs on building Africa’s internet. Alphabet is now at work on Project Taara, another “moonshot,” which aims to repurpose the Loon balloons’ airborne lasers. Meta — previously Facebook — has also floated airborne internet delivery systems, including using a satellite that would beam data to Africa from space (which was abandoned when the rocket carrying it was engulfed in flames on the launchpad) and its Aquila solar-powered drones (which were grounded after disappointing performances, including a crash landing). Elon Musk’s SpaceX seems to have had better luck, having now launched over 1,700 small satellites as part of its Starlink constellation, although it won’t begin providing internet service in Africa to consumers until later in 2023.
But beneath these shiny objects in the sky — laid, in fact, on the ocean floor — are a series of more traditional and likely much more transformative efforts to bridge the connected and the unconnected. After years of anticipation, massive undersea fiber-optic cables, stretching thousands of miles, have begun arriving on African and European shores.
In March, the Ile de Bréhat, a high-tech ship registered in France to Alcatel Submarine Networks, which is under the umbrella of telecomms giant Nokia, spooled out a cable as thick as a paper towel roll toward the beach in Lomé, the capital of Togo. Once it was secured on land, an executive from Google posed for photos alongside a “golden buoy,” marking a key milestone in the company’s ambitious project — known as Equiano — to link Africa and Europe using an undersea link of unprecedented capacity.
Lomé was a prelude. In April, Google and its “landing partner,” West Indian Ocean Cable Company (WIOCC), erected a temporary auditorium on Elegushi Beach, in Lagos, Nigeria. Inside, a crowd packed with traditional rulers, high-ranking government officials, and executives of cloud and data center companies celebrated Equiano’s imminent arrival. (The actual cable landing is still ongoing.) “What we jointly dreamt three years ago is becoming a reality,” said governor of Lagos State, Babajide Sanwo-Olu.
The expected impacts will be substantial. According to research commissioned by Google, the Equiano cable will improve median download speeds in Nigeria by up to six times, reduce retail data prices by 21 per cent, and create economic activity that will indirectly result in US$10 billion added to the Nigerian GDP and 1.6 million new jobs. The intention is to provide more international bandwidth, said Juliet Ehimuan, director of business strategy in West Africa at Google. “The cable leverages state-of-the-art technology to provide approximately 20 times more network capacity than the last cable built to serve the region,” she said. The Equiano cable will make two more landings after Nigeria — first in Swakopmund, Namibia, and later at Melkbosstrand, South Africa — before connecting back to European networks in Sesimbra, Portugal.
Meta, not to be outdone, had its own ceremony in April, 2,600 miles due north, when Alcatel’s Ile de Batz spooled out another cable toward the beach in Genoa, Italy. Meta’s 2Africa cable, announced in 2020, will fully encircle the continent; like Equiano, it will connect in Portugal and head south off Africa’s Atlantic coast, but it will then continue counterclockwise around the Cape of Good Hope, past South Africa, up the Indian Ocean coast, crossing Egypt overland and connecting back to Europe on the Italian, French, and Spanish shores. Meta expects the construction of the cable and its 46 landing segments to be complete in 2024.
“That something that is so profound for public communication is completely controlled by a private company is worrisome.”
These two new pieces of infrastructure will connect Africa to the global internet more robustly than ever, but they will also place an unprecedented level of control in the hands of the US-based tech giants. Google and Meta’s ambitions to build and own global data links mark a tectonic shift in how the internet works and who controls it.
The internet’s initial promise was to decentralize telecommunications, releasing consumers from the monopoly grip of telecomms incumbents. Over the last 30 years, the internet has done that, and much more. But undersea cables, owned by the internet’s behemoths, hint at a return to where we started: a near future in which a select group of massive corporations have not merely tightened their hold on our online activity but have deliberately rebuilt the internet for their own use, according to their own specifications, from the ocean floor up.
“That something that is so profound for public communication is completely controlled by a private company is worrisome,” Steve Song, a policy advisor at the Mozilla Foundation and a longtime advocate for internet access across Africa, told Rest of World.
If the giants’ cable building continues as planned, the future internet will be less a network of interconnected networks, as it was originally conceived and as it has grown, and more like a supranet, dominated by a handful of mega networks operating upon their own global physical infrastructure.
To understand the implications of that shift, Rest of World reported over two years on three continents, conducting more than two dozen interviews with policymakers, telecomms industrialists, data center , and net advocates. Overwhelmingly, people who spoke to Rest of World agreed that the construction of this huge new internet capacity will be a boon for hundreds of millions of internet users — present and future — across the continent. But the secondary consequences of Google and Meta’s ownership are less clear, and more worrying.
For regions with established internet networks, the evolution of tech giants like Google and Facebook into ISPs and internet infrastructure providers is alarming — and sets the stage for antitrust concerns and gladiator battles between industry giants. But in a region where about 33 per cent of the population is online, and connectivity is often limited, Meta and Google’s installation of Equiano and 2Africa suggests a more definitive goal: the lasting control of the global link for 1.4 billion people. In Africa, the least connected continent, the stakes of Google and Meta’s ambitions are stark.
For about 150 years, telecomms companies owned all the world’s telecomms cables, selling the right to transmit a message or make a call. Whether government-backed monopolies or publicly traded corporations, they mostly abstained from meddling in the content of the communication itself, leaving that to newspaper owners and broadcasters — the original “content providers.”
Google, founded as a search engine on a single bank of servers in 1998, was a new kind of company. Its powerful advertising business shook what became known as “legacy” media, and its acquisition of YouTube in 2006 gave it additional hold on our attention. All the while, it rapidly constructed a global physical infrastructure of astonishing breadth.
Google first began investing in the undersea cables that form the backbone of its international communications in 2008, partnering with a consortium of Asian telecomms companies to build Unity, a $300 million link that stretched from California to Japan. Over the next decade, more investments followed, with further cables across the Pacific and other cables connecting Florida and Brazil, and Australia, Singapore, and Indonesia. These cables were laid in partnership with telecomms giants like Singtel, KDDI, and China Telecom.
But in 2018, Google announced that it would be expanding its ambitions: rather than merely partnering with telecomms on new cables, it would plan a series of wholly owned new transoceanic cables, becoming “the first major non-telecom company” to do so, according to its own accounting. No longer merely a “content provider,” much less a search engine, Google would also become the owner and operator of its own global infrastructure, a far-reaching and distinct network effectively parallel to the internet itself.
“Africa is the least served of all continents in terms of capacity, so anything that brings additional infrastructure investment to Africa has to be a good thing.”
Google’s undersea cable-building program was formidable from the start. In a characteristic nod to its corporate self-image as a band of playful geeks, the new cables would soon be named alphabetically after “historical luminaries.” The first, a cable from California to Chile, was named Curie, after the chemist and physicist Marie Curie; a cable from Virginia to France was dubbed Dunant, for Red Cross co-founder Henry Dunant. The “E” cable, stretching from Portugal to South Africa, was named for Olaudah Equiano, an 18th century author and abolitionist. Then there’s Firmina, running from the eastern U.S. to Argentina, and Grace Hopper, stretching from the US to the UK and Spain. While Google won’t confirm any plans, the naming scheme leaves 19 letters left for future cables and plenty of ocean in which to lay them.
Google isn’t the only internet giant putting infrastructure in the water. Amazon, Microsoft, and Meta have all invested in cables across oceans, usually partnering with one another or traditional telecomms. Their motivation in doing so is straightforward enough: the appetite for internet content is exploding globally; the existing cables are insufficient to deliver their products to their users; and, with troves of cash at their disposal, they see no reason to wait for telecomms companies to build new ones. According to the research firm TeleGeography, in 2010, content providers consumed 6.3 per cent of total international cable capacity, leaving most of the rest to telecomms companies and other entities, like research-educational networks and governments. By 2021, that number had leapt to 69 per cent — a figure TeleGeography expects to grow again to 78 per cent by 2027. Implicit in those statistics is how that traffic now mostly travels on cables wholly or partly owned by the content providers themselves. In 2010, Google, Meta, Microsoft, and Amazon had invested in only one long-distance cable; by 2024 they will own all or portions of more than 30.
Not long after Google’s Equiano cable completes its southward run along the Atlantic seabed, Meta will follow with 2Africa. In September 2021, the company announced a further expansion of the cable: 2Africa Pearls, which will branch off near the Horn of Africa and connect India, Pakistan, and nearly all of the Gulf states. When complete, the combined system will be the longest undersea data cable in the world, bringing “seamless international connectivity” to 33 countries home to 3 billion people, or 36 per cent of the world’s total population.
Africa — or at least its coastal countries — has been reasonably connected by undersea links since the tail end of the 2000s, when traditional telecomms and private consortiums built cables like MainOne to Nigeria, or Seacom and Eastern Africa Submarine Cable System (EASSy) to Kenya and South Africa. But with gargantuan capacity enabled by the latest technology, Google’s Equiano and Meta’s 2Africa will make those older cables functionally obsolete — old mini buses in a fleet of shiny coaches.
For Meta, in particular, the value of 2Africa’s new bandwidth appears straightforward: more eyeballs on Facebook, Instagram, and WhatsApp. Africa’s 400 million internet users consume only a fraction of the data that a European or American user might. As more Africans come online and data consumption grows, demand for bandwidth is set to explode. “There will not be enough capacity on the subsea to be able to handle this kind of traffic growth,” director of network investments in Europe, Middle East and Africa at Meta, Ibrahima Ba told Rest of World in a Zoom call from his home in Virginia. “And you can solve it only by a cable like this one.”
In blog posts and videos, both Google and Meta emphasize the societal benefits of their cables, suggesting that they are acting, at least in part, altruistically. Google says the purpose of its cables is to “interconnect humanity”; Meta says 2Africa will ensure “everyone can benefit from the economic, educational, and social advantages of a digitally connected world.”
The reality is obviously more complicated. “No question about it — they’re doing it for selfish motives,” said former CEO of Seacom, a telecommunications venture that owns East Africa’s first broadband submarine cable system, which went online in 2009, Byron Clatterbuck. “But, by doing that, they’re also enabling more internet connectivity in Africa.”
Lead digital development specialist at the World Bank, Tim Kelly says that, despite the profit motive, Google and Meta’s involvement in cables is a net positive. “Africa is the least served of all continents in terms of capacity,” he said, “so anything that brings additional infrastructure investment to Africa has to be a good thing.”
But owning the infrastructure gives the tech companies more control. Meta is partnering with seven telecomms companies to build its cable, in what’s known in the industry as a “club,” and promises that bandwidth will be available “on a fair and equitable basis.” Google is more abstruse in its collaborations, publicly relying on partners for the landing segments in each country, while maintaining possession over the “deepwater” segments. Owning Equiano — like owning all of its undersea cables — gives Google technical and operational autonomy. It can do what it wants, where it wants, and when it wants.
“True, we will get connected, but the question is, what kinds of public policy will keep that power in check?”
“Since we control the design and construction process, we can fully define the cable’s technical specifications, streamline deployment and deliver service to users and customers faster,” explained Google’s vice president of engineering, Ben Treynor Sloss, in a 2018 blog post. Google can plan its own underwater routes, serving the destinations in greatest need of additional bandwidth. Google can also draw on its colossal capital reserves — Alphabet held nearly US$150 billion in cash in the third quarter of 2021, more than the GDP of all but four African nations — to overbuild capacity, allowing room for future growth. And it can connect markets ahead of any clear immediate demand — an especially powerful lever in Africa. Once the cable is up and running, Google’s network managers can prioritize Google’s own data traffic, ensuring the performance and reliability of their cloud services — while using that reliability as the services’ key selling point.
In a rapidly digitizing continent, Google and Meta’s plans worry Nanjira Sambuli, a tech/ICT policy analyst based in Nairobi who was a member of the influential UN Secretary-General’s High-Level Panel on Digital Cooperation. While she recognizes the value, infrastructure-wise, of these companies’ involvement in cable building in Africa, she worries about the implications. “True, we will get connected,” she said. “But the question is, what kinds of public policy will keep that power in check?”
As more services migrate online, the stakes rise. The Equiano and 2Africa cables will not only carry social media and internet searches (and the ads that help monetize them) but also entire countries’ health, governance, and security data. In the absence — for now — of large African data centers to store it regionally, these digital services remain offshore. For Sambuli, the key need is for African governments to put forward policies that protect their citizens under the thumb of internet giants with a history of putting profits before society.
“It’s one thing if Facebook is answerable to a country in terms of social media, but it’s another if they’re also controlling infrastructure,” she said. If they control the cables, internet companies could effectively dictate internet policy, even if that is not their explicit intent. Meta has confronted similar challenges with its Free Basics and Discover programs, which have been successful in bringing millions of people online but with content limitations that have drawn accusations of “digital colonialism” and being “poor internet for poor people.”
“The internet, left unchecked, is a monopoly-making machine, an engine designed to concentrate power, attention, and more in the hands of those who already have it,” observed James Ball in his 2020 book, The System.
Much depends on how Google and Meta exploit their control. “We don’t know all of the implications of what a proprietary cable would mean,” said the World Bank’s Kelly. “Maybe Facebook will do no evil — or is it Google that does no evil?” But ultimately, he added, “they make their money through advertising, and they make their advertising by control of data, and therefore, one of the reasons that incentivizes them to own their own networks is because they can suck up every single bit of data from those networks.”
Notably, Google and Meta have no intention of making a profit by selling the bandwidth itself. For its part, since Google is not a telecomms company, it has no plans to displace telecomms in the value chain. “[Equiano] provides international bandwidth, but to get to the end user, you still need metro access; you need to cater for the last mile, which is the space that a lot of telcos and service providers play [in],” Google’s Ehimuan explained to Rest of World. “The intention is to collaborate with them.”
CEO of Pure Infrastructure Limited, a cloud consulting company in Nairobi, Phares Kariuki sees Google and Meta’s cable building as a longer-term effort at commercial control. “They started out because the internet was open, and now they are quickly creating a moat by closing it up and ensuring no one else can be as competitive as them,” he said. As a relatively small-scale internet entrepreneur, he has little patience for the way the giants couch their efforts in claims of do-gooderism. “Just say that it’s your own private infrastructure, don’t sort of say that it’s to help infrastructure on the continent,” Kariuki says. “This is Google helping Google get more Google content on the continent.”
In 2010, Google, Meta, Microsoft, and Amazon collectively owned only one long-distance cable; by 2024 they will own all or portions of more than 30
Google’s Equiano cable will run from Portugal along the Atlantic coast of Africa to the small town of Melkbosstrand in South Africa
Meta’s 2Africa cable will also connect in Portugal before fully encircling the African continent, crossing Egypt and connecting back to Europe
Mombasa, Kenya, normally a cheery tourist haunt, was cloudy, dour, and gray on the morning Rest of World visited in September 2021. A few years back, the Mombasa governor directed all the buildings in the city’s central business district to be painted in the same hues of blue and white. One such building, built in the old Swahili architectural style, serves a very contemporary purpose: it is the landing station for Seacom.
From inside Seacom’s boardroom, a lone boatman was visible in the ocean, rowing, and behind him, a lighthouse. It felt appropriate: the Seacom landing station is, after all, a different kind of lighthouse — one that pulses invisible flickers through the hair-thin strands of fiber-optic cables.
To reach here, the Seacom cable snaked its way underwater from South Africa, veering North, touching Mozambique and Tanzania, before landing in Mombasa. After Mombasa, it went around the Horn of Africa, splitting to Djibouti and India, before crossing Egypt and ending up in Marseille in France. The cable first arrived in Mombasa on a July morning, a little more than a decade ago. From the ship, it was brought ashore and one end was affixed through a conduit under the beach and into a manhole that connects to this building — a process known as “landing” the cable.
In a room inside the landing station hum the machines that Kenya’s internet users depend upon: generators, fans, racks of servers, and line-terminating equipment. A sample of undersea cable was kept on hand for visitors; when Rest of World lifted it, it was surprisingly light.
Seacom was the first of many cables to land in Kenya: the East African Marine Cable System (TEAMS) also landed in 2009, followed by EASSy in 2010, the Lower Indian Ocean NetWork II (LION2) in 2012, and the Djibouti Africa Regional Express (DARE1) in 2021. But the arrival of Equiano on the Atlantic coast this year and of 2Africa right here in Mombasa, likely in 2023, will mean a more substantial shift — in both the business of telecommunications and the politics that pervade the industry.
“Subsea,” as it’s known, is traditionally a challenging business, with multi-hundred-million-dollar capital costs, followed by a relentless need for operational upgrades. Customers always demand more bandwidth, at lower prices. The more the customer base grows, the more you have to upgrade your system, eating into any growth in profits. Currently, Seacom and its competitors vie with one another to profitably sell bandwidth to African ISPs, international content providers, and farther-reaching global telecomms networks.
From the perspective of the terrestrial network owners, that effort can feel crushing. “Historically, one company has controlled [the cable], and it can charge an extortionate amount of money just to connect between the terrestrial network and the submarine fiber,” CEO of WIOCC, which is a partner on both Equiano and 2Africa and has investments in numerous existing African cables, Chris Wood told Rest of World.
But Equiano and 2Africa, bought and paid for with the titanic profits of Meta and Google, derived primarily from their advertising businesses, won’t have to be profitable at all. Clatterbuck, the former Seacom CEO, saw what that meant for his business: obsolescence. “Do I bemoan the Yellow Pages not getting delivered every year?” he told Rest of World in June 2020, one month after Meta’s 2Africa announcement and before his resignation as CEO in March 2021. “Do I bemoan the Encyclopedia Britannica? We used to buy it every three years. I don’t need it anymore. I think we move on, right?”
Facebook and Meta are offering traditional telecomms an alternative business model: the opportunity to play middleman between their cables and inland users who will continue to depend on terrestrial cables deep in the continent. In exchange for bearing an unspecified bulk of the capital costs, Meta, for example, is getting direct links to the national and regional networks that connect the new international cables to national eyeballs.
Ibrahima Ba, in pushing the cable project for Meta, tunes his pitch to appeal to the regional telecomms. “Do you want to make money on the subsea side?” he asks his local partners, “Or do you want to invest into the access and the retail side?” Ba suggests the future rests with the latter option. “A lot of operators now recognize that the investment needs to move on the retail side; that’s really where the growth is; that’s where the differentiation is,” Ba said.
Rather than pursue the traditional model of creating a consortium or “club,” Google is instead selectively partnering with telecomms only for the “branching units” that connect the main ocean trunk of the cable to the spurs that land in each country. This allows Google to avoid running the regulatory maze of being a “landing party,” which is often subject to strict country-specific regulations and keeps most of the company’s infrastructure investment in risk-free international waters.
Google can then mint its own kind of currency: it can trade the Equiano cable’s international bandwidth for domestic terrestrial bandwidth across Africa, and the complicated business of negotiating the construction, ownership, and regulatory challenges of fragmented assets in numerous countries that comes with it.
In Nigeria, for example, WIOCC is the “landing party” for the Equiano cable — and it has also “acquired” a fiber pair on the cable. Each strand of fiber carries pulses of information-carrying light transmitted in one direction, so they always work in pairs; WIOCC’s pair is just one of Equiano’s 12. When “lit” with the latest and greatest fiber-optic equipment, the pair’s 12-terabit-per-second capacity dwarfs the bandwidth of all the pairs on existing undersea cable systems — and gives WIOCC the international bandwidth it needs. “We will connect to our national network in Nigeria, and then we’ll sell services on that national network which would use our fiber pair to get to the rest of the world,” CEO Chris Wood told Rest of World. “So that’s how each company would then monetize its asset, its investment.”
Liquid Intelligent Technologies, a digital infrastructure company that operates cables and data centers across Africa, is pursuing a similar deal with Meta. In March 2022, the company announced that it has also “acquired” a fiber pair on Google’s Equiano cable. “We are working with Meta and Google and bringing internet cables inland,” said Ben Roberts, its chief technology and innovation officer. In July 2021, Liquid announced its plans to build a 2,000 kilometer long-haul and metro fiber-optic cable network in the Democratic Republic of the Congo. The cable will start at Muanda, where the 2Africa cable is slated to land, and run to the center of the country. Once completed, it will be a part of a bigger network extending to Congo-Brazzaville, Angola, Rwanda, Uganda, Tanzania, and Zambia. It will cross the African continent terrestrially, from the Indian Ocean to the Atlantic Ocean, “from Mombasa to Kinshasa,” as Roberts put it. “If you look at the dots we are trying to join up, we are having a terrestrial network, and we are trying to bring the networks inland, because if it’s just at the landing station, it’s no help to anyone,” he said.
With their new international bandwidth, both WIOCC and Liquid will have the capacity they need to feed their hungry terrestrial networks. In the old telecomms regime, handing over this much capacity to a competitor would be commercially foolish. But Google and Meta’s cables needn’t be profitable on their own but, rather, merely serve a far broader strategy — one that highly values the terrestrial links to a billion “eyeballs.” If the old undersea cables operated by telecomms see Meta and Google’s giant checkbook as a death warrant, the African terrestrial networks are eagerly embracing the symbiotic relationship they can have with Google and Meta.
Liquid has been upgrading its existing routes in anticipation of the new cables. Roberts said that, in less than two years, “we’re going to be having 5G networks.” These networks, invariably, will be key to the metaverse or whatever vision of the internet Meta, Google, and the other tech behemoths have in mind for Africa. What remains to be determined is how that meshes with the visions Africa’s internet builders have for themselves.
Since 2008, Michuki Mwangi has been regional development manager for Africa at the Internet Society, an international nonprofit founded in 1992 to promote the growth of the internet. In 2020, he wanted to measure and compare the total amount of data center floor space in Africa (a reasonable approximation of hosting capacity). The number he came up with — 80,000 square meters — astonished him. The U.K. alone had 500,000 square meters. “And I was like, ‘For a billion people, we have less than 100,000?’” Mwangi said.
The dearth of data center space was partly a reflection of low internet usage, but it also highlighted a separate issue: Africa’s internet was disproportionately imported. When a user in London navigated to a web page or service, in all likelihood that data was stored in a data center — a kind of warehouse — close by. But in Nairobi, and even more so in African cities further inland, such as Kampala or Lilongwe, the lack of local or even regional data centers meant that any request had to travel thousands of miles across undersea cables. As a result, performance suffered. Even at the speed of light, the journey took time — and it was expensive.
In 2008, when Mwangi started at the Internet Society, 99 per cent of African data traffic came from outside the continent, hosted on servers in Europe, America, and Asia. In coastal countries like Morocco, Egypt, and Tunisia, where undersea fiber-optic connections to Europe were available, he recalls that internet transit cost an average of US$600 per megabit per second per month — 60 times what it would cost in New York or London. In landlocked countries like Uganda, Zambia, and Botswana, where satellite was the only option, it was more than twice that, upward of US$1,500 per megabit. Those were wholesale costs; internet service providers still had to transport those bits to users’ homes and offices.
Since then, as the number of data centers in Africa has increased, more data is being hosted on servers locally, and more data is being exchanged among African networks, rather than traveling back and forth — “tromboning,” in internet jargon — to London, Frankfurt, or Marseille. This is primarily accomplished through the creation of internet “exchange points,” which allow networks to connect directly to each other. Encouraging their growth has been a major aim of the Internet Society, which in 2010 had set the goal — the “grand vision,” Mwangi called it — of making 80 per cent of internet content locally accessible in Africa by 2020.
They came close. According to its 2021 report, nearly 70 per cent of traffic in both Kenya and Nigeria was localized. Partly owing to the falling costs that localization engendered, internet use in Kenya rose from 8.8 per cent of the population in 2012 to 17.8 per cent in 2017; in Nigeria it went from 16 per cent in 2012 to 42 per cent in 2020. “You still have a population over a billion people that would need to be supported,” said Mwangi. “We haven’t scratched the surface in Africa.”
What must come next in Africa is data centers. The new cables from Meta and Google will be “fat pipes” to the rest of the global internet, offering an abundance of bandwidth that would have been difficult to imagine in 2008. But perhaps counterintuitively, they are also expected to accelerate the shift to locally hosted data. A node on a network becomes more useful with more connections, a truism that becomes very practical for companies deciding where to store their digital assets. Multinational players can be assured of robust connections back to their existing hubs; regional operators can feed off of the beefed-up presence of those multinationals. “The moment you put data centers and low-cost international submarine cable capacity, then a lot of things start becoming more possible,” said Mwangi.
“We haven’t scratched the surface in Africa.”
Even before the new cables have arrived, there has been a significant boom in the number of data centers under construction. Icolo.io, which operates two data centers in Kenya — one in Nairobi and one in Mombasa — is doubling its capacity in the country, building a new one in each city and a data center in Mozambique. Liquid’s Africa Data Centres division operates nine locations in Nigeria, Togo, South Africa, and Kenya and has plans to build 10 new facilities in 10 African countries over the next two years. Raxio, which operates a facility in Uganda, is planning new buildings in markets like Ethiopia, Mozambique, the Democratic Republic of Congo, Tanzania, and Côte d’Ivoire. Pan African Internet Exchange Data Centres (PAIX), which currently operates data centers in Kenya and Ghana, has plans to roll out to five other African countries. In November 2021, WIOCC announced plans to build more than 20 data centers across the continent, including one in Mogadishu, expected to be complete in 2022.
These companies are following the cables and looking to build the facilities they expect the international giants to want, as they increase their data center footprints to serve their newly burgeoning user bases.
During a visit to PAIX’s Nairobi data center, Vincent Camadro, PAIX Kenya’s managing director, laid out his company’s strategy. In the US, he said, Google and Meta build their own data centers; “in Africa, they’re going to rely on people like us to host them.”
But just as with undersea cables, local and regional ownership is unlikely to last forever. In April 2022, Equinix, a California-based digital infrastructure behemoth with Fortune 500 status, finalized its US$320 million acquisition of Nigerian-headquartered MainOne, which has data centers in Nigeria, Ghana, and Côte d’Ivoire. The Equinix vice president for growth and emerging markets, Judith Gardiner saw plenty of room for further expansion. MainOne’s land alone, she noted, was enough to build another ten data centers. Notably, Equinix’s deal for MainOne came less than a year after it announced plans to build a data center in Genoa, Italy, that it said would “serve as a strategic gateway for the 2Africa subsea cable system.” The internet may be a network of networks — but it has also proven to be, in James Ball’s phrase, “a monopoly-making machine,” with power often residing outside of Africa.
Early in the pandemic, at a moment when the digital transformations it had wrought were still jagged, the Internet Society hosted a webinar to celebrate its localization report and the progress that had been made in keeping African traffic local. Google had been a significant contributor to that long-term effort, investing in the network beachheads that serve its users, and Sylvie LaPerrière, a network strategist at Google and a leader of their Equiano effort, was on the panel. When, toward the end of the lunchtime session, Mwangi asked what can be done to further improve capacity, LaPerrière laid out Google’s ambition with unusual candor.
“I think we need stronger infrastructure,” she said. “We need more data centers, more intra-network in the data centers. We need more internet exchanges. We need more content platforms. More cloud operators. We need more of everything. In 15 years Africa will be 20 per cent of the world population. Now’s the time to really make a big bet.”
Back on Elegushi Beach, in Lagos, nearly two years later, Google’s bold intentions had become physically apparent. As the launch event for Equiano wound down inside the temporary auditorium, Rest of World walked down to the ocean’s edge. The Ile de Bréhat, carrying the cable, had not yet appeared on the horizon. But just above the beach, engineers and contractors were putting the finishing touches on a three-story complex belonging to Open Access Data Centres (OADC), which operates under the WIOCC umbrella. It will support the terminating equipment for Equiano as well as a bloom of new servers installed to feed off of its prodigious bandwidth. The estimated US$200 million facility, with an expected 20 megawatts of available power and room for 3,200 racks of servers, was merely OADC’s starting point; the company has publicly committed to US$500 million in further data center investment across Africa.
“It will change completely the digital economy and the digital transformation of not only Lagos but indeed Nigeria,” Sanwo-Olu, told those attending the launch event.
This stretch of sand is owned by the Lagosian royal family — the Elegushis. Behind a security fence, families picnicked and played in the surf. Beneath them, silent but bright, Google’s cable would soon change their connection to the world.
