Nigeria and West Africa’s digital transformation stands at a pivotal crossroads. Nigeria and the broader region are experiencing considerable upward growth in connectivity, data consumption, and digital adoption, yet persistent challenges around reliability, cost, power supply, foreign exchange risks, and data sovereignty threaten to undermine these gains. As CEO of TelCables Nigeria, a subsidiary of the Angola Cables global network, I see the immense opportunities, and the urgent need for localised, resilient solutions that put African businesses and governments in control of their digital futures.

Investments by telcos like MTN’s FiberX and WIOCC’s Open Access Metro sit alongside government initiatives such as Project Bridge by the Ministry of Communications, Innovation and Digital Economy, a US$2 billion investment into 90,000 km of terrestrial last-mile infrastructure expanding from Points of Presence (PoPs), hubs, cable landing stations (CLS), and data centres for businesses across the region. This is critical. According to statistics recently released by the Nigerian Communications Commission (NCC), Nigeria now has around 192 million mobile subscriptions and broadband penetration climbing beyond 55% in recent years. Alongside surging data usage (averaging nearly 10 GB per user monthly), reliable backbone connectivity is no longer optional, it is a national imperative for the country’s economic diversification beyond oil.

Cloud adoption critical for social and economic development of the region

Cloud computing adoption is accelerating, particularly in the financial services industry (FSI), which is encouraging businesses, both public and private, to leverage IaaS, backup, disaster recovery, and containerised solutions like Kubernetes, within their enterprise environments. However, the heavy reliance on international hyperscalers such as AWS and Azure introduces substantial financial volatility, foreign currency challenges, and compliance risks under Nigeria’s Data Protection Act (NDPA).

Data hosted offshore creates latency, unpredictable egress costs, and jurisdictional vulnerabilities that clash with national priorities around sovereignty and security. To counter this, and bring stability and security, the Central Bank of Nigeria has mandated that all banks, fintechs, and payment service providers store and manage all local payment transaction data within Nigeria by 1 January 2027.

This shift in policy is a catalyst for local innovation to come to the fore. TelCables Nigeria’s cloud compute solution, Clouds2Africa addresses these gaps directly: featuring two nodes in Lagos, billing in NGN to mitigate currency risks, and offering a full suite of enterprise services including IaaS, secure backups, and disaster recovery.

Crucially, Clouds2Africa is fully AWS-compatible, enabling seamless integration and effortless transition for businesses already familiar with public cloud architecture. To ensure a frictionless journey, TelCables offers a dedicated secure workload migration program to assist these enterprises in migrating workloads from diverse, heterogeneous environments to a localised data solution, safely and without disruptive downtime.

Furthermore, leveraging Angola Cables’ massive global backbone capacity and extensive submarine cable network (SACS, WACS, MONET), Clouds2Africa eliminates one of the biggest hidden operational costs of cloud adoption by offering zero egress fees for most of the use cases. Coupled to their strategic PoPs and partnerships with tier-three local data centres such as Rack Centre and Medallion, TelCables can deliver ultra-low latency, high-resilience connectivity that keeps Lagos’ digital core connected to the world while maintaining critical workloads securely within regional borders.

The path forward demands collaboration

Yes, challenges persist: power reliability, last-mile gaps in underserved areas, and the continuous need for technical skills development. If we are to succeed in overcoming these hurdles, both government and industry must collaborate to support initiatives that accelerate the adoption of local infrastructure and secure Nigeria’s digital future.

Industry players, regulators (NCC, NITDA), and government must prioritise open-access models, incentives for local data centres, and public-private partnerships like Project Bridge. At TelCables, we remain “Africa’s Most Interconnected Operator,” investing in terabit-scale rings interconnecting major Lagos data centres and continually expanding our footprint.

Our goal is clear: as a business and solutions provider, we want to assist in promoting the widespread adoption of capable, local cloud infrastructure that delivers secure performance, affordability, and control. By localising traffic, eliminating transfer costs, securing data, and building resilient networks, we can help unlock productivity, create jobs, foster innovation, and position West Africa as a competitive digital hub within the global geography.

The partnership was announced from Dubai, United Arab Emirates, in a release dated 14 September 2026. Yadea will supply electric two-wheelers and related products for Spiro’s markets; Spiro will integrate them into its swapping and energy infrastructure; and the 2 companies say they will co-develop two-wheeler platforms engineered for African road conditions and commercial use.

Yadea says it has sold more than 100 million vehicles in over 100 countries, holds more than 2,000 patents in electric-vehicle technology and runs 10 production facilities.

Chinese capital in June, Chinese hardware in September

The release ties the deal to Spiro’s latest funding, and the sequence is worth setting out. In June we reported that Spiro closed a $270 million round with a fresh $55 million from NewTrails Capital, a growth-stage fund based in Shanghai and Shenzhen with an office in Nigeria.

Three months later, the supply agreement is with a Chinese manufacturer. The release says the partnership “deepens the China-Africa connection”, which is accurate as far as it goes. It also means the capital and the hardware now come from the same direction.

The question about the factories

Spiro operates in 7 countries: Kenya, Rwanda, Uganda, Togo, Benin, Nigeria and Cameroon. It has manufacturing plants in Kenya, Rwanda and Uganda and a battery-recycling facility in Nigeria, and says it has deployed 100,000 electric vehicles across 2,500 swap stations.

So a company that already assembles vehicles on the continent has agreed to buy them in. The release does not say where the units bound for Spiro will be built, nor whether they will supplement local output, replace it, or be assembled under licence at the existing plants. Yadea says it runs 10 production facilities globally and does not say where they are.

That distinction decides what the deal means. Licensed local assembly of Yadea platforms would deepen African manufacturing; imported finished units would substitute for it. Both are commercially rational and they point in opposite directions for industrial policy, which is why the continent’s battery-manufacturing push is the relevant backdrop.

What was not disclosed

The release did not disclose purchase volume, contract value, timeline, or a list of markets, and it did not name which of Spiro’s 7 countries will receive Yadea vehicles first. The 3 “main areas of collaboration” it refers to are supply, integration, and co-development, as described above.

Anant Badjatya, Spiro’s chief executive, said the partnership “helps us meet that demand at scale”. Gagan Gupta, the company’s founder and chairman of Equitane, called it “a major endorsement of our execution to date”. Wang Jiazhong, senior vice president of Yadea Technology Group, said Africa “represents a massive frontier for zero-emission transport”.

The disclosure worth waiting for is a unit number attached to a plant. Until then, the deal is a supply agreement with an open question about where the supplying happens.

The government release states that “the legal authority for selling the service should rest with ORTEL, under the supervision of the State Secretariat for Artificial Intelligence”. APO Group distributed it on behalf of the government’s official web page.

The seller is the regulator

ORTEL is the Órgano Regulador de las Telecomunicaciones, Equatorial Guinea’s telecommunications sector regulator. The state-owned operator is GETESA. So the entity being given legal authority to sell the service is the one that supervises the market it would be selling into.

The release does not address that, and the wording reaching us is an English translation of a Spanish-language government statement, so it is worth reading it as published rather than inferring intent. The direction is clear: one designated channel, with the regulator named in it.

The supervising body is the State Secretariat for Artificial Intelligence, which is an unusual home for satellite broadband distribution.

Existing customers lose their contracts

The session also ordered the cancellation of contracts individuals had acquired outside official channels, a specific bank account to control Starlink’s income and service payments, and the training of Equatorial Guinean technicians to install and maintain equipment.

That first order is the one that immediately affects people. It implies a population already using the service, which is possible before an official launch because Starlink sells regional and roaming plans that work across borders. Those customers should now be moved to the state channel.

The country is not yet a live market

Equatorial Guinea sits in the coming in 2026 group on our Starlink availability tracker, not among the 26 African countries where the service is live. A government is therefore setting the terms of distribution for a service that has not formally launched there.

Nothing in the release gives a price, a licensing fee, an availability date, contract terms or an implementation timeline. Without those, the practical question for a household in Malabo, Equatorial Guinea, is unanswered: what the state channel will charge, and when.

The pattern is worth watching beyond one country. Satellite internet reaches consumers without the local infrastructure a government can license, tax or switch off, and states that want a hand on it have to intervene at the point of sale. Equatorial Guinea has chosen a direct form of that intervention, short of blocking the service outright, and readers weighing the alternatives to Starlink in Central Africa should expect the question to recur.

Chief executive Mike Ogbalu III gave the numbers at a media briefing in Lagos, Nigeria. PAPSS is a platform of Afreximbank, the African Export-Import Bank, which distributed the figures on its behalf.

A 1,000% rise means roughly 11 times as many payments; a 120% rise means roughly 2.2 times the value. Divide one by the other and the typical PAPSS payment is now about a fifth of the size it was.

Nigeria, which PAPSS singles out as a significant contributor, shows the same pattern, slightly sharper: volumes up about 1,100% and values up about 125%, an average payment roughly 81% smaller.

Why smaller payments are the good news

A settlement system that moves a few very large sums between central banks is plumbing. One that moves many small sums is infrastructure people actually use. A collapsing average transaction size, on rising totals, is what the shift from the first to the second looks like in the numbers.

That reading fits what PAPSS says it will do next. From 2027, the stated focus moves from building the network to “activating” it, through banks, fintechs and switches, and into “the channels businesses and individuals use every day”.

The network, and how it has moved since July

PAPSS now operates in more than 30 African countries across all 5 regions, connecting 24 national and regional central banks, more than 200 commercial banks and payment service providers, and 16 switches. Partnerships extend a termination footprint to more than 300 financial institutions. About 10 countries joined during 2026.

When we reported the Bank of Central African States joining in July, the figures were 28 countries, more than 190 commercial banks and fintechs, 16 switches and a 250-institution termination reach. So in the 2 months since, they’ve added a couple of countries, about 10 banks, and 50 reachable institutions, with the switch count unchanged.

That is steady rather than dramatic, which is worth saying plainly given the headline percentages.

What the percentages do not say

No absolute figures were given for either volumes or values, and “comparable periods” is not defined. A 1,000% increase is arithmetically the same whether the base was 10 payments or 10 million, and without the base neither the growth nor the shrinking average can be checked.

PAPSS also claims cost savings of 92% to 95% per transaction, a 99.99% reduction in processing time and up to 80% lower foreign-exchange requirements. Those are PAPSS’s own figures against an unstated comparator, and we have not verified them.

The platform runs 3 products: the PAPSS Instant Payment System, the African Currency Marketplace and PAPSSCARD. Further solutions are in pilot, with announcements promised later this year. The next phase is due to be set out at the PAPSS COWRY conference in Addis Ababa, Ethiopia, on 26 and 27 November 2026, co-hosted with the National Bank of Ethiopia.

The number worth watching is not the growth rate. It is whether the average payment keeps falling, because that measures whether a system built for central banks is reaching the businesses the continent’s payment rails have mostly not served.

Nigeria Data Centre 1, or NDC1, was announced at ITW Africa in Nairobi, Kenya on 8 September 2026. Digital Parks Africa describes itself as the data centre business associated with the Master Power Technologies Group, whose presence in Nigeria dates to 2001, and presents NDC1 as an extension from power and technical work into data centre services.

Menno Parsons, founder and chief executive, said the investment “is much more than a new facility” and reflects “our longstanding commitment to the region”. The release names financial institutions, banks, cloud providers and connectivity providers as the intended customers.

What the announcement does not contain

The announcement includes no figure for IT load in megawatts, no rack or floor-space count, no capital cost, no named site within Lagos, no completion date, and no carrier named on a facility whose selling point is carrier neutrality. On the numbers, it is an announcement of intent.

That makes 4 in a quarter. MTN’s data centre joint venture in August disclosed no capital, capacity or timeline.

A South African consortium including Digital Parks Africa claimed 400 NVIDIA GPUs made Africa’s most powerful AI cloud, a claim that did not survive examination. Vodafone and Cassava announced Egypt’s first national AI factory this week without a GPU count, a site or a date.

The Tier IV claim, and what it actually says

The release says Digital Parks Africa “engaged with Uptime Institute for the Tier IV Certification of Constructed facility assessment”, marked by a ceremonial signing with Tarik Bahanniss, managing director and vice president for Africa at Uptime Institute.

Read carefully: that is an engagement for an assessment, not an award. The distinction matters more than it looks, because Uptime issues 2 separate certifications, and most projects stop in the gap between them.

A Tier Certification of Design Documents assesses drawings. A Tier Certification of Constructed Facility assesses the building after it is built, under live demonstration.

Uptime’s published award list for Nigeria shows 15 facilities holding design certificates and only 5 that have completed the constructed-facility step. Exactly one of those 5 is Tier IV: Galaxy Backbone’s data centre in Kano, a government facility.

No data centre in Lagos holds a Tier IV Constructed Facility certification. The commercial colocation sites that have completed constructed certification there, including Rack Centre, MainOne’s Lekki facility and MTN’s Dabengwa data centre, are Tier III.

The company’s own record

Digital Parks Africa’s existing facility at Samrand, north of Johannesburg, appears on Uptime’s South African list with a Tier III Certification of Design Documents. We reported it as a Tier III site when it opened in 2023. It does not appear on the constructed-facility list at any tier.

So the company is pursuing the harder certification, one tier above anything it has certified before, in a market it has not built in. That is an ambitious target rather than a dishonest one, and the release does not claim otherwise. But a reader skimming it will take away the words Tier IV, and the accurate reading today is that an assessment has been commissioned.

The disclosures that would make this assessable are the ordinary ones: a site, an IT load in megawatts, a commissioning date, and in due course whether the constructed-facility certification is awarded. Until then, the useful fact is the one Uptime’s own list supplies: in 25 years of Nigerian data centre building, one facility has achieved what NDC1 is aiming at.

The project targets “an initial operational capacity of 20 MW within its first three years”, the partners said, “on a clear expansion path toward 200 MW”. Phase one alone “is expected to draw approximately USD 200 million in foreign direct investment”, with “total investment reaching USD 1 billion once full capacity is realized”.

The signing took place at the Grand Egyptian Museum in the presence of Raafat Hendy, Egypt’s minister of communications and information technology. The partners project up to 200 direct jobs and around 2,000 indirect jobs.

Three companies, three contributions

Cassava Technologies supplies the AI layer and is NVIDIA’s first cloud partner in Africa, selling accelerated computing as a service. Vodafone Business brings the enterprise customer base and Egyptian market position. Elsewedy Electric brings what the partners describe as “expertise in energy and large-scale project execution”.

That third name is the one worth noting. Elsewedy is an Egyptian electrical and infrastructure group, and its presence signals that the partners have treated power and construction as a first-order problem rather than an afterthought.

Why power is the real constraint

A 20 MW facility rising to 200 MW is a serious commitment by African standards, and it is the electricity rather than the silicon that decides whether it happens. South Africa’s own boom has become a transmission story, with operators commissioning their own solar because the grid cannot carry the load.

Egypt’s grid is in better condition than South Africa’s, and bringing an energy contractor into the shareholding rather than hiring one is a more credible answer to that problem than most announcements offer.

What is still missing

Two things are absent. There is no GPU count, which for a facility whose selling point is AI capacity is the number that describes what it can actually do. And no site has been named beyond the country, so the facility cannot yet be placed against a substation, a cable landing or an existing campus.

Compute built in-country matters for reasons beyond capacity. Latency to a European GPU cluster is survivable for training and awkward for inference, and data-residency rules increasingly make the location of the hardware a legal question rather than an engineering one.

Egypt has subsea landings on both the Mediterranean and the Red Sea, a large domestic market and a government pushing digital sovereignty.

The disclosures worth waiting for are the site, the GPU count and a commissioning date.

Correction, 10 September 2026: This article originally reported that the announcement gave no capacity, no investment figure and no employment figures, and grouped it with recent African AI announcements made without numbers. That was based on the wire release we drafted from. The partners’ own announcement does state capacity of 20 MW rising to 200 MW, phase-one investment of approximately USD 200 million and USD 1 billion at full capacity, and projected employment, and it names Elsewedy Electric as a third partner. The article has been rewritten to carry those figures. The absence of a GPU count and of a named site within Egypt, noted in the original, still stands.

Ericsson said on 3 September 2026 that its platform carried about $80 billion in transactions a month and served 131 million people who had transacted in the previous 90 days, as it agreed to pre-integrate payments and card issuing from FIS into the stack. Neither company gave a regional split, and nothing in the announcement referred to Africa.

Asked directly, Ericsson said the figures “relate primarily to the Middle East and Africa regions as this is where most of Ericsson Fintech Platform’s install base exists”, and that “the African continent accounts for approximately about 70-75 percent of the numbers cited”.

Taken at face value, that is 92 million to 98 million African users transacting in a 90-day window, and something in the order of $52 billion moving each month.

Whose volume this actually is

Ericsson also named platform customers across the Middle East and Africa for the first time: Tmcel in Mozambique and MTN Group across multiple African markets, alongside e& Money in the United Arab Emirates, Mobily Pay in Saudi Arabia and Easypaisa Digital Bank in Pakistan.

That list changes how the headline number should be read. This is white-label infrastructure, so the African volume is not a separate market Ericsson has built. It is largely the mobile money business of operators readers already know, running on plumbing they do not see.

MTN alone reports mobile money users in the tens of millions.

A figure that does not reconcile cleanly

The GSMA put global mobile money at more than $2 trillion in transaction value in 2025, an average near $167 billion a month across every market where it operates. An African share of about $52 billion a month would still mean roughly a third of all global mobile money value passing through one vendor’s platform on one continent.

That is possible, but the 2 figures count different things, which Ericsson has since confirmed. Asked how its numbers relate to the GSMA’s, the company said the “definition and categorization of ‘mobile money’ can exclude certain providers who use the Ericsson platform”, so “a direct 100% co-relation between the numbers stated by GSMA and what the Ericsson technology processes cannot be made”.

A platform ledger also records movements that a mobile money industry survey does not. The comparison is offered here as a reason to treat the $80 billion as a platform metric rather than a market one, not as a contradiction.

Still no launch market

On the question that mattered most in the original announcement, whether any African operator or market will carry the bundled FIS stack, the answer has not moved. Ericsson said it is “the current aim of the Ericsson-FIS collaboration to develop and offer a partnership platform for commercial launch which would be announced in due course”.

So the partnership remains unannounced in any specific market, and the case for it is still the one the companies make in the abstract: that pre-integrating wallets, ledgers, identity and card issuing, the layers cross-border tie-ups like M-Pesa and PayPal in Tanzania each had to solve separately, shortens the path from a financial-services idea to a working product for operators, retailers, healthcare providers and government.

What has changed is the scale of what is already there. An install base that is three-quarters African is a stronger argument for the continent mattering to this partnership than anything in the original release, which did not mention Africa at all.

Correction, 15 September 2026: This article originally reported that Africa accounts for roughly 70% to 75% of both the users and the transaction value on the Ericsson Fintech Platform, and derived from that a figure of $56 billion to $60 billion a month. Ericsson has since told tech.africa that the 70% to 75% range applies to the 131 million users, while the African share of transaction value is approximately 65%, which puts the monthly figure at about $52 billion. The article has been updated, and now also carries Ericsson’s explanation of why its numbers cannot be reconciled directly with the GSMA’s, which the original said it had not provided.

ICASA hosts a Regulatory Sandbox Study Visit in Pretoria from 8 to 10 September, with the International Telecommunication Union’s Telecommunication Development Bureau, the United Kingdom’s Foreign, Commonwealth and Development Office, the UK communications regulator Ofcom, the Communications Authority of Kenya and South Africa’s Council for Scientific and Industrial Research. The visit is funded through the ITU and FCDO partnership under the UK’s Digital Access Programme, which runs in Brazil, Indonesia, Kenya, Nigeria and South Africa.

What the sandbox actually tested

ICASA says it completed trials of Dynamic and Opportunistic Spectrum Assignment between 2025 and March 2026, run inside a regulatory sandbox. The three-day agenda covers how such a sandbox is designed and run, simulations drawn from those spectrum trials, Kenya’s own sandbox experience, and international perspectives.

The principle is straightforward and the engineering is not. Spectrum that a primary licensee holds but does not use at a particular place and time can be assigned to someone else on a secondary basis, provided the primary user is protected from interference. Television white spaces was the first application of the idea in South Africa. The draft regulations extend it into what ICASA calls the innovation spectrum: 3800 to 4200 MHz and 5925 to 6425 MHz.

Those two ranges are not idle choices. The first sits in the upper part of the C band, where satellite downlinks operate. The second is the lower 6 GHz band, the same territory Wi-Fi 6E was given. Both are contested, which is precisely why a regulator would want to test assignment rules before imposing them.

The paperwork has not moved since October

ICASA published the draft regulations on dynamic spectrum access and opportunistic spectrum management on 28 March 2025, in Government Gazette 52415 under Notice 6066 of 2025. Written representations were originally due on 30 May 2025 and the deadline was extended to 13 June. Public hearings were held from 1 to 3 October 2025.

As of 7 September 2026, neither ICASA’s inquiry record for the process nor its published list of final regulations carries a final version. The only finalised white spaces instrument on that list remains the Regulations on the Use of Television White Spaces from 2018, in Gazette 41512. That is roughly seventeen months since the draft was gazetted and eleven since the hearings closed. Neither this week’s announcement nor the inquiry record gives a date for publication.

Why it matters beyond the paperwork

Shared spectrum is the cheapest capacity a country can add. It requires no auction, no new allocation and no licensee to give anything up permanently. For wireless internet providers working outside the metros, the difference between dynamic access and the status quo is the difference between using capacity that is sitting idle and waiting for the next assignment round.

It also sits directly alongside the rest of ICASA’s spectrum workload. The regulator mapped its mobile spectrum plan out to 2030 in August, and heard argument on satellite spectrum rules the same month. Dynamic assignment in the C band touches both files.

The study visit is a real signal even with the regulations unfinished. Kenya is bringing its own sandbox experience rather than only observing, and the ITU is paying for the exchange. What is being exported this week is the method of testing a rule before writing it, rather than the rule itself.

Key takeaways

• EG-IX carries close to a terabit at peak on a membership of just 18 networks, against 143 at Nairobi’s KIXP and 605 at NAPAfrica Johannesburg.
• Only 5 of the 18 are Egyptian. The rest are global content, CDN and cloud networks, plus two Jordanian carriers and a root server.
• Members hold about 3 Tbps of provisioned ports across 31 connections, and every single port runs IPv6.
• The exchange is owned by Telecom Egypt, the state-controlled incumbent, and operated on AMS-IX’s IX-as-a-Service platform. Most large African exchanges are member-owned or run by a neutral third party.
• The growth is real but concentrated: the traffic curve steps up sharply rather than climbing, which is the signature of individual large networks arriving.

Egypt’s main internet exchange moves nearly as much traffic as the biggest exchanges in sub-Saharan Africa, and it does so with 18 member networks. Nairobi’s KIXP needs 143 to fill its fabric. NAPAfrica in Johannesburg has 605.

That gap is the most interesting aspect of EG-IX, and it is no accident. It follows directly from what the exchange is: a carrier-owned platform built to attract a small number of very large content networks, rather than a community exchange assembled from a national ISP base.

What can be verified today

EG-IX publishes live traffic on its own site. On 7 September 2026 it showed 528.80 Gb/s flowing and a peak of 722.19 Gb/s over the preceding day, with average traffic in each direction of about 443 Gb/s.

AMS-IX has separately put the longer arc at 250 Gb/s to 875 Gb/s of peak traffic in under two years, describing EG-IX as one of the largest exchanges in the Middle East and North Africa. That 875 figure is a peak across the year rather than a current reading, and it has not appeared in a formal company statement, so it is best read alongside the exchange’s own live counter rather than in place of it.

Either way, the direction is not in dispute. What the traffic curve shows, though, is not steady accumulation. It sits flat through most of 2025, then steps up twice in the first half of 2026 and holds. Curves shaped like that are usually the record of individual large networks turning up and lighting ports, not of a broad base growing together.

Eighteen networks, three terabits of ports

The membership is published through routing data rather than by the exchange, and it is small enough to list in full. Across 31 ports, the members hold roughly 3 Tbps of provisioned capacity, which is three to four times the traffic the platform actually peaks at, and every port carries IPv6 as well as IPv4.

Five of the 18 are Egyptian: Telecom Egypt itself, its TE Data subsidiary, Vodafone Egypt, Etisalat Egypt and Link Egypt. Two are Jordanian carriers. One is a root name server. The remaining ten are global content, CDN and cloud platforms.

Read as a market: an exchange where international content comes to reach Egyptian eyeballs, and where the Egyptian eyeball networks are represented mostly by the incumbent and the two mobile operators. It is not yet a place where Egypt’s smaller ISPs meet.

A different kind of exchange

Most of Africa’s large exchanges are neutral in both structure and name. KIXP is run by TESPOK, an industry association of Kenyan operators. JINX, CINX, and DINX belong to INX-ZA, which is run by the ISP association of South Africa. NAPAfrica is operated by Teraco, a data centre business that sells space to everyone and competes with no one.

EG-IX is a different arrangement. It is owned by Telecom Egypt, sits inside Telecom Egypt’s Regional Data Hub at Smart Village in west Cairo, and runs on AMS-IX’s IX-as-a-Service platform, which provides switching, operations, and brand discipline without taking ownership. The incumbent carrier owns the meeting place; a foreign exchange operator runs it.

This has real consequences rather than merely symbolic ones. Policy at a member-owned exchange is set by the members, who are also the competitors; at a carrier-owned one it is set by the owner. A content network weighing a port cares mostly about reach and cost, and Egypt’s reach runs through Telecom Egypt either way, which is why the content side of the table filled up quickly.

A smaller Egyptian ISP faces a different calculation. Joining puts its traffic inside a competitor’s building, and the traffic it most wants to exchange locally is with networks that are largely absent. The 18-member roster is what that calculation looks like in aggregate.

None of this makes the exchange less useful. Keeping Meta, Amazon, Microsoft, and the large CDNs on Egyptian soil shortens paths for every Egyptian user, whichever ISP they buy from, and Telecom Egypt’s 14 subsea cable systems at the same site make Cairo a plausible interconnection point for the wider region. It does mean that EG-IX and KIXP are doing different jobs, and comparing their member counts without saying so is misleading.

Where it sits in Africa

By members, EG-IX is not close to the continent’s largest exchanges. By traffic, it is in the company. NAPAfrica remains far ahead on both counts, but the Egyptian platform reaching three quarters of a terabit at peak in four years puts it among the busiest north of the equator.

It also confirms a pattern tech.africa has tracked since the big three exchanges began expanding into Africa: the growth is increasingly delivered by international operators running platforms on local soil, rather than by local associations building their own. AMS-IX runs fabrics in Lagos and Djibouti. LINX runs Nairobi and Mombasa. That model brings capacity quickly, and it moves the governance offshore.

The counterweight is that community exchanges keep growing too, and they grow where the smaller networks are. KIXP interconnected its two Mombasa data centres this year, Kinshasa’s KINIX expanded across two buildings, and CINX added Cape Town sites without any of them adding a terabit of content ports.

What to watch

Three things will show whether EG-IX becomes an Egyptian exchange as well as an international one.

The first is whether Egyptian ISPs, beyond the incumbent and mobile operators, appear on the roster. The second is whether the traffic curve starts climbing between the steps, which would indicate the existing members growing rather than new ones arriving. The third is the subsea count: Telecom Egypt said at launch that the site would reach 18 cable systems, and each one strengthens the case for regional networks to interconnect in Cairo rather than in Marseille or Amsterdam.

For now, EG-IX is a fast, well-provisioned exchange that a small number of very large networks find worth joining. Whether it becomes the place where Egypt’s internet meets itself is another question, and the membership list is where the answer will appear first.

Digital Realty, the American data centre operator that trades on the New York Stock Exchange as DLR, announced the opening of its Nairobi Two facility, NBO2, alongside the existing Nairobi One site. In the same statement the company confirmed that iColo, the Kenyan operator now inside the group, will trade under the Digital Realty name in both Kenya and Mozambique.

What NBO2 adds

NBO2 is built beside NBO1 and extends what Digital Realty describes as a campus rather than a single building. The company said customers on the campus can reach more than 100 networks, two internet exchange points (IXPs) and a satellite teleport, the last of which it presents as a route to locations where terrestrial fibre is thin.

Neither the capital cost nor the phasing of the 6.4 MW was disclosed. The company also did not name the two exchanges on the campus, nor say how much of the new capacity is already contracted.

“The opening of NBO2 and our transition to Digital Realty are part of one story: the continued growth of Kenya’s digital economy and iColo’s evolution within a global platform,” said Wanja Muriithi, country general manager for Kenya at Digital Realty.

The iColo name goes in two countries, not one

iColo reached Digital Realty by inheritance rather than by direct purchase. Interxion completed the acquisition of a 40% stake in the Kenyan operator on 17 April 2019, and Digital Realty completed its own combination with Interxion on 12 March 2020, both dates recorded in the company’s filings with the United States Securities and Exchange Commission. The Kenyan business has therefore been part of the group for more than six years under a name that is only now being withdrawn.

The rebrand is the more consequential half of the announcement for the region, because it covers Mozambique as well as Kenya. iColo built the Maputo site that carries the group’s Mozambican presence, and it operates two facilities in Mombasa in addition to the Nairobi campus.

Local brands acquired by global platforms usually survive for a period precisely because the local name carries the customer relationships. Dropping it in two markets at once signals that Digital Realty now expects the global name to do that work, and it closes the book on an operator that built much of the Kenyan colocation market the group now serves.

Where Nairobi’s peering density actually sits

The claim worth testing is the interconnection one. Kenya Internet Exchange Point (KIXP) publishes which of its members take a port at which building, and that list is a measure of peering gravity rather than floor space.

As of 7 September 2026, it lists 130 networks at Africa Data Centres’ Nairobi facility, compared with 15 at NBO1, 9 at PAIX Kenya, and 4 at iXAfrica’s NBOX1. Digital Realty’s “more than 100 networks” refers to everything reachable across the campus, including carriers and cloud on-ramps, so the two figures measure different things. They still point the same way: Nairobi’s exchange traffic concentrates at a rival’s address, and capacity alone does not move it.

That is the number to watch over the next year. Peering density is sticky; it accrues slowly, and it is the difference between a building that holds servers and one that anchors a market.

Who was in the room

The guest list at the Nairobi ceremony is worth reading, because it says more about how this investment is being positioned than the capacity figure does.

Digital Realty listed the Kenyan state across three of its arms: John Tanui, principal secretary in the State Department for ICT; David Mugonyi, director general of the Communications Authority of Kenya; and John Walubengo, the deputy data protection commissioner. Philip Thigo, Kenya’s special envoy on technology, also attended.

So did the United States. Frank Garcia, sworn in as assistant secretary of state for African affairs on 1 June 2026, attended alongside Susan M. Burns, chargé d’affaires at the American embassy in Nairobi. The State Department’s own public schedule had Garcia travelling in Zambia, Ethiopia and Kenya between 30 August and 10 September. A serving assistant secretary of state at the opening of a commercial data centre is not routine diplomatic business, and it reflects how far digital infrastructure has climbed the agenda between Washington and African capitals.

The local industry was represented by Fiona Asonga, chief executive of the Technology Service Providers of Kenya, which runs KIXP and hosts the Kenya Peering Forum in Nairobi on 11 September, and by James Turuthi of Frontier Optical Networks and KeNIC. Michael Jacobs of Remgro Infrastructure Managers attended for the investor side.

The presence of the data protection regulator alongside the ICT ministry is the clearest signal of what the company is selling here. The pitch is not cheap racks. It is Kenyan data, kept in Kenya, next to the networks and clouds that need to reach it.

The African estate Digital Realty has assembled

Kenya and Mozambique join a footprint built almost entirely by acquisition, and on the public record it is the largest interconnection estate in Africa. Digital Realty took a 55% stake in South Africa’s Teraco in 2022, holds the former Medallion sites in Lagos and Abuja bought the year before, runs a facility in Accra and a new build at Lekki outside Lagos, and operates the Lisbon campus it positions as a landing point for African traffic.

According to PeeringDB, where operators register their own sites, the group runs 14 data centres in five African countries, carrying roughly 1,050 network records between them. The nearest comparisons on the same register are Africa Data Centres with 5 African sites and 195 records, Raxio with 6 sites and 22, and Equinix with 4 and 107. tech.africa’s own infrastructure graph, which derives presence from exchange member lists rather than self-registration, finds 709 distinct networks inside Digital Realty buildings; a network present in two countries is counted once there and twice in the table below.

The concentration is stark. Teraco’s Johannesburg campus alone lists 390 networks in PeeringDB, more than a third of the group’s African total and, by some distance, its densest site on the continent. It is also the campus that hosts NAPAfrica, Africa’s largest exchange by membership. Nairobi is the second front, and it is an order of magnitude smaller.

NBO2 opens nearly empty, which is normal

PeeringDB currently records two networks at NBO2, against 62 at NBO1 next door and 94 at the older Mombasa site. That is not a criticism of a building that opened this week: interconnection is contractual, and cross-connects follow tenants over quarters rather than days.

It does frame the question the campus now faces. The 6.4 MW has been built, and the neighbouring building already supports the ecosystem. Whether NBO2 is filled with networks or simply with racks determines the difference between an extension of the Nairobi interconnection point and a well-connected warehouse.

The Kenyan sites have been quietly busy through that period. LINX launched its Nairobi exchange in partnership with iColo in 2023, and KIXP interconnected its two Mombasa data centres this year, both of which are iColo buildings now carrying a new name.

What the company has not yet said is whether the Digital Realty brand will extend further in Africa, or whether Teraco, by far its largest asset on the continent, will keep its own name indefinitely.