Google Just Picked the Eastern Cape to Build South Africa's Newest Gateway to the World's Internet

At the Google Cloud Summit in Johannesburg on 1-2 July 2026, Google confirmed something South Africa hasn't had before: it will host the company's first Digital Exchange Port on the African continent, built in the Eastern Cape. The announcement didn't come with a groundbreaking ceremony or a ribbon-cutting - permits for the site are already secured, but construction hasn't started, the exact location hasn't been made public, and no completion date has been given. That makes it easy to skim past as another vague corporate infrastructure pledge.

It shouldn't be skimmed past. Strip away the marketing language and this is South Africa being chosen as a physical landing point for two new pieces of undersea cable - one connecting directly to Australia for the first time ever, and another opening a new route to India - plus a data centre and possible internet exchange point bolted onto the same site. Google Cloud SVP James Manyika framed it as an investment "to improve the connectivity for businesses, while supporting Google's cloud platform," and President Cyril Ramaphosa tied it to "job creation," "SME growth," and "global competitiveness" in his own remarks at the summit.

This piece explains what a Digital Exchange Port actually is, how it connects to Google's Umoja cable project and the new India route, why the Eastern Cape specifically was chosen over Cape Town (where Google already has infrastructure), and - most importantly for FastestFibre readers - what realistic effect this has on the fibre line running into your house, and on what timeline.

"Digital Exchange Port" is Google's own term, and it's doing more work than it first sounds like. Reporting on the announcement describes the Eastern Cape site as functioning simultaneously as three things:

• A subsea cable landing station - the physical building on the coast where an undersea fibre-optic cable comes ashore and is spliced into terrestrial network infrastructure. This is the same role Openserve's Melkbosstrand facility plays for Google's existing Equiano cable.
• An edge data centre - compute and storage capacity physically close to the cable landing, so traffic (and increasingly, AI workloads) doesn't have to travel far inland before it can be processed.
• A possible internet exchange point - a facility where different networks can interconnect and exchange traffic directly, rather than routing through a third party, which tends to improve local latency and reduce transit costs for the networks that connect there.

Bundling those three functions into one site is the more interesting part of the story than the cable itself. A landing station alone just gets international bandwidth to the coast. A landing station plus an edge data centre plus an exchange point starts to look like a genuine piece of internet backbone infrastructure - the kind of thing that historically only a handful of coastal cities in Africa have had (Cape Town, Mombasa, Lagos, Djibouti). If the internet exchange piece actually materialises, it would give South African and regional networks a new place to peer directly rather than backhauling traffic to existing exchanges.

The centrepiece of the Eastern Cape site is Umoja (Swahili for "unity"), a subsea and terrestrial cable project Google first announced in May 2024. Umoja is the first ever fibre-optic route to directly connect Africa with Australia - every other path between the two continents today runs indirectly, typically via Asia or the Middle East, adding distance and latency.

Umoja's route is unusual in that most of it isn't undersea at all. The cable is anchored in Kenya and runs overland through Uganda, Rwanda, the Democratic Republic of the Congo, Zambia, Zimbabwe and South Africa as a terrestrial fibre route, built in partnership with Liquid Intelligent Technologies. Only once it reaches South Africa's coast does it become a subsea cable, crossing the Indian Ocean to Australia - built and installed by SubCom, the same undersea-cable specialist behind several of Africa's other major systems. Google has said Umoja is expected to go live in 2027.

The Eastern Cape Digital Exchange Port is where that terrestrial leg ends and the subsea leg to Australia begins - the physical point where a data packet travelling from, say, Kigali or Lusaka switches from over-land fibre to undersea cable for the final ocean crossing. That's why Google needed a South African landing site at all: Umoja was always going to come ashore somewhere on this coast, and the Eastern Cape is now the confirmed answer.

Umoja isn't the only cable using the new site. Google also confirmed a new subsea fibre-optic cable connecting the Eastern Cape directly to Visakhapatnam (Vizag) and Chennai in India. Reporting places this within the context of Google's much larger US$15 billion, five-year investment commitment in India - the India route shouldn't be read as a South Africa-specific spend, but South Africa is one of the physical endpoints that investment runs through.

Taken together, the two new cables mean the Eastern Cape site isn't just "another landing station" - it's being positioned as a switching point between three continents at once: the terrestrial African network arriving from the north, a new subsea link heading east to India, and a new subsea link heading further east to Australia. That's a materially different role to the cables South Africa already hosts, which mostly run north-south along the Atlantic coast toward Europe.

Google already has a cable landing station in South Africa: Equiano, which came ashore at Openserve's Melkbosstrand facility near Cape Town in August 2022, connecting South Africa to Portugal (and, via branches, to Nigeria, Togo, Namibia and St Helena) with a design capacity of 144Tbit/s across 12 fibre pairs - at the time, the highest-capacity cable ever landed on African shores.

Melkbosstrand made sense for Equiano because that cable runs up the Atlantic coast toward Europe. Umoja's Australia leg does the opposite: it needs to cross the Indian Ocean, which means it needs a landing point that actually faces east and south, not west. The Eastern Cape - further along South Africa's coastline than Cape Town, angled toward the Indian Ocean - is the geographically logical choice for a cable whose entire purpose is reaching Australia. The same logic applies to the new India route, which is also an Indian-Ocean-facing crossing.

Google's own framing backs this up: Africa MD Alex Okosi described the Eastern Cape hub as the "southern anchor" of a planned four-hub network across the continent, with further Digital Exchange Ports planned for East Africa, West Africa and North Africa - "We're still going to have one in East Africa, West Africa and a little bit more in North Africa," in his words. None of those other three locations has been confirmed yet; the Eastern Cape is the first of the four to be publicly named.

South Africa already sits at the end of several major international cable systems - SAT-3/WASC, WACS, ACE and EASSy on the older side, plus newer high-capacity systems including Seacom, 2Africa (backed by Meta and a telecoms consortium) and Google's own Equiano. FastestFibre covered in mid-July why that existing cable network still has a real weak point: just one permanently stationed repair ship covers the entire African coastline when any of those cables actually breaks, and repairs can take weeks or months once a fault happens far from Cape Town.

A new, geographically distinct cable route doesn't fix that repair-capacity problem - Umoja and the India cable will still eventually need repair ships if they're damaged, same as every other system. What it does do is add a genuinely different physical route in a part of the coast that currently has none of South Africa's major hyperscaler cable infrastructure, which is a real, separate form of resilience: if a fault or an attack affects the Atlantic-side cluster around Cape Town, traffic has another physical path via the Eastern Cape to reroute through, rather than everything funnelling through the same stretch of coastline. More landing points, on more of the coast, is one of the few structural fixes available for a country as reliant on a handful of international cables as South Africa is.

It's worth being direct about the timeline and the audience here, because coverage of announcements like this can make it sound more immediately consumer-relevant than it is. A few honest points:

• This is years away, not months. The Umoja cable isn't expected to be live until 2027, the Eastern Cape site hasn't broken ground, and no completion date has been given for the Digital Exchange Port itself. Nothing about your current fibre plan, speed or price changes because of this announcement today.
• The primary customers are businesses and cloud workloads, not home fibre users directly. Google has been explicit that the hub supports its cloud platform, AI infrastructure (including Gemini Enterprise access) and enterprise connectivity - this is backbone and business infrastructure first, matching Google's separate push into AI compute and skills investment across Africa, including a Soweto digital innovation centre with WeThinkCode and an AI research lab in Accra, Ghana.
• Any effect on consumer pricing would be indirect and gradual. More international cable capacity landing in South Africa tends to increase competition among wholesale transit providers and can put downward pressure on the cost of international bandwidth over time - that's part of what happened after Equiano landed in 2022, which was explicitly framed as making connectivity "more accessible and affordable." But that's a multi-year market effect that plays out through your ISP's own upstream costs, not a discount that shows up on your next invoice.
• It doesn't change what "uncapped fibre" means for you today. As with the cable repair-capacity story, this is about the international layer above your ISP, not the Openserve, Vumatel, Frogfoot or Octotel line running to your house. See our explainer on what uncapped and unshaped fibre actually promises if you want the full picture on where your ISP's responsibility starts and ends.

The Eastern Cape hub is one piece of a broader pattern of Google infrastructure spending across the continent, most of it aimed at AI and cloud capability rather than consumer connectivity directly. Alongside the cable and data-centre investment, reporting on the same announcement wave references a digital innovation centre in Soweto (a partnership with WeThinkCode at the George Tabor campus of South West Gauteng College in Dube) and a separate US$37 million commitment to AI skills and research centred on an applied AI lab at the Accra AI Community Centre in Ghana. Google has described its cumulative Africa infrastructure commitment as exceeding US$1 billion.

The strategic logic Google has given publicly is that lowering the cost of connectivity for terrestrial providers - by adding more cable capacity landing on the continent - increases digital adoption among African businesses and consumers, which in turn grows the market for Google's own cloud and AI products. It's an investment made in Google's commercial interest as much as an act of infrastructure philanthropy, and both things can be true at once: the capacity is real, and it also serves Google's own growth plans in Africa.

To keep this piece honest about the limits of what's been confirmed: Google has not disclosed the exact location of the Eastern Cape site, hasn't announced a completion date for the Digital Exchange Port itself (separate from Umoja's 2027 target), hasn't published capacity figures for either new cable, and hasn't confirmed an investment amount specific to the Eastern Cape site - the R3 million Soweto centre and US$37 million Accra AI lab figures are separate line items, not the cost of the cable hub. Whether the internet exchange point function actually launches, and which local networks would connect to it, also hasn't been detailed. Those are the numbers worth watching for as the project moves from "permits secured" to an actual construction start.

A few practical things worth keeping an eye on as this story develops:

• A construction start date and exact site. "Permits secured, ground not yet broken" is an early-stage announcement. The next real milestone is a confirmed build start.
• Which South African company lands the site work. Equiano's local landing partner was Openserve; expect a similar local telco or infrastructure partner to be named for the Eastern Cape site once construction begins.
• Umoja's 2027 launch date. That's the more concrete deadline to track, since the Eastern Cape site's usefulness is tied directly to the cable being ready.
• Whether your own ISP mentions upstream transit or peering changes. If new capacity does start to affect wholesale pricing, it will show up first in ISP-level commentary about transit costs, not in headline consumer price cuts - worth watching your provider's own announcements rather than expecting an automatic saving.