South Africa’s fibre market grows fast enough to look healthy on every local chart, yet the global speed table heads the other way. That split has less to do with a broken network than with who is joining, what package they buy, and how speed gets measured once the signal leaves the fibre line and starts fighting with the home router.
The numbers are awkward only if you treat them as one thing. They are not. More homes are being connected, average speeds are still rising, and the country is still sliding down the international ranking because other markets are improving faster.
Fibre keeps growing
Fibre-to-the-home and fibre-to-the-building subscriptions climbed from 2.47 million in 2024 to just over 3 million in 2025. That is a sharp jump by any standard, showing that fibre is still taking over the fixed-line market.
Overall fixed broadband subscriptions also grew by 19.3% over roughly the same period. That matters because it tells a cleaner story than the speed ranking does. The market is expanding, not shrinking. More households are buying fixed access, and fibre is doing the heavy lifting inside that growth.
Openserve, Vumatel, Frogfoot, MetroFibre, and Herotel sit behind much of that expansion. Their networks give internet providers room to sell more lines, widen coverage, and offer more choice. Openserve still carries the weight of a large legacy footprint. Vumatel remains a big residential player. Frogfoot has a strong metro and secondary-city presence. MetroFibre has pushed hard on residential and business fibre. Herotel has helped extend service into smaller towns that used to have fewer options.
The fibre footprint is wider, and more connections are being sold through it.
The speed numbers are doing two jobs
The global fixed-broadband ranking fell from 102nd to 115th over the same broad period, even though measured national download speed kept improving. That sounds like a contradiction until you remember that a ranking is comparative. A country can get faster and still lose places if the rest of the field gets faster first.
This seems to be happening here. The country is not standing still. It is being overtaken by markets where faster fibre tiers, cheaper gigabit offers, and broader take-up of high-end packages are moving the benchmark upward more quickly.
Once you separate the ranking from the underlying speed, the picture becomes less dramatic and more useful. A drop in place does not automatically mean the fibre line in the street became slower. It often means the gap between local progress and global progress widened.
That distinction matters because speed tables are easy to misread. They compress many different realities into one number and pretend that number tells the whole story. It does not.
Package choice is pulling the median down
The next piece is the one many glossy fibre adverts skip. New customers do not all arrive on 200Mbps or 1Gbps plans. Plenty of them choose 25Mbps or 50Mbps packages because those tiers are cheaper and still cover streaming, schoolwork, WhatsApp, and browsing without drama.
That buying pattern lifts subscription numbers while holding back the national median. If a wave of new sign-ups lands on entry-level packages, the average experience shifts downward even though the network itself keeps expanding and the top end stays fast.
The market structure encourages that mix. Wholesale operators such as Openserve, Vumatel, Frogfoot, MetroFibre, and Herotel give retail ISPs a wide menu, from low-cost starter tiers to high-speed options. The consumer then chooses the price point, not the engineer. In a country where many households are still price-sensitive, the first choice is often the cheaper one.
The speed story is being shaped by demand as much as by supply. More fibre is available, but a growing share of it is being sold at speeds that are good enough for the household budget, not necessarily high enough to push the national median sharply upward.
Home gear still gets in the way
The other quiet culprit is the house itself. A fibre line can be healthy and still produce disappointing speed tests if the router is old, badly placed, or overwhelmed by too many devices. Wi-Fi walls are not the same as fibre walls. The first can ruin a test in seconds.
Tests run over Ethernet and over Wi-Fi across 25Mbps, 50Mbps, 100Mbps, and 1Gbps tiers usually tell two different stories. Ethernet is the cleaner measurement because it removes the wireless layer. Wi-Fi has to deal with interference, distance, thick walls, neighbouring networks, and routers that were never meant to carry modern speeds properly.
That is why a 100Mbps package can feel fine at the line but sluggish in the lounge. It is also why some people blame the ISP when the real bottleneck is a router that should have been retired years ago. Congestion inside an ISP’s network, especially during busy periods, can also trim results before traffic even reaches the home network.
The measured speed that appears in public rankings is therefore a mix of access network quality and in-home behaviour. It is not a pure reading of the fibre cable alone.
The real race is outside the border
The central question is whether the problem is weak rollout, low package speeds, bad Wi-Fi, or the failure to keep pace with cheaper and faster markets elsewhere. The answer is annoyingly ordinary: it is all of them, in different proportions.
The rollout is not the main failure point. The subscription data says fibre is still spreading. The bigger drag comes from the package mix at the lower end and from the fact that other countries are moving faster up the speed ladder. A home on 50Mbps fibre will not look impressive in a global comparison if a growing number of markets are pushing much more of their customer base into 500Mbps and gigabit territory.
That leaves the industry with a practical job rather than a rhetorical one. It needs more homes passed, yes, but it also needs more households moved onto faster plans, better routers in those homes, and less confusion between line speed, Wi-Fi speed, and ranking position.
The fibre boom is real. The ranking slide is real too. They just describe different parts of the same market.
