Fibretime challenges South Africa’s population figures, says country may have over 95‑million people
Fibretime, a private fibre-network operator working deep inside South Africa’s townships and peri-urban communities, has thrown a bold challenge at the state’s official population figures, arguing that the country may be far more crowded than government data suggests.
The company, founded by telecoms entrepreneur Alan Knott-Craig, believes the national headcount could be above 95‑million people — far higher than the official estimates — and says its own operational data, backed by aerial imaging, drone footage and AI modelling, shows a pattern of serious undercounting on the ground.
Fibretime builds and operates fibre infrastructure, with a particular focus on townships, informal settlements and lower-income areas often overlooked by traditional operators. As part of its rollout model, the business physically counts homes before deploying its network, using those numbers to plan infrastructure and manage investment decisions.
“Every time we go into a township — and we’ve worked in 32 townships and nine cities across the country — we do a physical count of homes. Because we count every home, we see the reality on the ground. Each time, the number of homes we find is almost double what the government records show," Knott-Craig told Business Day.
“After surveying 39 townships, with another 30 in planning, we realised the official national population figure looks too low. Using rigorous data science — aerial imaging, drone footage, AI and careful extrapolation from our rollout samples — and adjusting for migration to avoid double-counting, we arrived at a range of estimates,“ he said.
According to Fibretime, these counts form part of an “operational sample” of more than 250,000 homes already connected, plus planning data for over 2‑million additional homes across the country. The company combined that internal dataset with government sources and then ran various scenarios to estimate what the national picture might look like if township realities were properly captured.
The result is startling. Fibretime’s work points to a possible population figure above 95‑million — and potentially far higher at the upper bound.
“We’re not declaring a definitive number; we’re asking whether the population could be as high as 95‑million, and our top‑end estimate is 124‑million.”
If the population were 95‑million, that would mean roughly 33‑million more people than currently counted — and most of those would not be undocumented immigrants but South Africans who have simply not been registered. Knott-Craig argues that the implications for planning, budgeting and service delivery are enormous.
“As an infrastructure company, we plan networks using official statistics, so if those stats are wrong, it affects decisions about schools, hospitals and roads. If the population were 95‑million, that would mean roughly 33‑million more people than currently counted — and most of those would not be undocumented immigrants but South Africans who have simply not been registered,” he said.
“Many children born in townships are never registered and therefore don’t appear in official records. This undercounting has been compounding for decades, and it means we are not planning properly for the people who actually live here.”
Fibretime set out its analysis on its public site, explaining how it extrapolated from its network rollout data and blended that with new forms of remote sensing. The company emphasised that physical home counts in “real streets, real shacks and real backyards” underpinned the findings. Those ground counts were then cross-checked with aerial imaging, drone footage and AI models to estimate densities and to correct for people moving between areas.
The company also said it adjusted for migration to avoid double-counting individuals who might be picked up across different locations as they move in search of work or better living conditions.
The claim, first reported by Moneyweb, lands at a time when the reliability of official statistics is increasingly under scrutiny from parts of the private sector. Earlier this year, Capitec outgoing CEO Gerrie Fourie suggested that South Africa’s unemployment rate may overstate joblessness because it does not fully capture informal work and self-employment, especially in townships and rural areas.
Fibretime’s assertion effectively serves as a stress test for Statistics SA (Stats SA), the underfunded national statistics agency that carries the responsibility for censuses, labour force surveys and other headline indicators.
Early last month, Business Day reported that the Statistics Council — a 25‑member oversight panel that signs off on national statistics — had raised alarm about high vacancy rates and a shrinking operational budget at Stats SA. The council warned that chronic underfunding and staff shortages were undermining the agency’s ability to produce robust, timely data.
Concerns about data quality are not new. The Medical Research Council (MRC) last year described the 2022 census as “a work of fiction”, accusing it of overstating the population and saying there had been a 31% undercount in mortality and labour data. That criticism suggested some statistics were inflated while others — particularly linked to deaths and employment — were understated.
The MRC’s harsh assessment placed Stats SA under intense pressure, as it defended the integrity of its census process and methodology. The agency has previously stood by the 2022 census results and rejected the “work of fiction” characterisation.
Stats SA did not immediately respond to a request for comment on Fibretime’s latest claim, but repeated challenges from respected institutions and private players risk eroding public confidence in the country’s core datasets. If population numbers, unemployment rates and mortality figures are all contested, it becomes harder for policymakers, investors and service providers to plan with certainty.
Knott-Craig, however, insists that Fibretime’s work is less about scoring political points and more about highlighting a practical problem that affects everything from school capacity to hospital beds.
By relying on official population and household counts that may be significantly too low, he argues, both the public and private sectors could be under-building and under-servicing large parts of the country. The people most affected would be exactly those living where Fibretime operates: dense informal settlements, backyard structures and township extensions that have often grown faster than the state’s record-keeping.
If Fibretime’s estimates are even partly correct, South Africa’s infrastructure plans — in telecommunications, energy, water, housing, health and education — could be based on a fundamental misreading of how many people actually live in the country, and where they are.
For now, the fibre operator is careful to present its numbers as a challenge rather than a final answer. The company says its objective is to spark a serious conversation about how to measure South Africa more accurately in the age of drones, satellites and AI, and to push for better integration of on-the-ground operational data into official planning.
Whether Stats SA chooses to engage publicly with the methodology, or to open the door to collaboration with private players like Fibretime, could shape how future censuses and surveys are designed — and how close South Africa gets to counting all of its people, not just those who appear on paper.
Follow Us on Twitter
