Building Where Incumbents Won't

Why fibertime bet on aerial fibre instead of wireless

fibertime's core insight is that wireless connectivity cannot solve the township internet problem, and only physical fibre into every home can. Knott-Craig is blunt about it: wireless always congests. "You go to any community anywhere in the world that's densely populated, anywhere in the world that has money, they're not using wireless," he said. No 5G, no Starlink. The speeds slow, the experience degrades, and most South Africans live in exactly the kind of dense communities where that congestion bites hardest.

The company also broke from the traditional fibre playbook in a second way. Rather than trench cable underground, which is around three to five times more expensive, fibertime strings aerial fibre along poles, the same approach cities such as Tokyo use for the highest fibre penetration in the world. In a place like Alexandra, up to 32 homes hang off a single pole, and trenching is close to impossible anyway. The result is internet Knott-Craig says is 85% cheaper than the leafy-suburb equivalent, with a UPS and a router installed on the wall of every connected home.

How a 5-rand voucher became a working business model

fibertime's revenue comes from prepaid vouchers rather than contracts, a deliberate choice built around how township customers already spend. People are already buying mobile airtime, Knott-Craig points out, so fibertime is not asking them to adopt a new habit, only to swap to far better value. A customer buys a five-rand voucher at a spaza shop or through the Capitec app and gets 24 hours of uncapped 100Mbps.

The value gap is stark. fibertime's biggest customer in September last year burned through 1.6 terabits of data in a month playing Call of Duty and FIFA, and paid roughly 85 rand in fibertime vouchers. The same usage on MTN mobile bundle rates, Knott-Craig said, would have cost 1,650 rand. The router itself doubles as the advertising: sitting on the wall of every home fibertime installs, it prompts residents to try their first voucher. On marketing and activation, he rates the company "1 out of 10," but the networks are already profitable, so every voucher sold adds margin.

Funding infrastructure at scale with a REIT model

The financing behind fibertime borrows from property, not telecoms. Knott-Craig grew tired of the "stop, start, stop, start" of raising money against his own balance sheet, so fibertime adapted the real estate investment trust (REIT) structure: as it builds a network it sells the infrastructure into a fund and rents it back, a sale-and-leaseback that keeps the debt off its own books and recycles capital straight back into the next build.

Getting there was a hard sell, and the obstacle was track record rather than the safety of the asset. "How do you create a track record unless they give you the funding, but they're not going to give you the funding unless you've got a track record," he said. Java Capital helped assemble the structure, FairVest came in as anchor, and RMB backs the business as a financial partner. With momentum, funders have grown comfortable that a fibre network is a 20-year, yield-generating asset. The numbers are large: roughly 2 billion rand deployed so far, with another 600,000 homes and 5 to 6 billion rand planned over the next 12 months, and a total addressable mission Knott-Craig sizes at something like 100 billion rand to connect all of South Africa's homes.

This episode was sponsored by

Silicon Overdrive Founders Series

Visit Silicon Overdrive↗

Why profit and purpose are not in conflict

Knott-Craig is direct that the mission depends on making money, not despite it. "Anyone can change the world and lose money," he said. The purpose is to connect every home, but the only place 100 billion rand of capital comes from is investors who expect a return, not donors. Profit, in his framing, is simply the mechanism that unlocks more capital to connect more homes.

That clarity shapes who he lets onto the cap table. After bruising experiences with shareholders in the past, he now vets backers hard, and says fibertime's investors, including RMB, understood the long game before joining. The pressure he fields is not to raise prices but to lower them: once economies of scale arrive, fibertime intends to cut the retail price while keeping its margins, putting more people online. He has held the same discipline on hard operational calls. When load shedding hit stage six and wiped out revenue as power vanished, fibertime added UPS units to keep routers alive without changing its pricing. Now, he says, revenue climbs during load shedding, because when the lights go out there is little to do but use fibertime.

What Mixit taught him about second chances

fibertime is the product of a career that included a very public failure. Knott-Craig bought the South African messaging platform Mixit in 2011, believing it could compete with WhatsApp, and by his own account got the strategy, the smartphone experience and the business model wrong. "I was grossly incompetent," he said. He was pushed out in 2012, lost a lot of money, and ended up renting a small cottage for eight years in a town small enough that everyone knew.

He drew three lessons he still applies: under-promise and over-deliver, because business runs on trust and trust cannot survive broken promises; be frugal, unlearning the free-spending habits of the mobile-operator world he grew up around; and above all, focus. "The hard part is not saying yes, the hard part is saying no," he said. He now declines almost everything that is not fibertime. The recovery ran through a non-profit, Project Isizwe, working with Tshwane municipality on free Wi-Fi, then Herotel, which he sold in 2022, rebuilding enough of a reputation for making investors money that raising capital for fibertime came more easily. Crucially, he says he could raise again after Mixit because he had been incompetent, not corrupt: "I hadn't stolen any money. You don't get second chances if you're dodgy."

[CHAPTERS]

Why can't wireless or satellite solve township internet?

Wireless and satellite cannot deliver reliable, affordable internet to dense townships because they congest, according to Knott-Craig. In any densely populated, moneyed community worldwide, he argues, people abandon 5G and satellite because shared spectrum slows to a crawl once enough users are on it. Fibre does not suffer the same ceiling, which is why he insists it is the only technology that gives township residents a genuinely world-class experience. For rural areas with no fibre backbone, fibertime has been in early talks with Starlink about transmission, but the metros, where roughly 70% of homes sit, are the priority.

How does fibertime make money at five rand a day?

fibertime makes money because the model only works at scale and its networks are already profitable per voucher sold. Knott-Craig concedes it would never work at 100,000 homes, but at hundreds of thousands the aerial-fibre cost base and prepaid volume produce, in his words, "world-class profit margins." The business sells into a REIT-style fund and rents the infrastructure back, recycling capital so it can keep building without repeatedly stopping to raise fresh rounds. The customer economics do the rest: at five rand a day for uncapped 100Mbps, fibertime undercuts mobile data so heavily that usage, and therefore voucher revenue, climbs on its own once a home is connected.

Could fibertime become the next Capitec?

Knott-Craig believes fibertime could become the next Capitec in terms of impact and value creation, and that is the comparison he reaches for when asked where the business goes. He hopes never to sell it, wants eventually to list it on the JSE, and says his real ambition is for township residents themselves to be able to buy shares and build wealth from a company serving their own communities. He is also eyeing the slums of Latin America, the Philippines, Indonesia, Bangladesh and the rest of Africa, while cautioning that South Africa must be "rock solid" first. The through-line is the same one that started it all: get people onto the internet, and make money doing it, so the mission can keep funding itself.