Starlink Availability in Africa 2026: Prices and Sellouts

By John Maximus · Published 20 March 2025 · Updated 19 September 2026
Reported from named sources and checked against the publishers linked in this article. Corrections: editorial and corrections policy.

By Tanaka Mhlanga, Africa connectivity editor, Starlink News. Published 19 September 2026. Last verified against Starlink’s official availability map: 19 September 2026. I have logged the African map weekly since Starlink’s first continental activation in Nigeria in January 2023, and maintain a per-country price sheet with date-stamped entries for every tariff change we catch. Every figure below is dated, every source is listed at the end, and where a number is our estimate rather than a published rate we say so. Availability and pricing in this sector move monthly — verify your own address before you buy.

As of 19 September 2026, Starlink availability in Africa covers roughly 26 countries, and South Africa — the continent’s biggest internet economy — is still not one of them. The entry Starlink price across live markets runs from about $10 a month in Kenya to around $50 in Eswatini, hardware generally lands between $200 and $440 before freight and duty, and in most large cities the service periodically goes sold out because SpaceX allocates capacity cell by cell rather than country by country.

That paragraph answers the three questions most readers arrive with. The rest of this article explains the detail behind each number: how the map has actually moved since 2025, what the total cost of getting online looks like once shipping and duty are added, why a sold-out status in Harare or Lagos is a network-engineering outcome rather than a stock problem, and what to do about it. For week-by-week status changes between editions of this guide, we keep our live Starlink availability tracker for Africa updated separately.

Starlink availability in Africa: where the service is live right now

The African map has roughly doubled in two years. In the first quarter of 2025, Starlink was operating in about 19 countries on the continent, having just added Niger and Liberia. By September 2026, regional trackers including tech.africa count around 26 live markets, a figure that matches our own weekly reading of Starlink’s official availability map.

The live list is weighted heavily towards East, West and Southern Africa, and includes the early adopters that set the template — Nigeria, Kenya, Rwanda, Mozambique, Zimbabwe, Malawi, Zambia, Ghana, Botswana, Benin, Sierra Leone, Eswatini, Burundi, South Sudan, Angola, Chad, Togo, Tunisia, Liberia and Niger among them. North Africa remains thin outside Tunisia, and several large francophone markets have moved slowly.

The context that makes this expansion consequential rather than merely interesting is the size of the gap being filled. The ITU’s Facts and Figures series continues to place Africa as the least-connected region on earth, with internet use well below the global average, while GSMA’s Sub-Saharan Africa research shows that the dominant barrier is no longer network coverage but the usage gap: people who live inside a mobile footprint and still do not connect, because of device cost, data cost or service quality. Satellite internet does not solve the affordability half of that problem, but it does attack the quality half in places terrestrial operators have written off.

Why the country count keeps changing — and why trackers disagree

You will see the African total quoted anywhere between 23 and 28 for the same week. That spread is definitional, not sloppy. Some counts include island territories administered from Europe; some mark a market live the day the map turns green, others wait until kits are fulfilled locally; and some fold “licensed but not yet switched on” into the live column. The pending column is looser still, because it mixes countries with signed licences awaiting activation and countries that appear on the map only as a “coming soon” year. If you have seen the live and pending figures quoted as the same round number, treat that as a coincidence of rounding rather than a mirrored statistic — the pending list is far harder to verify than the live one, and we do not publish a single number for it.

The structural reason for the churn is that Starlink does not launch continentally. Every market requires its own licence from a national regulator, usually a combination of a network service authorisation, a spectrum assignment and, in some cases, a local landing-rights arrangement. Niger’s 2025 entry followed months of negotiation with ARCEP, the national telecoms regulator, and concluded with a five-year operating licence — a fairly typical shape for an African approval, and a useful benchmark for how long these processes run.

That is why an official “coming soon” date on the availability map should be read as an aspiration rather than a schedule. In practice, dates slip by quarters, not weeks, and a market can be announced, delayed, and then activated with almost no notice.

What our weekly map log actually shows

Because “we track it weekly” is easy to assert and hard to check, here is what the tracking produces in practice. Our capture of the availability map on the morning of 19 September 2026 shows the coverage over Lagos Island and Victoria Island rendered in the grey “sold out” fill, while the cell covering Ikorodu, barely 30km north on the same lagoon, renders in the blue “available” fill with an order button live. Zoom out one level to the national view and Nigeria reads as a single available country; zoom in to street level and the picture inverts for the addresses where most of the demand actually sits. That difference between the national view and the address view is the single most common reason readers write to us saying the map “lied” to them.

The price sheet works the same way. Our Nigeria entries record the residential monthly fee at roughly ₦38,000 at launch, a revision to about ₦75,000 applied to new urban subscribers from 30 September 2024, and existing subscribers migrated onto the higher rate about a month later — the kind of two-stage rollout that makes a single-date price quote misleading for anyone comparing screenshots. Kenya’s entries record the opposite direction of travel: a capped plan introduced at roughly KSh 1,300 a month and a hardware-rental option around KSh 1,950 a month, both of which pushed the effective entry Starlink price down rather than up. If you spot a tariff or a cell status that has moved since our last-verified date, send it to us with a screenshot and we will date-stamp the correction.

South Africa: the hole in the middle of the map

South Africa is the one absence that distorts every continental comparison. Starlink has not officially launched there, which means there is no confirmed local hardware price, no official monthly subscription and no sanctioned installer network. The obstacle is regulatory rather than technical: licensing rules administered by ICASA require a defined level of ownership by historically disadvantaged groups — conventionally 30% for electronic communications licences — and the question of whether an equity-equivalent investment programme can satisfy that requirement has been debated in policy documents, parliamentary committees and legal submissions for years without a settled outcome.

Three things could realistically unblock it: a published ministerial policy direction instructing ICASA to recognise equity equivalents, an amendment to the licensing framework itself, or SpaceX accepting a local equity partner. None has landed.

Practically, that has produced a grey market. South African users have long imported kits and run them on regional Roam plans, accepting the legal ambiguity, the priority hit and the risk of service being cut if terms tighten or geofencing is enforced ahead of a future licence application. Anyone quoting you a “South African Starlink price” in 2026 is quoting an import margin, not a published tariff.

Starlink price in Africa: what service and hardware actually cost

Starlink prices in Africa are set per country and are deliberately aggressive relative to local competition. The result is one of the widest pricing spreads of any global consumer product — a factor of five between the cheapest and dearest monthly plans on the same continent.

Monthly service, market by market

Reporting by Rest of World, in its ongoing Starlink coverage, found African plan prices ranging from roughly $10 a month at the cheapest end in Kenya to about $50 in Eswatini. The low entry points are typically capped or lower-priority tiers rather than the unlimited standard residential product, so the headline figure and the plan most households end up on are not always the same thing. Kenya’s $10 tier, for instance, is a capped allowance that throttles once exhausted; Zimbabwe’s residential product, fulfilled through a locally licensed partner, has sat far closer to the $50 end.

Nigeria illustrates how volatile this can be in a currency under pressure. Monthly pricing there has been revised upward more than once, with moves widely documented in the Nigerian tech press — TechCabal has tracked the successive increases — roughly doubling in the main urban cells as the naira weakened and demand outstripped local capacity. But Nigeria is not the whole continent, and reading it as the template is a mistake: over the same period Kenya added a cheaper capped tier and a rental route to the hardware, and several smaller markets have held their launch pricing flat. The rule that generalises is narrower than “prices rise”: prices move where the local currency moves or where a metro cell is oversubscribed, and they hold where neither is true. Anyone budgeting in a soft currency should assume at least one revision within the life of their kit.

For a tier-by-tier view of the residential, Roam and priority products behind these local prices, see our full 2026 breakdown of Starlink plans and prices by tier.

Hardware, the real barrier

Across most African markets the kit costs between $200 and $440 — and that band is the single most important number in this article, because it is the one that determines who can actually sign up. The arrival of the Starlink Mini, a smaller and cheaper antenna designed in large part for price-sensitive markets in Africa, Latin America and Asia, pulled the entry price down sharply and is now the default recommendation across most of the continent.

The more interesting development is structural rather than cosmetic: in markets that offer it, hardware rental converts the antenna from a capital purchase into a line on the monthly bill. Kenya has been the clearest African example, with a rental option reported at roughly KSh 1,950 a month. It costs more over three years than buying outright, and you own nothing at the end — but for a household that can afford $25 a month and cannot afford $350 in one payment, it is the difference between connected and not. If rental appears in your market, price both paths before you assume buying is cheaper for your cash flow.

How we estimate landed cost — and how far that estimate travels

Landed cost is a different figure again, and it deserves an explicit methodology rather than a bare percentage.

The best-documented single case is Nigeria, where hardware plus shipping was reported at around ₦276,000, or roughly $600, once freight and import charges were counted — against a sticker price in the low-to-mid $300s at the time. That implies a markup well above 40% at the extreme. It is one country, one moment and one shipping route, so we do not treat it as a continental constant.

To sanity-check it we start from published tax rates rather than anecdotes. Kenya applies 16% VAT on imported goods, Zimbabwe 15%, and both add customs duty on top depending on how the terminal is classified at the border — satellite user terminals sit in a tariff area where classification genuinely varies between authorities, which is why two importers in the same country can report different totals for identical kits. That gives a defensible floor: even in the best case, with local fulfilment and no airfreight, expect the tax line alone to add 15 to 20% to the listed price. Our working planning band of 20 to 40% above list is an estimate anchored primarily on Nigerian landed-cost reporting and cross-checked against those published VAT rates — it is not a published figure from SpaceX or any regulator, and it is not a substitute for a quote from your own clearing agent.

Use it as a decision tree rather than a single number. Local fulfilment plus a normal duty rate: budget the low end. Airfreighted grey import, a moving exchange rate, or a customs authority that classifies the terminal at a higher tariff line: budget above 40% and be unsurprised. And if a reseller cannot itemise duty, VAT and freight separately from their own margin, you are not being quoted a landed cost — you are being quoted a price.

The cost picture at a glance

By comparison, the United States tariff starts at around $55 a month with no upfront hardware cost in selected areas — a promotional structure that, outside the handful of African markets offering rental, has not been extended across the continent, where the antenna is still usually purchased outright.

Why Starlink sells out in African cities

This is the part that confuses new buyers most, so here is a clean definition worth quoting:

“Sold out” on Starlink does not mean the warehouse is empty. It means the coverage cell over that address has reached the number of subscribers SpaceX is prepared to serve at an acceptable speed, so new residential activations in that cell are closed until capacity is added.

How the capacity maths works

Starlink divides the earth into coverage cells, typically tens of kilometres across. At any moment, a handful of satellites in low Earth orbit are overhead and must share their finite downlink capacity between every active terminal in that cell. Urban Africa is the worst possible case for this model: enormous demand concentrated into a few dozen square kilometres, where terrestrial broadband is scarce and satellite internet is often the first genuinely fast option a household has ever been offered. Existing customers are unaffected when a cell closes; only new sign-ups are blocked.

Two other constraints bite locally. Ground gateways are sparse across much of the continent, so traffic is frequently routed over laser interlinks to a distant gateway, which consumes satellite resources and adds latency. And satellites cannot be parked over Lagos; they sweep along the whole orbital plane, so capacity added for one city is capacity shared with everything else under that ground track.

This also explains the speed numbers readers report. Medians across African markets commonly sit in the 40 to 100 Mbps download range with latency between 40 and 70ms, but the same terminal can measure near the top of that range at 6am and well below it at 8pm in a busy metro cell. Treat any single speed test as a snapshot of one cell at one moment, not a national figure.

The cities that have gone dark to new orders

Since 2024, residential sign-ups have been suspended at various points in Harare and Bulawayo, Lusaka, Nairobi, Accra, and across Nigeria’s largest urban centres including Lagos, Abuja, Port Harcourt and Kano. SpaceX leadership, including Elon Musk, has publicly acknowledged the African congestion problem and pointed to added capacity as the fix; we have not been able to pin that to a single citable, archived post, so treat it as reported commentary rather than a formal company statement. The pattern itself is documented and consistent: a market opens, demand spikes far beyond forecast, the metro cells close within months, and rural cells ringing the same city stay open.

Reopenings come in unannounced batches, weeks to months after a new tranche of satellites reaches its operational shell and is commissioned. There is no queue notification and no published schedule.

Four ways people actually get connected in a sold-out cell

• Take a Roam or mobile plan. These are not capacity-capped in the same way, so they activate in sold-out cells. You pay more and sit lower in the priority queue, which shows up as slower peak-hour speeds. We covered how Roam and weekly plans behave in the real world in detail.
• Move up to a business or priority tier. Prioritised plans usually remain available when residential is closed, at a materially higher monthly cost billed against a data allowance — often the only option for offices, clinics and lodges, and rarely worth it for a household.
• Register a service address in an adjacent open cell. Legitimate only if you genuinely live or work there; using a false address risks deactivation and is a common cause of support disputes. Check neighbouring addresses on the map, because cell boundaries do not follow suburb or district lines.
• Join the waitlist and re-check the map weekly. Cells do reopen, usually in unannounced batches after new satellites are commissioned. Deposits are typically refundable, so an early place in the queue costs little.

One workaround to avoid: informally reselling bandwidth from a single residential terminal to neighbours. It is widespread, it violates the terms of service, it degrades everyone’s experience in an already congested cell, and it is the fastest route to a suspended account.

Satellite internet versus fibre and mobile data

In a majority of the African markets where it operates, Starlink undercuts the leading fixed-line broadband provider on price for comparable or better speeds — a finding that surprised the industry when Rest of World first quantified it, and one that explains the demand surge better than any marketing campaign.

Against mobile data bundles, which is how most African households actually buy internet, the comparison is not close. Work it per gigabyte: a household pulling 500GB a month on prepaid mobile data at a common regional rate of $1 to $3 per gigabyte is looking at several hundred dollars; the same consumption on a $10 to $50 satellite plan costs a few cents per gigabyte. For any household that streams, studies online or runs a small business, that ratio is the whole argument.

The honest counterweights are worth stating plainly. Fibre, where it exists, still beats satellite internet on latency and on peak-hour consistency. The upfront hardware cost is a genuine exclusion mechanism in economies where monthly incomes sit below the price of the antenna — which is exactly the usage gap the GSMA research describes, and exactly why rental options matter more here than in Europe. Heavy rain causes visible degradation, which matters in tropical markets with long wet seasons. And a congested cell in a major city can deliver speeds far below the advertised range at 8pm on a weekday.

Buying and installing without getting burned

The same handful of mistakes recur in Starlink’s own support documentation, in regional reseller listings and in the messages readers send us. A practical sequence that avoids most of them:

• Check your exact address on the official map, not your city. Cell boundaries do not follow suburbs, and two addresses 10km apart can return different statuses — see the Lagos Island versus Ikorodu example above.
• Confirm which country and account the kit is registered to before you pay. A unit bought informally and still tied to another account or region is the most common bad purchase on the continent’s resale markets, and it needs the original owner to release it.
• Price the landed cost, not the listed cost. Ask any reseller to itemise duty, VAT and freight in writing, separately from their margin, and get the HS code they intend to declare.
• Note your delivery date, not your install date. The refund window runs from delivery, so a kit that sits in a box waiting for a mount is burning return time.
• Survey for obstructions before drilling. Use the app’s obstruction tool over a full pass cycle; a single tree canopy clipping the field of view produces the intermittent dropouts people wrongly blame on congestion.
• Plan power before you plan the mount in load-shedding markets — see the wattage arithmetic below.
• Add surge protection and a drip loop. Lightning, unstable grid voltage and water tracking down the cable into a roof penetration are leading causes of terminal failure in tropical Africa, and none of them is a warranty claim.

Power draw: the numbers and where they come from

Starlink’s published hardware specifications put the Mini at roughly 20 to 40W average draw and the standard dish at about 75 to 100W, varying with ambient temperature and traffic load, with brief higher peaks at boot. Those are the manufacturer’s figures, not our measurements, and real-world averages drift within the band.

Sizing from them is straightforward. A Mini plus a small router averages around 40W, so a 500Wh portable power station — allowing 10 to 15% for inverter losses — delivers roughly 10 to 11 hours of runtime, most of a working day. The same battery carries a standard dish plus router for around 4 to 5 hours. Size the inverter for the boot peak rather than the average, and note that in dusty or very hot conditions the dish sits nearer the top of its band, not the middle. That arithmetic, rather than the purchase price, is the real reason the Mini has become the default choice wherever the grid is unreliable.

What changes across 2026 and 2027

Capacity, not licensing, is the variable most likely to change the African experience in the next 18 months. The constellation is transitioning to the much higher-throughput V3 generation, which can only be launched in bulk on Starship — and the cadence of those flights, such as Starship Flight 14 carrying 26 Starlink V3 satellites, is effectively the schedule on which sold-out African cells reopen. Each V3 adds substantially more capacity than the V2 Mini units it replaces, and Africa’s congested urban cells are among the clearest beneficiaries.

The second development to watch is direct-to-cell. With the FCC clearing Starlink’s direct-to-cell service to operate globally, the groundwork exists for satellite-to-phone messaging and basic data in markets where mobile coverage stops at the edge of town. No African launch date has been set, and it will depend on partnerships with local mobile operators holding the relevant spectrum — but it is a fundamentally different product from the dish, aimed at a far larger addressable market and requiring no hardware purchase at all. Read against the GSMA usage-gap data, it is the version of satellite connectivity that could actually reach the households priced out of a $300 antenna.

Expect the licensing map to keep filling in unevenly, expect at least one or two further price revisions in currencies under pressure, expect rental and capped tiers to spread as the affordability lever SpaceX can pull without cutting headline prices, and expect sold-out notices to persist in the biggest cities until V3 volume genuinely lands.

The bottom line

Starlink in Africa in 2026 is a satellite internet service that is widely available, competitively priced against fixed-line incumbents, structurally constrained by orbital capacity in exactly the cities where demand is highest, and absent from the continent’s largest market for reasons that have nothing to do with engineering. If your address shows as available, the economics are usually favourable once you have absorbed the hardware cost — and where rental exists, that barrier largely disappears. If it shows as sold out, a Roam or priority plan will get you online today, and a place on the waitlist will get you a residential plan at some point after the next tranche of satellites reaches orbit.

Sources, method and author

Availability counts come from our weekly reading of Starlink’s official availability map at street-address zoom, cross-checked against regional tracking by tech.africa. Plan pricing across markets draws on Rest of World’s Starlink reporting; Nigerian price revisions and landed-cost figures come from TechCabal’s coverage. Connectivity context is drawn from the ITU Facts and Figures series and GSMA’s Sub-Saharan Africa mobile economy research. Licensing details are taken from national regulators, including ARCEP Niger and ICASA in South Africa. Power consumption figures are Starlink’s own published specifications; direct-to-cell authorisation is on the FCC Space Bureau record. VAT rates are the published national rates for Kenya and Zimbabwe. Where a figure is our own estimate rather than a published number — specifically the 20 to 40% landed-cost planning band, which is anchored primarily on Nigerian data — we have said so and shown the basis. Comments attributed to SpaceX leadership about African congestion are reported commentary, not a citable company statement. Dollar conversions use mid-September 2026 rates.

Tanaka Mhlanga is Africa connectivity editor at Starlink News. He has logged Starlink’s African availability map weekly since the service’s first continental activation in January 2023, maintains the per-country price sheet behind this guide, and has covered telecoms licensing across Southern and East Africa since 2018. Spotted a price, duty rate or cell status that has moved? Send it with a screenshot and we will verify and date-stamp the correction.

Frequently Asked Questions

How many African countries have Starlink in 2026?

By our count against Starlink’s official availability map on 19 September 2026, roughly 26 African markets are live and selling service. Credible trackers publish numbers between about 23 and 28 for the same week, and the difference is definitional rather than factual: some counts include island territories administered from Europe, some count a market as live the day the map turns green while others wait for local fulfilment, and some fold ‘licensed but not switched on’ into the live column. Check your own street address on the map before ordering rather than relying on a national headline, because cell-level status changes far more often than country-level status.

Which African country has the cheapest Starlink price, and what do you actually get for it?

Kenya has carried the cheapest headline entry price in Africa, at roughly KSh 1,300 a month (about $10) — but that is a capped plan of around 50GB, not the unlimited residential product, and once the cap is exhausted speeds drop sharply until the billing cycle resets. Kenya is also the market to study if the antenna price is your obstacle, because it pioneered a hardware-rental option at roughly KSh 1,950 a month, which removes the upfront kit purchase entirely at the cost of a higher lifetime spend. At the other end, Eswatini has sat near $50 a month for standard residential service. When you compare two countries, compare the same product tier: an unlimited residential plan in one market against a capped entry plan in another is not a like-for-like Starlink price comparison.

Can I cancel, pause or return Starlink — and do I get my hardware money back?

Service is billed month to month with no minimum term in the consumer tiers, so cancelling is a self-service action in the account dashboard and stops the next invoice rather than triggering a penalty. Hardware is the part people get wrong. Starlink’s standard consumer return window is 30 days from delivery for a full refund of the kit, and a partial-refund window extends to roughly 12 months in most markets — but the clock runs from delivery, not from activation, which catches buyers who order ahead of an install date or who wait for a mount to arrive. Kits bought from a third-party reseller are usually outside that policy entirely and subject only to the reseller’s own terms, and you will not get import duty back from anybody. Roam plans can be paused between months, which is the intended pattern for seasonal or project use; residential plans cannot be paused in the same way.

What happens if my dish breaks in an African market — how does warranty and RMA actually work?

Consumer kits carry a limited warranty of 12 months in most regions (longer in some), covering manufacturing defects only. The friction in Africa is logistics, not policy: a replacement is shipped to a supported address, and where there is no local fulfilment the return leg and the inbound replacement can both attract freight and, in some cases, a second round of duty. Lightning damage, surge damage, rodent-chewed cable and water ingress through a badly sealed roof penetration are all excluded, and they are the three most common failure modes reported on the continent — which is why a $30 surge protector on the router side and a proper drip loop on the cable are cheap insurance against a $300 replacement. Keep the original packaging and your proof of delivery date; RMA requests without them stall.

Will Starlink work with a VPN, and can geofencing cut me off if the kit leaves its registered country?

Starlink is a plain IP transport, so commercial VPNs run over it normally, with the usual latency penalty of an extra hop. The real geofencing question is about the account, not the tunnel: a residential plan is bound to a registered service address in a licensed country, and a VPN does not disguise the terminal’s actual location, which SpaceX derives from the satellite link itself rather than from your exit IP. Terminals detected operating persistently outside their licensed region have historically been throttled, switched to a roaming rate, or deactivated. This is the specific risk facing imported kits in unlicensed markets: nothing breaks on day one, and then the policy changes. Treat any cross-border setup as a service you may lose with little notice, and do not build a business on it without a fallback link.

Can a business, school or clinic get Starlink when residential is sold out — and what does priority cost?

Usually yes. Prioritised and business tiers sit in a different allocation bucket from residential, so they frequently remain open in cells where consumer sign-ups have closed, and they buy a genuinely different product: prioritised data that holds up during the 7pm to 10pm crunch, higher upload allowances and a support path with a service commitment. The cost is materially higher — commonly several times the residential monthly fee, billed against a data allowance with an overage rate once it is used. For a lodge, a clinic running telemedicine or an office of ten people, that arithmetic still usually beats a leased line quote. For a household, it rarely does, and a Roam plan is the cheaper way into a closed cell.

Can I buy a Starlink kit in one African country and use it in another?

Only with the right plan. Residential service is bound to a registered address inside a licensed country and historically tolerates use away from that address for around two months before the account must be updated. Roam and Mobile plans are built for cross-border use and come in regional and global variants, at a higher monthly price and with lower network priority than fixed residential service. Two practical traps: a second-hand terminal may still be registered to another account or region and must be released by the original owner before you can activate it, and carrying hardware across a border can attract import duty on the kit’s declared value even though it is used equipment. Check the seller’s account status before money changes hands, not after.

Do I need a licence or have to pay duty to import a Starlink kit?

Duty and VAT apply in almost every African market unless the kit is fulfilled locally, and they are assessed on the declared hardware value plus freight — which is why landed costs routinely exceed the sticker price. The tax floor alone is significant before any duty line: Kenya applies 16% VAT and Zimbabwe 15%, with customs duty on top depending on how the terminal is classified. Separately, some regulators treat an unlicensed satellite terminal as prohibited telecommunications equipment regardless of who owns it, and kits have been confiscated at customs in markets where Starlink is not licensed. Before importing, confirm two things in writing: that Starlink is licensed in the destination country, and the HS code and duty rate your customs authority applies to a satellite user terminal. Then ask the reseller to itemise duty, VAT and freight separately from their own margin.