Solutions — ke-1a Nairobi, tz-1a Dar es Salaam, ng-1a Lagos
For companies whose customers are there and whose servers are not
You do not need a cloud in these three countries. You need the part of your system that touches users or personal data to be inside one of them, and the rest to carry on where it already works. That is what ke-1a, tz-1a and ng-1a are for: an application tier, a read replica, a webhook receiver, an object store, or a whole product for one market — deployed from the same account and API as everything else, billed in dollars, and held under the law of the country you chose. Cloud servers from $13.89 a month.
For startups
Launch for $13.51, pay in local currency, and scale into the full stack when you win.
For startupsFor agencies
Run every client on one platform, white-labelled, with margin you keep.
For agenciesFor enterprise
Reliability, dedicated hardware, and data that stays in-country, under SLA.
For enterpriseColocation
Bring your own hardware into our Tier III facilities.
ColocationWho deploys here from somewhere else
Eight ways a company outside these countries ends up inside them
Described by what the system does all day rather than by the industry on the pitch deck. Find the one that looks like your quarter.
SaaS with enough customers in one market to notice
You run in Frankfurt, Ireland or northern Virginia, and your Nairobi or Lagos accounts have started mentioning the loading time in renewal calls. The answer is almost never a migration. It is a regional tier: application servers and a read replica in-country, session and cache beside them, background workers next to the queue they drain, the control plane and the warehouse left where they are, and DNS sending each user to the nearer origin. Two instances and a replication link. A cloud server is $13.89 a month, so the pilot costs less than the meeting about it.
Fintech and payments entering a market
Payment API endpoints and webhook receivers taking callbacks that originate inside the country, ledger and reconciliation databases, KYC and document stores, fraud scoring, and transaction logs that only ever grow. Nigeria is the sharpest case: a Central Bank circular of 15 June 2026 directs that payment transaction data generated in Nigeria be stored and managed there from 1 January 2027, across banks, mobile money operators, fintechs, switching companies, payment service providers and super agents. In Kenya the mobile-money callback originates on a Kenyan network with a retry policy pointed at your endpoint. Bare metal from $309 a month where the ledger wants single tenancy.
Development programmes and donor-funded work
Health information platforms, form servers taking submissions from enumerators on mobile data, file sync for staff working days from a city, monitoring dashboards, grant accounting. Two things decide this purchase and neither is price. Beneficiary records are personal data, so the country the database sits in is a board question — and in Tanzania, moving that data out later needs a permit from the regulator rather than a form. The other is the calendar: annual billing at ten months for twelve gives one approval, one payment and one document for the grant file.
Media and streaming reaching local audiences
Media is origin storage plus egress, the shape distance punishes hardest, and a cache only helps if there is one nearby. The large content networks publish African edge locations in Cape Town, Johannesburg, Nairobi, Lagos and Mombasa — none in Tanzania, so a Dar es Salaam viewer is served from another country however the CDN is configured. Catalogue and entitlement services, transcoding pipelines, and an object store the cache fills from next door. S3-compatible storage with free transfer and API requests is $0.061 per GB a month, $0.035 at scale.
Agencies, integrators and ISVs with clients in-region
Your client's tender says the data is held in-country and your architecture diagram says Frankfurt. Multi-tenant client hosting, per-tenant environments provisioned through the API, staging, Git-based deploys and an in-country address so sites resolve locally. One account covers all three countries — one supplier, one API and one renewal date rather than three of each — and a second client in the same country is a script rather than a procurement exercise.
Meeting a customer's data-residency requirement
Sometimes the region is chosen by someone else entirely: a customer's data processing agreement, a bank's onboarding questionnaire, a security review, or a tender clause that disqualifies you in the first round. The work is to identify what actually has to be in-country — usually the personal data and the records that reference it, rarely the whole platform — deploy that tier into the matching region, and write down what is in it. That list is what you will be asked for, repeatedly, by people who did not read the diagram.
Enterprises consolidating what they already have
A rack a local partner manages in Lagos, a server under a desk in a Nairobi office nobody has rebooted since the last person left, and a Tanzanian shared-hosting account whose login lives in a former employee's inbox. Three regions on one account, one console and one invoice replace three suppliers and three renewal dates you keep discovering by email. The Nigerian half carries a cost argument: the grid collapsed in December 2025 and again in January 2026, and a machine in your own building is kept alive by a generator and a fuel account.
Companies with people and equipment on the ground
Mining and energy contractors, freight forwarders, engineering firms, survey and inspection businesses — foreign-owned, operating inside the country, generating data faster than a thin link can move it abroad. Telematics ingest from trucks on Tanzanian roads, site-to-cloud aggregation over satellite links, customs and port middleware, geological stores that grow by the survey, safety systems holding worker records. The upload path stops crossing an ocean, and the records stay in the country the workers are in.
Cloud servers from $13.89/month, Linux VPS from $21.95, Windows VPS from $31.95 with the licence included, managed Kubernetes from $27.86, dedicated LineServe Core hardware from $309 plus a one-time $99 setup, object storage from $0.061 per GB, and colocation quoted per footprint.
Network
The three countries are not interchangeable, and the differences change your design
Distance is charged per round trip. A brochure page makes a handful and the reader forgives all of them; a checkout makes dozens in series at the moment a customer decides whether to trust you with money; a webhook receiver has an operator's retry policy pointed at it and a reconciliation job waiting behind it. The workloads that make thousands of round trips are where an in-country origin stops being a preference and becomes the design. But which country you deploy into changes what else you have to think about, and treating the three as one region with three names is the mistake that surfaces later.
Kenya's international capacity lands at Mombasa and is carried roughly 480 km inland to Nairobi, while domestic traffic is exchanged inside the city at KIXP — 136 peer networks and 2.9 Tbps on PeeringDB, with the large content networks present. Lineserve peers there. The audience is enormous and mobile: 84,090,298 active SIMs and 52,852,505 mobile broadband subscriptions in the quarter to March 2026. Mobile money penetration passed 100% of the population, which is why so many Kenyan integrations are callbacks arriving from inside the country rather than requests going out of it.
Tanzania is the one with a genuine hole in the map. Its cables come ashore in Dar es Salaam itself, so there is no inland haul, and 13,820 km of national backbone fibre carries a single origin to Mwanza, Arusha, Dodoma and Mbeya over domestic infrastructure. What the country has no trace of is a content-network edge, and the nearest cloud region is in southern Africa — so for a Tanzanian user a cache in Nairobi or Mombasa is a foreign hop and a border crossing for the data. TCRA counted 111.9 million subscriptions across four mobile networks with none above 32.3%, so no single carrier relationship gets you most of the country.
Nigeria is the largest audience and the densest connectivity: eight subsea systems ashore in and around Lagos, a national exchange running 13 points of presence across seven states with more than 130 connected networks, and 189.68 million active mobile subscriptions as of May 2026, MTN and Airtel holding roughly 86% of lines. It also has the shortest path from a technical decision to a commercial one, because power is a real cost line — the grid collapsed in December 2025 and again in January 2026, and twenty-four companies listed on the Nigerian Exchange spent ₦400.83 billion on diesel, gas and alternative energy in the first quarter of 2026 alone.
84.1M
Active SIMs in Kenya (CA, Q1 2026)
111.9M
Telecom subscriptions in Tanzania (TCRA, Q1 2026)
189.7M
Active mobile subscriptions in Nigeria (NCC, May 2026)
0
Hyperscale cloud regions in the three countries
Some workloads have no meaning outside the country
A RADIUS server for a Nairobi WISP, a resolver for Tanzanian subscribers, a USSD gateway backend, a NetFlow collector watching Lagos links. If your product is sold to networks in these countries, hosting it elsewhere is not a compromise but a defect — and the buyer runs a traceroute before reading a word of marketing, which is the correct order.
Invoicing
One supplier, one dollar invoice, three countries on it
Buying from outside the three countries, you are quoted, invoiced and settled in US dollars by card or bank transfer, on one account however many regions you use. That is deliberately unremarkable: $13.89 on the pricing card is $13.89 on the invoice, from a dollar price list rather than a conversion applied the moment you click. Displayed prices exclude tax, and where VAT applies it is calculated and shown separately at checkout. Annual billing runs at ten months for twelve on eligible plans — the shape most useful when spend has to be committed inside a budget window.
The alternative is worth knowing about before you order rather than after. If a subsidiary in one of the three countries should be the entity paying, it can hold its own account in that country's market and settle in KES, TZS or NGN, with the country's own tax treatment following, and each price list is set per currency rather than converted. Send [email protected] the registered names, the tax identifiers for any entity that will appear on an invoice and any purchase-order reference your system needs, and the account is set up once instead of corrected later.
Colocation and larger dedicated builds are quoted rather than listed. Send rack, power, connectivity and remote-hands requirements to [email protected] and you get a written quote against them.
Data residency
It matters enormously to some of these buyers and not at all to others
Sort your own case into one of three buckets before designing anything. The first is mandate: Nigeria's Central Bank has directed that Nigeria-generated payment transaction data be stored and managed in Nigeria from 1 January 2027, and Kenya's Regulation 26 wants in-country processing, or at least one serving copy held in Kenya, for six named purposes including basic education, primary and secondary health care and public finance administration. Inside one of those, the region is a deadline rather than a design choice.
The second is friction. Tanzania's Personal Data Protection Act, 2022 puts transfers of personal data out of the country behind a permit from the Commission, applied for with the recipient, the data, the purpose and the destination security arrangements — prior authorisation that finishes when a regulator says so. Kenya's sections 48 and 49 and Nigeria's section 41 are lighter but still something to choose, paper and defend. Hold the data in-country and none of that machinery has anything to act on for records that never leave.
The third is commercial, and it is where most international buyers live. No regulator is asking, but a customer is: a data processing agreement, a bank's onboarding form, a security review, a tender clause. Answering with the name of a city closes that conversation instead of opening one. What a region supplies is location — your data held in Nairobi, Dar es Salaam or Lagos and moved only when you move it. The controller obligations stay with you, in whatever country the server sits.
Hosting abroad does not move the obligation
Kenya's Data Protection Act reaches a controller neither established nor ordinarily resident in Kenya that processes the personal data of people located in Kenya. A company in Amsterdam or Austin with Kenyan users sits inside that Act whether the servers are in Frankfurt or Nairobi. What changes when you deploy in-country is that the cross-border transfer analysis stops applying to the data that stays put.
Getting there
Start with the country that is asking
The pilot: one country, one tier
Almost every deployment on this page began as one instance in one region, running alongside whatever was already there. Keep your existing cloud for everything that works, and put the in-country tier underneath the part that has to be local: the user-facing servers, the read replica, the webhook receiver, the object store your customer's regulator asks about. An IPsec or WireGuard tunnel between the two, replication across it, DNS routing each user to the right origin. Nothing has to be migrated, and you end up with a region in a country where your current provider does not have one.
Then decide whether the other two follow
The second and third countries are the same configuration with a different region code, on the same account and invoice — which makes it tempting to do all three at once and then discover only one was ever needed. Let demand decide. Two rules save most of the rework: put the origin where the audience is rather than at an imagined midpoint, since traffic between African cities frequently transits Europe; and treat any copy of personal data in a second country as a cross-border transfer under the first country's law, because that is what it is.
Migration planning and assistance are included at no extra charge, whether you are coming from a hyperscaler region on another continent, a colocation cage you have outgrown, or a local provider in one of the three countries. Tell [email protected] what you run and where, and the reply is an architecture rather than a quote.
FAQ
Questions, answered
Yes, and most customers on this page are in exactly that position. You are quoted, invoiced and settled in US dollars by card or bank transfer, and you choose the region per instance. Switching to the Kenya, Tanzania or Nigeria market changes the currency to KES, TZS or NGN.
No, and it is rarely the right answer. The usual shape is a regional tier — application servers, a read replica, a webhook receiver, an object store — alongside whatever you already run, joined by a tunnel and split by DNS. Keep the control plane, the analytics and the CI where they work.
The one your users and your data belong to: Kenyan in ke-1a, Tanzanian in tz-1a, Nigerian in ng-1a. If you serve two of the three, run two regions rather than picking a midpoint — traffic between African cities frequently routes through Europe, so a compromise location is a compromise for both audiences. Per-unit prices are identical everywhere.
Usually less than the first draft assumes: the personal data and the records that reference it, the endpoints receiving callbacks from inside the country, and whatever a contract or a regulator has named. Machine images, build artefacts, aggregate analytics and telemetry stripped of personal data generally do not. Writing that boundary down is the useful hour.
It supplies the location, which is the part a hosting provider can supply: the data is held in Nairobi, Dar es Salaam or Lagos, under that country's law, and it is not moved out without your instruction. Whether your architecture as a whole meets the clause is a reading your own counsel does — and the region is what makes the answer possible at all.
A Central Bank of Nigeria circular of 15 June 2026 directs that payment transaction data generated in Nigeria be stored and managed there from 1 January 2027, reaching banks, mobile money operators, fintechs, switching companies, payment service providers and super agents. ng-1a in Lagos is the location input; how your systems and retention periods map onto the circular is for your compliance function. Start early — the date is fixed and the architecture work is not.
For static assets in Kenya and Nigeria it helps. For Tanzania there is no edge in the country at all, so a Dar es Salaam viewer is served from abroad whatever you configure. And a cache never runs your database, receives your callbacks or holds your customer records — which is the half of the problem that decides where the region goes.
Yes: one account, one API, one USD invoice, with the region chosen per instance and object storage on per-region endpoints. If you would rather each subsidiary held its own account in its own currency so the cost lands in the right ledger, that works too — set it up before you order rather than after.
Displayed prices exclude tax and where VAT applies it is calculated and shown separately at checkout. The domestic rates differ — Kenya 16%, Tanzania 18% on the mainland, Nigeria 7.5% — and which one reaches your invoice follows which market bills you, not which region your instance runs in.
Only for the part that belongs here. Put the origin where the audience is, and use an in-country region for the tier that serves local users, receives local callbacks or holds data a contract or a regulator has placed in that country. A two-region layout is a deployment decision on one account, not a second supplier relationship.
Owned hardware is colocation, quoted per footprint rather than listed, because rack units, power, connectivity and remote hands vary too much for a price card; colocated kit can sit on private networking beside cloud instances in the same region. Customer references are available under NDA — tell [email protected] what you are building and which country it is for.