ISP Launch & Growth
From intent to operation, or from stagnation to growth — feasibility, licensing, number resources, network build-out, and the operational systems that make an ISP a business rather than an experiment.
Networks are built once. Operations are run every day.
01 — The Reality
Why ISP Launches Stall — and Established ISPs Stagnate
Most new ISPs do not fail on radio planning or fibre routes. They fail on everything around the network: licences applied for in the wrong order, IP resources requested too late, equipment bought on price instead of roadmap, billing bolted on after launch, and a team trained on whatever happened to be installed.
Established operators hit a different wall. The network that carried the first thousand subscribers — consumer-grade radios, flat topology, manual provisioning — quietly becomes the ceiling on the next five thousand. Growth stalls not for lack of demand but because the infrastructure and the organization were never designed to scale.
We work both ends of this problem. At SkyTrend Networks in Kenya, a CAK-registered ISP holding a Network Facilities Provider (NFP) licence engaged us to overhaul its infrastructure and plan nationwide expansion — we re-designed the fibre/wireless network and systems architecture, deployed and optimised it, and trained the technical team. At DSI in Eastern DRC, we carried out a full infrastructure overhaul for growth: a carrier-grade RF and core upgrade, IP and RF planning for residential expansion, AfriNIC public IP and ASN acquisition and origination, and training for both the technical and business teams.
02 — Groundwork
Groundwork: What Must Be Right Before Anything Is Built
The unglamorous work that determines whether the build succeeds
Feasibility & Architecture
- Market and coverage feasibility analysis
- Target network architecture (access, aggregation, core)
- Business architecture: pricing, cost structure, break-even
- Capacity and upstream transit planning
- Failure domains and redundancy from day one
Licensing & Regulator Engagement
- Licence-category mapping to the actual business plan
- Application preparation and technical documentation
- Spectrum and infrastructure obligations review
- Compliance posture designed in, not retrofitted
AfriNIC Number Resources
- AfriNIC membership and justified IPv4 requests
- IPv6 allocation and deployment planning from the start
- ASN acquisition and BGP origination
- Addressing plans that survive growth, not just launch
Technology & Vendor Selection
- Access, backhaul, and core platform evaluation
- Consumer-grade to carrier-grade migration paths
- Licensing, support, and spare-parts economics
- Interoperability and vendor-lock-in exposure
At DSI, this discipline meant migrating the access and core estate from Ubiquiti to Cambium and Cisco — a deliberate step-change in reliability, planned around budget reality.
03 — Build vs Lease
Build, Lease, or Both
Infrastructure economics decide more launch outcomes than technology does
Own-Build
Building your own fibre or wireless plant maximises margin and control — and capital exposure. It makes sense where you own the customer relationship long-term, where no credible wholesale option exists, or where the asset itself is the business.
Lease & Open-Access
Across the continent, open-access and wholesale fibre operators now let ISPs reach customers without carrying the full build cost. Leasing shifts capital into customer acquisition and operations — at the price of margin and dependency on someone else’s repair clock.
The Hybrid Answer
In practice, most viable plans mix the two: build where density and ownership justify it, lease where speed to market matters, and keep the design clean enough that segments can move from leased to owned as revenue proves out.
The right answer is a spreadsheet and a topology diagram agreeing with each other — not an ideology.
04 — Engagement
How a Launch or Growth Engagement Runs
- 01
Baseline & Feasibility
For a new ISP: market, coverage, and financial feasibility. For an existing one: a Phase 0 review of what actually exists — network, systems, and operations — before anything is proposed.
- 02
Licences & Number Resources
Regulator engagement and licence applications in the correct order, alongside AfriNIC membership, IP justification, and ASN acquisition — the long-lead items that stall launches when started late.
- 03
Design & Vendor Selection
Target architecture for access, core, and operational systems; structured vendor evaluation; and a phased bill of materials matched to the funding plan.
- 04
Phased Build-Out
Access, core, and systems deployed in phases that each carry revenue — with BGP origination, monitoring, provisioning, and security hardening landing with the network, not after it.
- 05
Launch Operations
NOC workflows, billing and AAA integrity, and ticketing live from the first paying customer. An ISP that launches without these is accumulating invisible debt.
- 06
Growth & Handover
Expansion planning, additional revenue lines, and — deliberately — training the technical and business teams so the operation runs without us.
Every phase is designed to leave the ISP more capable of running itself than the phase before.
05 — Day One
What Must Be Live on Day One
Launch day is an operations event, not a network event. Before the first customer pays, these must already work:
- Monitoring and alerting that sees the whole network — with someone responsible for watching it
- Billing, AAA, and provisioning that agree with each other, so every active subscriber is a billed subscriber
- A ticketing and escalation path from the first support call
- Configuration backup and change tracking on every device that matters
- Security hardening of the management plane before it is exposed to the world
- Documented runbooks for the failures you already know will happen
At DSI we built exactly this stack — planning, monitoring, and ticketing systems, with redundant systems on AWS EC2 — so growth did not outrun visibility.
06 — Growth Levers
Growth Levers Beyond Bandwidth
Connectivity is the entry product, not the whole business
Virtualization & Hosted Services
An ISP already owns power, cooling, connectivity, and customer trust — the ingredients of a local hosting business. At DSI, server virtualization served internal systems and was leased to customers, cutting power costs while opening a recurring revenue line. Local hosting also answers a demand public clouds cannot: data that stays in-country.
Expansion Planning
Growth is an engineering exercise before it is a sales exercise: IP and RF planning for new residential zones, capacity headroom in the core, and build phases sequenced so each new area carries its own economics — the same discipline behind the residential expansion planning we delivered at DSI and the nationwide expansion plan at SkyTrend.
The Team as Infrastructure
Equipment depreciates; a trained team compounds. We train technical teams on the network they actually run and business teams on the systems that bill for it — as we did at both SkyTrend and DSI. From the BSS era onward, our work has centred on product development and technical capacity enhancement for ISPs across West, East, and Central Africa, backed by partnerships with Cisco, MikroTik, Cambium Networks, Microsoft, Google Cloud, and AWS.
A licence makes you legal. An ASN makes you a network. Operations make you a business.
In 2026, AfriNIC is effectively the last registry where justified IPv4 requests can still be met from a free pool — but the soft-landing exhaustion phase means new operators must plan for smaller allocations and treat IPv6 as launch scope, not a someday project. Meanwhile, the spread of open-access wholesale fibre across African markets has changed launch economics: the credible question is no longer "can we afford to build?" but "which segments should we build, and which should we lease?" — while data-sovereignty demand is making locally hosted services one of the few high-margin revenue lines an ISP can add without new spectrum or new trenches.
Launching an ISP — or Restarting Its Growth?
Start with a factual baseline of your market position, licence path, and technical estate before committing capital.