Ethiopia's software market, in numbers

Every pitch deck about Ethiopian tech opens with a big number. The trouble is that the big numbers disagree with each other — sometimes by tens of millions of people. The Ministry of Innovation and Technology puts internet users at over 42 million8; DataReportal, working from operator and ITU data, puts them at 29.5 million1. Both are published. Only one can be the planning assumption.

So this piece does something narrow: it takes the figures we can source to a primary publisher, shows them plainly, and says what each one does and doesn't support. These are the numbers we use when we scope a product for this market.

Connections are not people

The single most misread statistic in this market is the mobile connection count. Ethiopia had 93.2 million active cellular connections in October 2025 — 68.4% of the population1. That number gets rounded up in conversation to "most Ethiopians are connected." It doesn't mean that. Connections count SIM cards, and one person can hold several.

The gap between the first bar and the second is the whole story of building software here. Coverage is not the constraint — GSMA finds that 76% of Ethiopians live inside mobile broadband coverage and still do not use mobile internet6. The constraint is devices, cost, literacy, and language. A product that assumes a smartphone, a data bundle, and English is addressing the third bar, not the first.

The line that actually moves

Internet-user estimates for Ethiopia are unstable enough that plotting them as a trend would be dishonest: the same publisher reported 29.83 million users in January 2022 and 20.86 million in January 20232, a revision in method rather than nine million people logging off. Connections, which come off operator reporting, are consistent year to year. So that is what we plot.

Four years, forty million connections. That is the demand curve worth designing against — not because everyone on it is a smartphone user, but because it is the population that can be reached at all, today, on the network as it exists.

Payments arrived before the internet did

The most striking thing in Ethiopian digital infrastructure is that financial rails have outrun general internet adoption. telebirr closed the 2025/26 fiscal year with 60.6 million customers, 2.61 billion transactions worth ETB 4.19 trillion, and a network of 410,700 agents and 440,100 merchants3. For the first time, data and internet services — 31.1% of revenue — overtook mobile voice as ethio telecom's largest revenue line.

The direction of travel is set by regulation as much as by demand. Phase one of the National Digital Payments Strategy took mobile banking accounts from 9.1 million to 54 million; phase two, running to 2030, targets full interoperability, digital-ID integration, and merchant acceptance among small businesses7.

What the money says

GSMA sizes Ethiopia's digital economy at ETB 700 billion in 2023, projected to pass ETB 1.3 trillion by 2028, adding over a million jobs and ETB 57 billion in tax revenue along the way6. Its projection assumes mobile internet users roughly double to over 50 million and the usage gap narrows from 76% to 66% — which makes the whole forecast a bet on devices and affordability, not on network build-out.

• The forecast is conditional — it depends on closing a usage gap, and the 40% gender gap inside it is the hardest part.
• The rails are ahead of the users — payments infrastructure is national; general internet use is not.
• Compliance is now real — the Personal Data Protection Proclamation No. 1321/2024 has been in force since July 2024, with 72-hour breach notification and restrictions on transferring personal data abroad.9

What we take from this

Four working assumptions come out of the data above, and they shape how we scope every product for this market.

1. Design for the connection, not the smartphone — low-end Android, small payloads, and a path that still works when the data session drops.
2. Treat payments as infrastructure, not a feature — telebirr and bank rails are where the users already are, so integration is a starting assumption.
3. Amharic first, not Amharic later — language is one of the barriers holding the usage gap open.
4. Assume the data law applies to you — breach response and data residency are architecture decisions, and they are cheaper made early.

If you are planning a product for this market and want to pressure-test the assumptions behind it, get in touch.