Democratic Republic of Congo, Central Africa
At the United States–Congo Investment Forum in Washington, Bob Mabiala Mvumbi, who heads the Democratic Republic of Congo's Inga Development Agency, made an argument that would have sounded fanciful five years ago and now reads as commercially serious. The Inga site on the lower Congo River, he told an audience of investors, is the best place on earth to put an artificial intelligence data centre. He listed the reasons in the order a hyperscaler would: abundant hydroelectric capacity, fibre already in the vicinity, and a river that solves cooling.
The physical case is hard to dispute. The Inga complex currently generates under 2 gigawatts. Its assessed potential is around 44 GW — roughly twice the output of China's Three Gorges Dam, and more firm, low-carbon capacity than the entire installed base of most African countries. Kinshasa is drafting a dedicated Inga Law to give private investors a defined regulatory and fiscal framework, covering design, construction and operation. The World Bank has committed $1 billion to advancing the site, with a first $250 million tranche expected to disburse this year. It is a genuinely attractive pitch. It is also the second entry in a ledger, and the first entry is considerably less flattering.
The domestic ledger: 21.5 percent
The DRC's national electrification rate stands at about 21.5 percent. The disaggregation is starker than the headline: roughly 45.4 percent in urban areas and about 1.1 percent in rural areas. Under its National Energy Compact, the government has committed to lifting national access to 62.5 percent by 2030 — a target that would require connecting tens of millions of people in under five years. Meanwhile the state utility SNEL delivers approximately 2,100 MW against an installed capacity of around 2,800 MW, a 25 percent gap that reflects deferred maintenance and transmission losses rather than any shortage of water.
So the country offering 44 GW to the global compute industry currently cannot reliably move 2.8 GW to its own population. That is not a rhetorical gotcha — large hydro for export or industrial offtake and distributed access for households are genuinely different engineering and financing problems, and solving one does not automatically solve the other. But it is the central political economy question hanging over the Inga pitch, and it has a precedent in the same country's minerals sector.
What the telecom numbers actually show
Congo's connectivity picture is improving quickly on some measures and barely at all on others. According to the Autorité de Régulation de la Poste et des Télécommunications du Congo (ARPTC), the country closed 2025 with 73.9 million active mobile subscriptions, a penetration rate of 65.9 percent. Mobile internet penetration, however, sat at roughly 33 percent, and overall internet penetration remains below 30 percent — leaving something close to seven in ten Congolese without a data connection.
The commercial composition of the sector tells you where the growth is. Total telecom revenue reached $2.394 billion in 2025, and data now accounts for about 53.8 percent of market revenue, against nearly 14 percent in 2016. In under a decade the industry has flipped from a voice business to a data business. That is the classic precondition for a domestic digital economy: operators now have a direct commercial interest in getting people online, because that is where their margin is.
Money moves faster than data
The most striking divergence in the Congolese stack is between payments and connectivity. ARPTC recorded roughly 34.2 to 34.3 million active mobile money subscriptions at the end of 2025 — depending on the counting basis, a penetration rate of between 30.6 and 34.8 percent. That is broadly level with mobile internet penetration, which is unusual: in most markets, financial adoption trails connectivity by years.
In cities the figure is far higher. A June 2026 study by Target Sarl found that 77 percent of respondents in major Congolese cities reported owning a mobile money account. The binding constraint at that level of adoption is no longer access but fragmentation — multiple closed-loop operator wallets, limited interoperability, and the transaction friction that comes from a market where moving value between providers is harder than moving it within one. This is a solvable regulatory problem, and it is a more immediately valuable one than any data centre: an interoperable retail payments layer would compound across every sector of the economy, including the informal one.
The satellite shortcut
Congo's geography — 2.3 million square kilometres, extensive rainforest, limited road network — makes terrestrial coverage economics brutal outside the corridors. The country has become the continent's testbed for the alternative. Starlink went live in the DRC in 2025, a month after receiving its operating licence. Then, on 18 August 2026, Airtel Africa and Starlink commercially launched satellite-to-mobile service in the DRC — the first commercial deployment of Starlink Mobile anywhere in Africa.
The proposition is that a compatible smartphone and a clear view of the sky are sufficient, with no additional equipment. It follows Airtel Africa's December 2025 agreement with SpaceX to roll out direct-to-cell across all 14 of its African markets by the end of 2026, and trials in Kenya in March 2026, Uganda in May 2026 and Madagascar in July 2026. Airtel DRC managing director Thierry Diasonama framed the launch around the country's size and geography; the named target users are transport and logistics operators, humanitarian organisations, health workers, farmers, mining operations and communities outside terrestrial reach. Eligible customers get a 30-day free trial via the MyAirtel app.
Direct-to-cell is not a broadband substitute — the bandwidth supports messaging and light data, not streaming or cloud work. But as a continuity layer it changes the risk profile of operating outside the cities, and the DRC being chosen as the first commercial market is a meaningful signal about where the operator believes the unserved demand is. It also illustrates a pattern worth naming: Congo's fastest connectivity gains are coming from capital and infrastructure it does not own and cannot direct.
The minerals precedent
Which brings the argument back to Inga, by way of cobalt. In 2026 the DRC set an annual cobalt export quota of 96,600 tonnes — 87,000 tonnes distributed to producers pro rata, with 9,600 tonnes retained under the discretionary control of ARECOMS, the state minerals regulator, for projects deemed nationally significant. That works out to roughly 7,250 tonnes per month, approximately half the average monthly export volume recorded in 2024. ARECOMS has signalled it will maintain the quota regime through 2027, with further reductions possible if the market requires rebalancing.
Whatever one thinks of supply management as price policy, the quota is evidence of something new: a Congolese state willing to use regulatory leverage over a resource the world needs, rather than accepting price-taking as a permanent condition. Cobalt is the input to the batteries in the devices; Inga would be the input to the compute those devices talk to. The question the Inga Law will answer is whether Kinshasa applies the same logic — negotiated terms, retained national share, conditions attached — to electricity, or whether it competes for hyperscaler investment on price alone.
What would make Inga a development story
There is a version of this in which the two ledgers reconcile. A large anchor offtaker at Inga — even one whose racks serve users on other continents — underwrites transmission investment, creates a creditworthy revenue stream against which distribution can be financed, and establishes the DRC as a location where serious digital infrastructure operates. Fibre laid for a data centre serves whatever is built near it. That is roughly how the Gulf states converted energy endowment into logistics and then services.
There is also a version in which it does not. Enclave infrastructure is the default outcome in the DRC's history, not the exception: generation dedicated to mining concessions, roads built to ports, value captured at the point of export. A data campus at Inga could easily become another enclave — one that raises national statistics without changing whether a household in Kasaï has light after sunset or a data connection at all.
The distinguishing variable will be whether the Inga Law contains domestic supply obligations, local grid co-investment requirements and transparent pricing — or whether it contains only incentives. On current evidence, the DRC is building the export layer of a digital economy considerably faster than the domestic one. Both ledgers are open. Only one of them is being pitched in Washington.
