Who Holds the Valve? Water, War and the Price of Peace -By Kator Ifyalem

On 1 October, Ethiopia ordered an Egyptian diplomat out of the country within 48 hours. Cairo answered the same day with a mirror-image expulsion of an Ethiopian embassy counsellor, and neither side publicly named a specific violation. The diplomats were incidental. The cause was a river.

Water now sits at the centre of African security politics. The pressure comes from three directions at once, dams being built faster than agreements are written, power stations being destroyed in wartime, and a continent where roughly 60 per cent of water resources cross a border. Rivers run through countries that signed different treaties, in different decades, with different armies behind them.

The Nile (a dam finished, an agreement not): Ethiopia inaugurated the Grand Ethiopian Renaissance Dam on 9 September 2025. It is designed to generate 5,150 megawatts, more than double the country’s previous capacity. For a country of over 120 million people with chronic power shortages, that is a development argument, and a serious one.

The counter-argument from Egypt is arithmetic. Irrigation Minister Hani Sewilam said in September that Egyptian per capita water availability has fallen to about 472 cubic metres a year, below the 500 cubic metre line used to define absolute scarcity. He put annual renewable supply at roughly 56.8 billion cubic metres against a need of about 123 billion. A country that short has no slack to absorb operating decisions from another country.

That is the real dispute, and it is a technical one. Cairo and Khartoum want a legally binding agreement on how the dam is filled and operated, while Addis Ababa insists on guidelines. The sticking points are the annual refilling, how much water Ethiopia releases downstream in a multi-year drought, and how future disputes get settled.

Anyone who has worked around reservoirs knows why those three items matter more than the headline storage volume. A dam is a valve. Whoever sets the release rule during the third dry year in a row holds the harvest of a downstream country in their hands. A guideline is a courtesy. A binding rule with a verifiable data feed is a protection.

The 2025 flood season showed what happens without one. In early October 2025, Egypt accused Ethiopia of reckless management, claiming abrupt releases after the inauguration produced a “man-made flood” that hit farmland in Sudan and the Nile Delta. Ethiopia replied that the dam had prevented something worse and pointed to other causes, including White Nile inflows, shifting rainfall and conflict-damaged infrastructure. Both accounts may hold some truth. The point is that nobody could settle it from shared, trusted, jointly audited flow data, because that data system does not exist.

2026 (diplomacy that kept widening the field): The year has been one of offers and escalations. In January, President Trump wrote to President Sisi offering to restart US mediation, and Egypt and Sudan both welcomed it. No settlement followed. In August, Ethiopia’s water and energy minister confirmed plans for three more hydropower dams on the Blue Nile system. They are Karadobi, Mandaya and Beko Abo. On 29 September, Egypt’s foreign minister told a meeting chaired by President Sisi that Cairo would not accept any new dams or a repeat of what happened with the current one.

Then the water dispute merged with the wider security picture. Ethiopia’s army chief accused Sudan, Egypt and other external powers of backing the forces fighting his troops, and Egypt rejected the claim. On the same day as Addis Ababa’s move against Egypt, it expelled ten Eritrean diplomats and closed its embassy in Asmara, and Eritrea cut diplomatic ties in response. On 4 October, President Sisi hosted the Eritrean and Somali presidents and Sudan’s army commander at New Alamein. Their joint declaration backed international law on transboundary rivers, opposed unilateral action affecting Egyptian and Sudanese water supplies, and asserted that Red Sea governance belongs to coastal states.

This is how water disputes mutate. A river argument that could have stayed with hydrologists and treaty lawyers has become a bloc-building exercise across the Horn. Once a technical question becomes a coalition question, compromise gets harder, because every concession is now read as a loss of face to somebody’s allies.

Sudan (when the infrastructure is the target): The Nile dispute is a threat. Sudan is already the proof of what comes next. Since the war began in April 2023, drones have turned power and water infrastructure into a battlefield. In January 2026, UN human rights chief Volker Türk toured the Merowe Dam power plant, the largest hydroelectric facility in Sudan, after repeated drone strikes had crippled national supply. Sudanese officials described the damage as a deliberate attempt to paralyse essential services. The United Nations reports that the conflict has devastated water and energy infrastructure in 13 states.

The link between a transformer and a tap is direct, and it is an engineering fact. Municipal water runs on pumps, pumps run on electricity, and treatment plants need continuous power to keep chlorination and filtration working. Hit the substation and the water system fails without a single shell landing on a pipe.

El Obeid shows the consequence. By July 2026, drone strikes on power infrastructure had cut electricity to the main pumps in the city. Its southern water basin had been repeatedly struck while the northern one remained under paramilitary control, and residents reported that the price of a jerrycan had doubled in a matter of weeks. The Sudan International NGO Forum warned that unsafe water sources could trigger waterborne disease outbreaks such as cholera. Aid agencies have already linked earlier outages to forced use of untreated river and shallow well water.

International humanitarian law is not silent here. It protects objects indispensable to civilian survival, and it places specific restrictions on attacks against works containing dangerous forces, which includes dams. The law exists. Enforcement does not. Nobody has yet paid a visible price for hitting a water system, and that is why the pattern is spreading.

Luanda (a forum with a good theme and a hard test): From this, the timing of the Biennale of Luanda is worth noting. Its fourth edition runs on 22 and 23 October 2026, with the theme of strengthening water governance in Africa as a tool for conflict prevention, mediation and resolution. President João Lourenço is due to give the opening address, alongside African Union Commission Chairperson Mahmoud Ali Youssouf and UNESCO Director-General Khaled El-Enany. A panel on water, peace and transboundary cooperation will examine water diplomacy, regional integration and a culture of peace.

The reframing is right. Shared water can be the cheapest place to build trust between states, because the interests are concrete and the data can be measured. But a forum can only produce statements. What would make Luanda useful is a short list of deliverables that can actually be checked.

The first is a continental norm that treats water and power infrastructure as protected in conflict, with a monitoring body that publishes violations. The second is a commitment to real-time hydrological data sharing across the major basins, which is cheap, technical and politically survivable. The third is binding operating rules for large dams in drought years, agreed before the drought arrives. Everything else is a communiqué.

The Niger Basin (the same story, slower): West Africa is not the Nile, but the structure is familiar. The Kandadji dam on the Niger River in Niger was meant to regulate flow, irrigate roughly 45,000 hectares and generate power. The project is expected to relocate about 50,000 people from 24 villages. Construction began in 2019 with a first phase completion previously expected in 2025 or 2026. It has been delayed.

On 10 July 2026, after a three-day national forum in Niamey, Prime Minister Ali Mahaman Lamine Zeine closed proceedings with recommendations that frame three restart options, reorganising cooperation with donors, mobilising more military engineering, and phasing funding from national resources. The government has designated Kandadji as the top energy priority for the country.

Nigeria should be paying attention. The Niger flows through the heartland of Nigeria and upstream regulation changes what arrives at Kainji and Jebba. Nigeria also knows the other side of the problem from the Benue. In 2022, Nigeria lost more than 600 people in its worst flooding in a decade, after heavy rain and releases from the Lagdo dam in Cameroon, and experts said the failure to complete a Nigerian dam meant to backstop Lagdo made the disaster worse. Abuja has since moved the Datsin Hausa dam in Adamawa State into a public-private partnership structure with a transaction adviser. Early warning, release protocols agreed with upstream operators, and storage of our own are the three things that separate a managed flood from a disaster. We have been slow on at least one of them.

What the pattern says: Four things connect these cases. The first is that volume is rarely the real fight. The Nile carries enough water for a negotiated deal. What is missing is a verifiable operating rule and a trusted data feed. That is an institutional deficit before it is a hydrological one. The second is that infrastructure outruns agreement. Dams get built on national timelines, and treaties get written on political ones. The gap between them is where distrust accumulates. The third is that war has made water systems targets. When a power station is cheap to strike and costly to repair, and nobody pays for striking it, strikes will continue. The fourth is that diplomacy follows capacity. The alignment of the Red Sea in Egypt and the pursuit of sea access by Ethiopia show what happens when a water dispute has no technical forum strong enough to contain it. It spills into whatever forum is available, and usually a worse one.

Peace over water is possible, and the building blocks are unglamorous, gauging stations, data protocols, drought year release tables, protected infrastructure norms and enough money to maintain them. None of that will trend on social media. All of it is cheaper than the alternative. Luanda opens next week. The test is whether it produces something an engineer could audit.