Nigeria is making its next economic error, and this time it is invisible. For generations, the nation exported crude oil while importing refined fuel, losing billions in the process. Now it is repeating the same pattern with data.
The first commercial oil well in Nigeria was discovered in 1956 at Oloibiri, a remote creek in the Niger Delta. The well produced 5,000 barrels per day, and by 1958, Nigeria began shipping crude overseas. What returned decades later was petrol, diesel, and jet fuel—refined abroad and priced in foreign currencies. In 2022, the country spent $23.3 billion importing petroleum products derived from its own crude. Even Oloibiri, the site of the first strike, lacked reliable electricity for years, relying on solar mini-grids long after the oil revenue flowed out.
It took 68 years to break this cycle. In 2013, Aliko Dangote unveiled plans for a refinery in Lekki, initially budgeted at $9 billion. By the time the facility opened in 2024, costs had swollen to $19 billion, and construction spanned over a decade. Within three years, however, the refinery reduced seaborne fuel imports by two-thirds, from nearly 400,000 barrels per day to under 130,000. By mid-2026, Nigeria began exporting refined fuel to Europe. The lesson was undeniable: profit lies not in crude extraction but in refining.
Today, Nigeria is replicating this error—this time with data.
Data as the new crude
Every voice recording in Yoruba, every financial transaction in Onitsha, every insurance claim, every Nollywood script, every hospital record, and every bank transfer in Nigeria represents raw data. With 200 million people, the country generates one of the world’s richest data pools. Yet, like crude oil in 1958, this data leaves unprocessed.
It is sent abroad for refinement, used to train foreign models, and resold as intelligence, through subscriptions, cloud fees, and per-token charges, all denominated in dollars. Nigerian businesses already spend roughly $850 million annually on foreign cloud services, according to estimates from the National Sovereign Cloud Initiative. Government data hosting alone costs near $1 billion yearly, with projections approaching $4 billion if trends persist.
On the surface, Nigeria’s digital economy appears robust. GDP growth reached 4.43% year-over-year in the second quarter of 2026, the fastest pace in five years. Services now contribute 56.62% of real economic output, with information and communication identified as a key growth driver. Inflation eased to 15.39% in August. A government official could cite these figures to argue the digital sector is thriving.
Only 10.1% of working-age Nigerians used generative AI tools in the first quarter of 2026, compared to 17.8% globally and 27.5% in the Global North, per Microsoft data. Nigeria ranks 110th out of 147 economies in AI adoption. The nation produces vast amounts of raw data but captures minimal value from AI refinement.
The issue extends beyond financial losses. It is a matter of influence. Every Nigerian interaction, queries, document uploads, corrections, improves foreign AI models, particularly for Nigerian-specific applications. This creates a self-sustaining loop where local models struggle to compete. Unlike crude oil, which does not learn from its sellers, data does. This makes the second Refinery Paradox more difficult to overcome than the first.
The missing dashboard
Nigeria meticulously tracks fuel imports. The Central Bank can report, down to the barrel, how much petroleum the country imported last month. No such transparency exists for data-related expenditures.
There is no dedicated entry in national accounts for dollars spent on AI intelligence. No balance-of-payments report includes this outflow. No ministerial brief acknowledges it. Economists refer to it as the token import bill, the fastest-growing invisible drain on Nigeria’s economy. Each AI subscription paid in dollars functions economically like a small fuel import. The difference is stark: petrol burns and vanishes, while intelligence learns from Nigerian inputs.
Every interaction with a foreign AI model enhances that system’s understanding of Nigeria, making it harder for local alternatives to catch up. Unlike the oil economy, where crude improved the refinery that purchased it, data refines itself against its sellers. This creates a trap with no easy exit.
Governance failures exacerbate the problem. Nigeria released a National AI Strategy. The country has established a National AI Trust, a National Centre for AI and Robotics, and pending legislation for a National AI Commission in the National Assembly. Despite these efforts, Oxford Insights ranks Nigeria 72nd of 195 governments in AI readiness and 35th in policy capacity.
The evidence lies in the Dangote refinery’s IPO. On October 13, 2026, subscriptions closed for a N2.15 trillion offering, 4.1 billion shares at N525 each, valuing the refinery at approximately $47–49 billion. Investors purchased the refining margin, not the crude itself.
