“It’s for the person in your village”: Oji Udezue’s $10mn Enzi Prize to crack AI’s cost wall for Africa
Ten dollars per million tokens sounds like a technical detail until Oji Udezue explains who gets left out when the price stays high. The former chief product officer at Typeform and Calendly, and product lead at Twitter, Udezue took the stage at Moonshot's Emerging Tech event to talk about a question he says keeps pulling him back to the continent: who is a technology actually for? His answer shapes everything about the Enzi Prize.
The prize is a $10 million competition, part of a roughly $14 million initiative, built around a single target. Running a frontier "Fable-class" AI model currently costs around $50 per million tokens by Udezue's benchmark. The Enzi Prize wants to bring that down to about $5, and ideally much lower. "We say 100x because maybe we're hedging our bets," he told the audience. The prize is backed by the Africa Deep Tech Foundation, TechCabal, Future Africa, and Talent Nation through the Enzi Prize Foundation, a Delaware non-profit.
Cost is only half the challenge. Udezue is clear that a cheap model scoring well on standard benchmarks means nothing if it has no knowledge of tropical diseases or the things, as he put it, "people in my village know that are not anywhere on the internet." Before any submission competes on cost, it must first clear a baseline he described as "a master's in intelligence and a PhD in African context and African languages." Entries, whether cloud-based or designed to run on a phone, are ranked on a public, open-source leaderboard with rotating sealed questions to stop teams from gaming it. Winners keep their intellectual property but must release an open, royalty-free version for use across Africa and the Global South.
The model Udezue is borrowing from is the Ansari X-Prize: put up a reward, let competition pull in far more investment than the prize itself is worth. He pointed out that a $20,000 pilot had already drawn serious talent and produced working demos. "What would $10 million do?" The purse is not yet fully funded, and Udezue mentioned he had spoken with OpenAI and Google the same day. He estimates a winner could emerge in three to six years.
His closing appeal was not really about the money. "It's for the person in your village who's not educated," he said. "Otherwise, we would be the most selfish generation that ever existed." Whether the funding follows the ambition is still an open question.
Originally published by TechCabal.