A Nigerian entrepreneur who co-founded two billion-dollar African startups and was a founding investor in a third is making another bold bet: that African software engineers will fill the rising global demand for AI talent.

Iyinoluwa Aboyeji has started Talent Nation to supply developers to multinationals, providing the human input that remains relevant to making usable AI products. The new venture mirrors Andela, a company he launched in 2014 and which — with the backing of investors like Mark Zuckerberg, Serena Williams, and former US Vice President Al Gore — trained a generation of software engineers in six African countries. The dizzying levels of capital being raised by AI companies in Silicon Valley, underscored by the buzz around Anthropic’s march to a potentially historic IPO, point to a market opportunity that could surpass his previous endeavors.

“We’ve got a goal of a million AI engineers over the next 10 years,” Aboyeji, who also co-founded Flutterwave and was one of the first checks into Moove, told Semafor in an interview. Most developers from his new company’s training programs will be African, though “I don’t know that we are going to constrain ourselves,” he said. Andela, which Aboyeji left in 2016, eventually moved beyond training only African software developers in 2021.

Talent Nation is taking off with seed money from African investor Enza Capital and London-headquartered fund LocalGlobe, Aboyeji said, but declined to say how much. The startup’s inaugural program, Learn2Earn, has received 63,000 applications, from which less than 3% were selected and are currently in training in multiple campuses. It does not charge fees for tuition.

“Developing talent at scale is an extraordinarily complex undertaking,” Kesiah Valkin, the company’s CEO and Talent Nation’s co-founder, said. “It is easy for us to say that Africa has enormous talent but what’s really hard is building the infrastructure to actually find that talent and develop them to globally competitive standards.”Africa’s AI scene is not recording anything close to the flood of money pouring into the frontier model companies from tech giants such as Microsoft, Google, and Nvidia.

Compared to the rush to build data centers in the US, high-level AI conversations in Africa’s leading technology ecosystems often focus on governance concerns and national policy documents, like those published by Nigeria and Kenya over the last two years.

Substantial infrastructural moves are being made, too. A company owned by Zimbabwean tech billionaire Strive Masiyiwa opened its first AI factory — composed of data centers and chips that process large language models — in Cape Town, South Africa this year and plans to build another in Johannesburg. But beyond the need for data centers and capital investment into companies building AI-enabled services, the supply of human talent is the largest global gap in AI development and one that Africa is positioned to fill, Talent Nation’s founders say. Africa has only 62,000 AI specialists, accounting for 5% of the global AI workforce, according to the Boston Consulting Group.

With the AI investment arms race has come a war for talent that involves the offer of multimillion-dollar pay packages. In the US, base salaries for AI engineers are now much higher than the mean for other jobs in the US, according to online learning platform Coursera. “When you see compensation behaving like that, it usually means supply is constrained,” Valkin said.

Some of the startup’s plans are yet to be finalized. For example, Aboyeji said he has not made a decision whether Talent Nation’s trained engineers will be the company’s full time employees or freelancers. But a key pillar of his calculation is that AI engineers can become world-class for less than $5,000 over a two-year training period, without paid instructors but based on a peer learning system.

“We don’t need $300,000 AI engineers. That’s not how this technology is going to find its way to mom and pop shops around the world,” Aboyeji said.