advertisement
Africa’s AI Talent Tug-of-War
In March 2022, Kenya’s then President Uhuru Kenyatta was shown around Microsoft’s new Africa Development Centre at Dunhill Towers in Westlands, Nairobi. The centre was the Kenyan half of a $100 million plan, announced in 2019, to employ African engineers to develop products used around the world.
By December 2022, Microsoft said it had more than 500 engineers in Nairobi. On Monday, 12 October this year, staff were told the engineering centre would close, a source at the centre told CIO Africa. WeeTracker reported that nearly 400 jobs were at risk.
Microsoft described the move as part of a wider restructuring but did not explicitly confirm or deny the closure. Its Lagos centre had already shut in 2024.
advertisement
The reported closure comes amid what is usually described as a talent shortage. Boston Consulting Group estimated in July that Africa has about 62,000 AI specialists, representing around 5 per cent of the global AI workforce.
As Africa Tech Festival continues in Cape Town this week, several groups are competing for a share of that talent pool: banks and telecommunications companies embedding AI in their operations, governments developing national AI strategies, start-ups building models and global firms recruiting African engineers, both in offices and increasingly through remote work.
For local employers, the most difficult competitor is the one that never opens an office. Foreign technology companies now pay African engineers in hard currency while they remain at home, ITWeb has reported. This means South African enterprises are competing directly with US and European employers and their stronger balance sheets.
advertisement
In markets where local currencies have weakened, the gap is wider still. A salary paid in Kenyan shillings or Nigerian naira loses value as those currencies depreciate. A remote salary paid in dollars offers a different measure of financial security.
Start-ups feel this pressure most acutely. A report published this year by GIZ’s AI Made in Africa initiative and Briter Bridges found that the main talent challenge for African AI businesses is the limited supply of qualified professionals. International technology companies can also offer significantly higher pay.
Fifty-one per cent of the companies surveyed identified AI research as the hardest role to fill, while a third cited machine learning operations and infrastructure. Only 22 per cent recruit internationally. The most common response, reported by 47 per cent, was to train existing employees.
advertisement
Nigerian health technology company MyItura hires junior developers and puts them through continuous, structured training. In Côte d’Ivoire, regulatory technology firm Danaya is developing young engineers in generative AI, computer vision and machine learning.
Larger enterprises are pursuing similar strategies at scale. Absa reported in August that more than 1,400 of its developers use AI coding tools, while more than 30,000 employees use Microsoft Copilot each month. Another 4,400 staff members are licensed to build their own automation agents.
In South Africa, AI specialists have joined the country’s ten hardest-to-fill skills, according to Xpatweb’s September survey. Governments are also working to expand the talent pipeline. Nigeria’s Three Million Technical Talent programme, for instance, is training people in digital skills at national scale.
There is evidence that African workers are moving faster than the institutions around them. In BCG’s 2025 AI at Work survey, 55 per cent of respondents in Africa said they had already upskilled in AI, the highest share of any region.
Kate Woska, Vice-President for Africa and the Middle East at talent company Andela, told ITWeb Africa that part of the answer lies in “upskilling the engineers who are already working, and doing it fast.”
The Microsoft closures add a lesson the talent debate has tended to overlook. Global companies compete for African engineers when it suits their business plans, then step back when those plans change.
The Lagos centre lasted about two years after opening. In Nairobi, the commitment ran for seven years. Microsoft remains a sponsor of African AI research events, including the Deep Learning Indaba, illustrating how corporate support for research and investment in local employment can move in different directions.
For chief information officers, this changes how the talent problem should be understood. Competition for AI engineers is real, but the supply is not fixed, and foreign firms are not the only ones determining where that talent goes.
Organisations that train their own people, measure their contributions and offer competitive compensation are better positioned to build teams that last.
Several hundred experienced engineers in Nairobi are being told this month that their jobs are going. Where they go next will depend partly on which African employers are ready to hire them.