About $760 to $1,510 per person in training cost, on Applied Frontier’s benchmark. That is roughly one-eighth to one-sixteenth of the cost of a four-year computing degree.
The benchmark takes the cost of teaching one computing student for one year at a Kenyan public university, KSh 408,000 or about $3,026, and applies it to a program of three to six months, the time needed to train most of the eight AI-economy job families. Three months costs about $760 and six months about $1,510. The four-year degree costs about $12,105.
Population sets the scale. If Kenya trained one in ten of each year’s labor-market entrants, 80,000 to 100,000 people, the training bill would be about $61 million to $151 million a year. For comparison, the National AI Strategy 2025–2030 has been costed at KSh 152 billion, about $1.18 billion, over its life.
This is a benchmark, not a measured result. Training cost only, at public teaching-cost rates. It excludes devices, connectivity, placement and the cost of the employer demand that turns a trained person into a job. Targeted pilots measure the full cost per placed worker.
$760–$1,510
training cost per person, 3 to 6 months
$12,105
four-year computing degree
$61 million–151 million
a year to train one in ten labor-market entrants
How a targeted pilot measures cost per job →