Nigeria has a population of more than 230 million people and fewer than 700 consultant radiologists. That works out to roughly one specialist for every 300,000 people — and most of them are concentrated in the country's major cities. A patient in a rural community might manage to get an X-ray, but there is a real chance no one qualified will read it anytime soon.
That is the specific problem South African company Nexus AI has set out to solve. The company has built an AI-powered medical imaging tool that analyses chest X-rays and flags possible abnormalities in approximately 45 seconds. Crucially, it is engineered to function in settings where reliable internet and electricity — the infrastructure most AI tools take for granted — simply do not exist.
A Gap Too Wide to Ignore
International guidelines recommend between 100 and 120 radiologists per million people. Countries like Germany and the United States broadly meet that benchmark. Across Sub-Saharan Africa, the ratio collapses to fewer than one radiologist per 300,000 to 500,000 people. Diagnostic imaging equipment is slowly expanding across the continent, but the specialist workforce to interpret results has not kept pace.
The consequences are predictable: delayed diagnoses, overworked specialists, and patients who fall through the cracks of overburdened systems — particularly in lung health, where early detection of conditions like tuberculosis is critical.
From Mobile Clinics in Mozambique to AI
Nexus AI did not emerge from a Silicon Valley lab. Its roots go back to 2013, when co-founders Gerhard Ferreira, a medical doctor who previously owned a large hospital group in South Africa, and Andries Vorster, an electronic engineer specialising in clinical and medical engineering, established LTE Medical Solutions. The company delivered primary healthcare services across more than 10 African countries.
Ferreira had sold his hospital group in 2008. Years later, he and Vorster returned to healthcare — Ferreira driven by what he describes as a deep passion for primary healthcare at the community level. Their fieldwork brought them into contact with organisations including the Global Fund, collaborating on mobile healthcare initiatives.
One early project placed mobile X-ray clinics in Mozambique. Getting equipment into underserved communities was a milestone, but it exposed a deeper operational mess.
"We found that there were a lot of errors, there was duplication, and it was very difficult to monitor the patients," — Gerhard Ferreira, co-founder, Nexus AI, describing the spreadsheet-based systems some projects were using at the time.
To fix that, the founders built Healthcare Framework, a clinical management software platform designed to digitise those fragmented processes. Ferreira says the platform is now active in approximately 20 African countries and has processed more than five million consultations.
The Bottleneck That Sparked Nexus AI
Healthcare Framework improved operations, but a stubborn problem remained. Large-scale screening programmes could identify people with potential abnormalities, yet patients were not reliably connected to the follow-up care they needed. The screening-to-treatment pipeline kept breaking down.
Nexus AI was built as a subsidiary specifically to address that bottleneck — using AI to accelerate the interpretation step that radiologist shortages had made so slow and inconsistent. The platform's ability to work offline and with intermittent power supply reflects the realities Ferreira and Vorster encountered firsthand during over a decade of field operations across the continent. For the kind of health-tech innovation that can genuinely shift outcomes in Africa, offline capability is not a bonus feature — it is the whole point.
The company's positioning sits at the intersection of an urgent public health need and a growing continental push to deploy AI in contexts where it can have measurable real-world impact. Whether Nexus AI can scale its tool widely enough to meaningfully shift diagnostic outcomes across Sub-Saharan Africa remains to be seen — but the infrastructure thinking behind it is already different from most tools in this space.