For more than a week, a machine-learning model sat running on a researcher’s laptop in Nairobi.

The task was ambitious. Analyse health data that could help scientists better understand why people die and how those deaths might be prevented. But despite the promise of the research, progress was painfully slow. The computer simply lacked the power needed to process the enormous volumes of data.

Across many African universities, this scene is familiar.

Researchers with world-class ideas often find themselves limited not by talent or ambition, but by access to high-performance computing infrastructure.

Now, a partnership between Strathmore Institute of Mathematical Sciences, under the leadership of the Dean, Dr. Godfrey Madigu, and technology company Udu Technologies is helping change that narrative.

Through the Africa GPU Hub (AGH), researchers and students at the university, including Cosmas Ochieng Kungu, Kennedy Senagi and Dr. Evans Omondi, have gained access to advanced Graphics Processing Units (GPUs), specialised computing hardware that powers modern artificial intelligence systems. The tasks that once took days or weeks can now be completed in hours.

The shift is enabling researchers to tackle some of society’s most pressing challenges, from public health and urban mobility to financial crime.

One of the projects currently under development is a traffic disruption prediction system. The system processes more than 1.5 gigabytes of transportation data. The goal is to identify patterns that signal potential disruptions before they occur, helping public transport operators respond more effectively and reduce commuter delays.

In another corner of the university, researchers are applying artificial intelligence to combat money laundering.

The project draws on research recently published by Strathmore University scholars in the international journal Machine Learning with Applications. The study developed an advanced anti-money laundering framework capable of analysing more than 54,000 real-world cross-border financial transactions from an East African commercial bank. The system achieved a precision rate of 99.63 per cent among high-priority alerts, significantly reducing false alarms while identifying suspicious activity.

The researchers also demonstrated that the system could process approximately 300,000 transactions in under five minutes. This highlights its potential for deployment in resource-constrained financial institutions.

For African economies increasingly integrated into global financial systems, such innovations could prove critical in strengthening compliance, improving transparency and detecting emerging financial crime schemes.

Perhaps the most socially significant application, however, lies in healthcare.

A Strathmore researcher is developing a hierarchical Bayesian model that uses data from a Kenyan Health and Demographic Surveillance System to predict causes of death. The objective is to generate insights that can support better public health interventions and help policymakers allocate resources more effectively.

Before access to GPU infrastructure, running the model on a personal laptop often took more than a week and sometimes failed to complete altogether. With AGH resources, the same workload can now be processed in a matter of hours.

For researchers, that difference is transformative. Faster processing means more experiments can be conducted, more hypotheses tested and more accurate models developed. It also means promising ideas can move from the laboratory to practical application far more quickly.

The developments at Strathmore reflect a broader conversation taking place across Africa’s higher education sector.

As artificial intelligence increasingly shapes industries, governments and economies, universities are under pressure to equip students and researchers with tools that match global standards. Yet many institutions continue to face significant infrastructure challenges.

Experts argue that access to computing power has become as important to modern research as access to laboratories.

The projects emerging from Strathmore University point to a challenge familiar across many African institutions. The continent is rich in talent and ideas, but researchers often lack access to the high-performance computing infrastructure needed to transform those ideas into scalable solutions.

That view is increasingly shared across the continent. Governments, universities and private-sector partners are exploring ways to strengthen Africa’s digital research capacity.

For Udu Technologies, the Africa GPU Hub represents one response to that challenge. By providing universities with access to advanced computing resources, the initiative aims to democratise AI research and enable institutions to compete on a global stage.

From predicting traffic disruptions and strengthening financial security to improving public health outcomes, the work underway demonstrates how artificial intelligence can be applied to solve practical African problems.

It is also a reminder that the future of AI will not be determined solely by algorithms. It will be shaped by whether researchers have the tools they need to build them.

Article wriiten by Stephen Wakhu.