Udu founder outlines GPU-cluster projects with about five partners

Udu Technologies expects to work with about five partners on their first GPU computing clusters, founder Alexander Tsado told Business Insider Africa in a profile published November 8. Each project would be worth $1 million to $2 million. His pitch is to help African organizations move from trying AI tools to building with their own data.

Buying computing in smaller pieces

Udu sells access to GPU chips used for AI computing, alongside computing hardware and advisory services. Tsado describes the mechanism as combining an organization’s data and understanding of that data with GPU computing to produce models or insights. The goal is not simply to copy an existing AI application, but to build around local needs.

The company says developers can use its cloud GPUs for less than $1 an hour. It also offers computing credits starting at $50 and $100, distributed through venture capital firms and accelerators. These smaller purchases are intended to reach startups before they become established enough to qualify for large credits from global cloud providers.

Tsado compares that approach to mobile-phone recharge cards: reduce the size of the purchase so more people can start using the service. He wants Udu to shorten AI adoption from 10 years to two or three. That is an ambition for the business, not a measured result from its deployments.

Installing chips is only part of the work

Tsado’s infrastructure experience includes work at Nvidia between 2016 and 2020, helping establish some of its early cloud GPU infrastructure for Amazon Web Services. At Udu, he says, deployment has required adapting to countries with different electricity costs, infrastructure conditions and levels of technical expertise.

He says Udu helped drive a GPU deployment project across six African countries. The work included training local teams to manage the equipment and eventually expand its capacity. His account puts ongoing operation—not just the initial delivery of hardware—at the center of making computing infrastructure useful.

In Nigeria, Udu is involved in implementing the AI Scaling Hub with partners including Lagos Business School and Data Science Nigeria. Tsado says it has also built software that lets institutions manage access to computing resources rather than handle those processes manually.

Sharing that capacity across national borders is another obstacle. Tsado says getting governments to share access has proved difficult, even as he argues that individual countries need to pool resources.

We're going to need to share resources.

Alexander Tsado, founder and CEO of Udu Technologies, speaking to Business Insider Africa

The customer still needs a business case

Tsado argues that businesses should judge AI against their own operations, rather than the spending of global technology giants. He points to bank transaction processing, suggesting GPUs could cut some reconciliation jobs from hours to about 30 minutes. Those are his illustrative performance claims, not independently verified results from a named customer.

His funding argument goes beyond governments buying equipment. Based on Udu’s calculations, he believes businesses should eventually spend 5% to 10% of revenue on AI, including training, software, tools and hardware. Governments, in his view, should fund research and infrastructure rather than issue AI strategies without financial backing.

There is also an ownership argument behind the infrastructure push. Tsado worries that collecting African-language data will not, by itself, create products and technology owned by Africans. He wants organizations to process that data locally, comparing computing infrastructure to a refinery that turns a raw resource into something more valuable.

Beyond the initial partner projects, Udu is targeting deals worth more than $100 million within three to five years, Tsado says. That remains a growth target, rather than a disclosed order.