Ten days ago, alongside Deon van Heerden, our CEO for South Africa, and Fabrice Zapfack, our CEO for Côte d’Ivoire, I spent time in Nairobi at Africa Forward and in Kigali at the Africa CEO Forum. We met dozens of clients, prospects, partners and policymakers: ministers, regulators, public sector leaders.

One conviction came back, sharper with every conversation: Africa is not behind on AI. Africa is on the cusp of a leapfrog.

We’ve seen this story before. The continent didn’t replicate the West’s path on payments — it skipped the bank-branch era and went straight to mobile money. M-Pesa wasn’t an incremental optimization of legacy banking; it was a different architecture for a different reality.

The same opportunity is now in front of Africa with AI.

The opportunity: agentic process reshape

The conversations I had on that trip were of a different nature. CEOs were not asking, “how do I sprinkle AI on my current workflow?” They were asking:

  • How do I rethink an entire industrial process to cut time-to-market in half?
  • How do I reduce maintenance costs by 20% across my asset base?
  • How do I redesign customer journeys so that agentic systems handle full end-to-end outcomes, not just isolated steps?

These are not optimization questions. They are reshape questions. And they are only answerable through rupture — by redesigning processes around what AI agents can now do, rather than retrofitting agents into processes designed for humans alone.

This is Africa’s leapfrog moment: less legacy to defend, fewer sunk costs to protect, more freedom to design native.

What it will take: the full stack, not just the model

Today’s debate often narrows to one question — local-language LLMs. They matter. But they are a small slice of what’s actually needed.

Look at how China has built its AI strength: not by training models alone, but by treating compute, sovereign data, agentic platforms and models as one integrated bet. Africa needs the same systemic posture — and what struck me on the ground is how quickly that dynamic is shifting from theory to balance sheet.

The compute layer is moving. In Morocco, a Nexus Core Systems consortium is building a 500 MW renewables-powered AI infrastructure project on the Atlantic coast, with its first 40 MW of NVIDIA Blackwell chips coming online — explicitly under Moroccan jurisdiction, not foreign law. In Kenya, Servernah Cloud has just launched as the country’s first sovereign AI cloud platform, hosted in Nairobi. Cassava and NVIDIA are rolling out AI Factory capacity across South Africa, Kenya, Nigeria, Egypt and Morocco. And the hyperscalers are scaling alongside, not against, this build-out: Google’s Johannesburg cloud region went live after a ~$148 million investment, and Microsoft has committed roughly $300 million more to South African cloud and AI infrastructure by 2027.

The sovereign-data layer is moving too — including in places most observers don’t look. The African Space Agency was inaugurated in Cairo in April 2025, coordinating 19 African countries that now operate 68 satellites, with over 120 more in development by 2030 and a €100 million Africa-EU Space Partnership behind it. This isn’t an afterthought; it’s the African data infrastructure being built in parallel with the compute one — earth observation for agriculture, climate, logistics, security. Data that the continent generates, owns and governs.

Where the gap remains is the agentic layer — the platforms where workflows are actually reimagined, not just augmented. That layer is still nascent, and it is precisely where the next wave of African ambition has to land.

And this stack won’t finish building itself. It requires international development finance institutions to evolve their instruments — financing infrastructure and platform layers, not just pilots and capacity building. It requires regulators to design for agentic AI from the start, rather than retrofit yesterday’s frameworks onto tomorrow’s systems.

The human layer is the real transformation

Let’s not romanticize the technology. The hardest part of agentic transformation is not the model — it’s the organization.

What does it mean to have AI agents working alongside human teams? How do you manage a hybrid workforce? How do you redesign roles, accountability, performance and trust when a meaningful share of your operations runs through autonomous systems? How do you upskill leaders to orchestrate this?

These are paradigm shifts, not project plans. And they are the part most likely to determine who wins.

The trap to avoid: importing the “optimization” playbook

There is one trap African leaders must resist. For the past five years, in more mature markets, AI has largely been deployed to optimize tasks and workflows: a smarter chatbot here, a better forecast there, a copilot inside an existing tool. Useful, but incremental. The org chart doesn’t change. The process doesn’t change. The KPIs barely move.

If African enterprises adopt that same logic, they will inherit its ceiling. The leapfrog only happens by skipping that stage, not replaying it.

Where Artefact stands

That agentic gap is precisely where we want to be useful. We work hands-on with CEOs who want to design the next operating model rather than patch the old one — reshaping industrial processes, customer journeys and entire P&Ls around what agentic systems can now deliver. That is the business we are in: not augmenting yesterday’s workflows, but architecting tomorrow’s.

And because none of that runs without the right foundations, we also build the environments that make it possible — the data, the platforms, the governance that turn ambition into operating reality.

Africa already proved, with mobile money, that leapfrogs happen when ambition meets architecture. The same window is open now, and it won’t stay open forever.

To the African CEOs, ministers, partners and friends we met in Nairobi and Kigali — thank you for the energy of these conversations. The continent isn’t catching up. It’s choosing a different starting line.