The future of AI in South Africa must be built, not imported

South Africa has learned repeatedly that access to strategic infrastructure and control over it are not the same thing. From mining and electricity to rail and telecommunications networks, the country’s economic trajectory has often been shaped by who owns critical capabilities and who captures the value they create.

Today, a new strategic asset is emerging, not beneath the ground or across railway tracks, but across algorithms, data and computing infrastructure. As artificial intelligence becomes embedded in every sector of the economy, a key national question is taking shape: who will own, control and benefit from the intelligence systems that increasingly power economic growth?

We have faced a similar crossroads before. As cloud computing became mainstream, South Africa adopted a cloud-first approach across the public sector. This accelerated digital transformation, but it also entrenched deep dependence on a small number of global providers. By failing to build meaningful national capability alongside adoption, we limited our ability to influence pricing, terms and, ultimately, our own digital destiny.

We cannot afford to repeat that mistake with AI.

AI is not simply another technology layer or productivity tool. It is fast becoming the operational intelligence layer of organisations, governments and economies. If we consume AI that is designed, trained and governed elsewhere, we risk borrowing someone else’s intelligence while giving away our competitive advantage. At the same time, we need to understand that sovereignty cannot be added later as a compliance exercise. Once control is surrendered, it becomes difficult to reclaim. Sovereignty must be part of the design from the start.

So, what does getting this right actually look like?

Building capability at home

Sovereign AI is not just about where AI runs, but also who builds it.

That means backing the researchers, entrepreneurs and developers building AI in South Africa. It means creating pathways for people to develop AI skills at every level. And it means ensuring that the models shaping our future are trained on our data, governed by our laws and aligned to our values.

This foundation is already beginning to take shape. The University of Cape Town’s AI compute initiative is expanding access to the computing power needed for AI research. Programmes such as the AI Hub for Development’s Compute Accelerator are helping local builders access the tools, expertise and support needed to innovate. And local ventures are finding ways to run advanced AI on smaller, more efficient infrastructure.

These efforts matter because they are helping build the skills, intellectual property and innovation ecosystems that allow South Africa to create, adapt and govern AI on its own terms.

Turning potential into advantage

South Africa is not starting from scratch. The country already has many of the ingredients it needs.

We continue to attract significant digital infrastructure investment and our universities are producing world-class AI research. South Africa also has unique national data assets, including one of the world’s largest public antiretroviral treatment programmes. Combined with our growing startup ecosystem and deep technical talent, these are significant advantages.

However, potential does not mean much if we do not act on it. The countries that are moving fastest on AI development, will be the ones that attract and retain the talent, the capital, and the startups that define the next era of economic growth. We have got to move with urgency.

Building for choice, not dependency

We also need to be realistic about the scale of the global AI race. The world’s largest model developers and cloud providers are investing at levels few countries can match. Trying to outspend them is not a winning strategy.

The question therefore should not be whether South Africa can build a model that outperforms the best global alternatives. It’s whether we can build the foundations that give our economy genuine capability, control and choice.

To achieve this objective we need to start by removing the barriers slowing us down. The IMF estimates that AI could lift Sub-Saharan Africa’s economy by around 4% over the next decade if the region improves electricity supply, internet access and digital skills. We need to accelerate the infrastructure and skilling projects, permitting processes and investment decisions that enable AI development at scale. We need to treat energy availability and cost as strategic issues, because compute cost is directly linked to energy cost. And we need to give local companies the capital and regulatory clarity to build and scale here, rather than elsewhere.

At the same time, we must be deliberate about which parts of the AI stack we own, control or strongly influence. Encouragingly, South Africa has already begun to recognise this through emerging sovereign AI initiatives aimed at strengthening national capability across both the public and private sectors.

Globally, there are also already useful models to learn from. Siemens Energy, for example, chose not to rely solely on external AI services. Working with HPE, it built a sovereign, AI-enabled engineering platform that gives it control over its data, intellectual property and critical systems, while using AI and high-performance computing to accelerate innovation. The platform supports digital twins, advanced simulations and predictive maintenance. AI is not treated as a standalone tool, but as a core part of how the business designs, operates and competes.

There is an important lesson here for South Africa. Similar approaches could be applied to strategic infrastructure operators such as Eskom. The goal is not isolation from global technology ecosystems. It is ensuring that we are building enough of a foundation that we create choice, not dependency.

Sovereign AI is no longer a future consideration. The infrastructure is being built, the models are being trained, and the decisions about who controls the resulting economic value are being made today. South Africa already has many of the ingredients needed to determine our own AI destiny. If we move with urgency, we have an opportunity not only to participate in the AI economy, but to help shape it on our own terms.

//By President Ntuli, Managing Director of HPE South Africa