By Eamonn Ryan, based on a presentation by Wojtek Piorko at the seminar AI Unlocked: What ‘AI-ready’ means for 2026 decisions
From digital convenience to critical capacity.
Wojtek Piorko, managing director for Africa at Vertiv. © RACA Journal
The conversation around artificial intelligence (AI) often focuses on algorithms, applications and the promise of automation, yet behind every AI model lies something far more tangible: physical infrastructure. Data centres, power systems, cooling technologies and connectivity networks form the backbone of the digital economy.
As AI adoption accelerates, understanding the foundations of this infrastructure – particularly in emerging markets like Africa – is becoming essential to future-ready decision-making.
Why the next wave of data centres will be faster, denser and more modular
Over the past few years, the importance of this infrastructure has moved from being a convenience to an absolute necessity. Before 2020, digital infrastructure was widely seen as an enabler of convenience. Online government services reduced paperwork, digital platforms simplified commerce, and communication tools made collaboration easier.
Then the world changed. The global disruption caused by the COVID-19 pandemic demonstrated how essential digital systems had become. Remote work, online communication and digital commerce suddenly became the primary mechanisms through which economies and societies continued to function.
Platforms such as WhatsApp, Microsoft Teams, Skype and FaceTime became the infrastructure of daily human interaction.
This shift accelerated investment in digital infrastructure worldwide, including in Africa.
Between 2021 and 2023, major international operators expanded aggressively across the continent. Global players such as Equinix, NTT, Digital Realty and Vantage Data Centers all recognised the long-term potential of Africa’s digital market.
Yet despite this investment, the continent still represents only a tiny share of global data centre capacity.
The scale gap Africa must close
Globally, active data centre capacity is estimated at around 100GW of white space (usable, IT-ready capacity inside a data centre). Africa’s share of that capacity remains extremely small.
Today, the entire continent has roughly half a gigawatt of active data centre capacity – less than many individual global cities. For perspective, Dubai alone has roughly the same amount of operational capacity as all Africa.
For a continent of more than 1.4 billion people, this disparity represents both a challenge and a major opportunity. AI is likely to accelerate the need for new infrastructure even further.
But expanding data centre capacity in Africa requires more than just investment. It also requires solving several structural challenges: reliable power supply, regulatory complexity across 54 countries, permitting processes and the development of technical skills.
Another issue is data sovereignty. Much of the data generated within Africa is currently stored and processed outside the continent, raising concerns about control, governance and economic value.
The technology shift driving AI infrastructure
The rapid growth of AI workloads is being driven by a fundamental shift in computing architecture. Traditional data centres relied primarily on central processing units (CPUs), which process tasks sequentially – handling one instruction at a time. Graphics processing units (GPUs), originally designed for rendering images, operate quite differently. They are built to process thousands of operations simultaneously.
A simple analogy illustrates the difference: imagine transporting 1 000 people to an airport – one extremely powerful vehicle might eventually complete the job, but it would take many trips. Alternatively, a fleet of smaller vehicles could move everyone at once, completing the task far more quickly. That parallel processing capability is what makes GPUs essential for AI workloads.
Importantly, CPUs are not disappearing. They remain vital for many applications and system operations. But the rapid expansion of GPU-driven computing is transforming the design requirements of modern data centres.
continue in part two…
