By Matthew Turner, Head of Innovation at SolarAfrica
It’s 2022, and generative Artificial Intelligence (AI) has officially arrived in the mainstream zeitgeist. While it looks very different to its early blockbuster portrayals (remember the malevolent supercomputer HAL of 2001: A Space Odyssey or the tech-paranoia of Black Mirror?), it is surprisingly far more useful than the bots we’ve become accustomed to. But…it talks a lot but doesn’t always offer much of substance. It makes mistakes. And while we may dabble here and there out of curiosity, it’s more a novelty than an everyday thing.
Fast forward to 2026, and in under four years AI has gotten smarter and found its way into just about every profession, across just about every sector. What many of us don’t see, however, is the vast physical infrastructure behind the scenes that powers every prompt and eats terawatts like toast. From the user perspective, AI feels seamless and instantaneous – when, in fact, every interaction relies on large data centres working around the clock to keep pace with demand, devouring huge amounts of energy. According to the African Energy Chamber, data centres are expected to consume more than 16 TWh of electricity annually by 2030 in South Africa alone – equivalent to around 6.5% of the country’s total electricity demand.
This growing demand presents several challenges. As AI evolves from an on-demand tool into an increasingly autonomous, always-on capability, both electricity consumption and cooling requirements are expected to rise substantially. Meeting this demand is becoming more expensive, with rising energy costs and tariff uncertainty complicating long-term planning. At the same time, electricity networks require significant investment to accommodate new load growth, while data centre operators face increasing scrutiny around sustainability and environmental impact.
Yet these challenges also present a significant opportunity for Africa. Unlike many developed economies that are attempting to decarbonise established energy systems, much of Africa’s future infrastructure is still being built. Combined with some of the world’s best solar and wind resources, this creates a unique opportunity not only to support the growth of AI, but to shape a more sustainable model for powering it.
The real question to ask: what should that energy solution actually look like?
The road to 24/7 renewable power
For many organisations, the starting point is still the grid. While this perhaps seems like the simplest option operationally, it also leaves businesses fully exposed to rising electricity costs, tariff volatility and an increasingly constrained network. And as demand grows, so too does the risk associated with relying on a single source of supply. Another option is combining on-site solar PV with grid electricity. However, while generating renewable energy where it’s consumed can reduce electricity costs by around 24%, it’s very much limited by physical footprint: large data centres have continuous, high energy requirements, meaning the amount of land required to generate a significant volume of solar power often far exceeds what is realistically available on-site.
Adding battery energy storage (BESS) to an on-site solar installation allows excess solar energy generated during the day to be stored and discharged during peak tariff periods, improving renewable penetration and increasing potential savings to almost 30%. Even so, the solution remains constrained by the physical footprint required for large-scale solar generation, while the batteries themselves are primarily delivering cost optimisation through time-of-use arbitrage (rather than long-duration energy security).
This is where wheeling begins to change the conversation. By generating renewable electricity where the country’s strongest solar resources exist and wheeling that power through the national grid, businesses no longer need to dedicate valuable operational land to large solar installations. Pairing off-site renewable energy with on-site battery storage allows organisations to maximise renewable penetration while using batteries strategically to reduce peak demand and optimise electricity costs. Based on SolarAfrica’s modelling, this combination offers the strongest balance between cost, scalability and sustainability, reducing electricity costs by around one-third while providing most of the data centre’s consumption with renewable energy.
Perhaps the most important thing to remember is that AI itself is not standing still; it is evolving at pace, becoming more capable and increasingly embedded in the economy. Anyone who has used AI will have realised that it doesn’t replace human potential so much as augment it, enabling people and organisations to do more than was previously possible. The same principle applies to energy: as AI becomes more autonomous, the energy systems supporting it will need to become more innovative and flexible to enhance its abilities.
Find an energy partner capable of building a dynamic power solution that can evolve alongside your business. Because if AI has taught us anything, it’s that the status quo does change – and fast.
