The nascent AI revolution is not just driving electricity consumption and massive demand for additional capacity—it is reshaping how power is built, maintained, and delivered. For Africa, the real risk is no longer just insufficient capacity—it is also losing control and ability to manage the capacity it already has.
By Lazarus Angbazo
In a recent article, I argued that geopolitical fragmentation may not necessarily push investment away from Africa’s infrastructure markets. In many cases, it could even create more room for Africa to engage different partners and attract new sources of capital.
But another kind of fragmentation is also taking shape—quieter, less visible, and potentially more serious. It is not political. It is technical. And it is being driven, almost entirely by the rise of artificial intelligence (AI). Unlike geopolitics, which shifts where attention goes, this new fragmentation is beginning to determine something more fundamental—who gets access to power infrastructure, and who does not.
Something fundamental and transformational is happening in the global power industry and the rules are changing. For years, the model was predictable. You built power plants, maintained them, and relied on a network of OEMs and independent service providers (ISPs) to keep things running. Even in difficult environments, there was a basic assumption: if you operated a turbine, the supply-chain system around it would continue to support it.
That assumption is breaking down. The surge in AI-driven infrastructure—especially data centers—has created a new kind of demand. It is large, constant, and extremely sensitive to downtime. And it needs power that can match that reliability. To meet that demand quickly, the world has turned again to industrial gas turbines. The impact has been immediate. OEMs are now fully booked, with backlogs stretching from five to seven years. Prices have risen sharply, in some cases nearly tripling since 2019.
But the real shift is not just cost or timing. It is focus and prioritization. OEMs are concentrating on new turbines and long-term service contracts where the revenue is predictable and sustained over decades. Supporting older equipment across many markets is no longer their priority.
What is driving this change can now be seen more clearly in the numbers that provide an unmistakable sense of the scale of the shift. Over the past three years alone—from 2022 to 2025—power demand linked to AI infrastructure has accelerated sharply. The increase is not gradual; it is step-change. AI queries can consume up to ten times more energy than a traditional internet search, and as usage scales, so does the underlying electricity requirement. More importantly, this demand is no longer driven mainly by occasional model training. The day-to-day running of AI systems—what is known as inference—now accounts for most of the energy consumption. In other words, the load on power systems is continuous, not intermittent.
This is why data centers have moved from being just another customer segment to becoming priority customers—driving urgent, large-scale procurement of reliable generation capacity across the world. The implications extend well beyond generation. Transmission and distribution systems—already under strain in many African markets—are not designed for this level of concentrated, high-density demand. In Nigeria, for example, peak transmission has only recently reached about 5.8 gigawatts, with average delivery closer to 4.5 gigawatts. That entire system would struggle to support even a modest cluster of hyperscale data centers without significant upgrades.
At the same time, the supply chain that supports generation, transmission, and distribution equipment is becoming less accessible. OEMs across the value chain are increasingly scaling down their footprint of service centers and personnel in Africa, focusing instead on markets where demand is larger, faster-moving, and backed by stronger purchasing power. The implication is clear: Africa cannot rely indefinitely on external support systems that are no longer prioritizing it. Localization—of supply chains, technical capability, and maintenance ecosystems—is no longer a long-term ambition. It is now an immediate requirement.
The industry is no longer operating as one unified system. The market is quietly splitting in two. On one side are new assets—modern turbines, long-term service agreements, and customers tied to the fast-growing AI economy. These customers get priority, attention, and guaranteed support.
On the other side are older fleets—machines still running, still critical, but increasingly difficult to maintain because OEM support is no longer forthcoming affordably or on time. Support is inconsistent. Spare parts are harder to find. Solutions are pieced together rather than planned.
This is not a temporary distortion or short-term imbalance. It is now more likely to be a permanent structural shift and reordering of how the industry works. Therefore, the question Africa must urgently address is what this means for Africa’s Power sector.
Africa—and Nigeria in particular—sits largely on the wrong side of this divide. Much of the continent’s power system depends on older gas turbine technologies—Frame 5, 6, 7, 9, and aeroderivative machines. These machines were built to last. But they were not designed for a world where the manufacturers would gradually step back from supporting them. From a distance, the system still appears intact. Plants are running. Megawatts are being generated. But up close, the picture is different.
What does this look like from the ground? From direct experience working across multiple customers’ operating sites—gas processing facilities, refineries, utility power plants, and large industrial operations—the pattern is becoming clear. What used to be routine is no longer routine. Maintenance plans are increasingly adjusted based on what can be found, not what should be done.
In one case, a straightforward intervention stretched into months simply because one component could not be sourced in time. An overhaul that should take weeks can now take months—or may not close at all. A planned major overhaul of a GE MS 5001 turbine at Gas Processing Plant began in December 2025. The scope was defined. The team was ready. What followed was a lesson in what happens when a capable operator meets a supply chain that has moved on. Broken bolts, structural repairs, component modifications — each challenge was solved, but each took longer than it should have. By March 2026, when the unit was finally ready to be returned to service, the auxiliary gearbox procurement was still open. Not because the team lacked competence. But because the part existed somewhere in a global supply chain now principally focused elsewhere. This is not an isolated failure. It is a structural condition playing out across ageing fleets across the continent.
In another situation, a Rotor needed an answer, but nobody could prioritize it. At the same operating cluster, a routine rotor inspection surfaced rust findings. The team identified them correctly, documented them thoroughly, and referred the matter to the OEM for formal technical disposition — exactly what responsible operators should do. The response took far longer than it should have, and the schedule paid for it. The machine was not broken. But it could not be cleared for operation without an OEM opinion that sits behind newer turbines and newer customers representing the long-cycle revenue OEMs are now structured around. A well-maintained asset, held in place by a prioritization decision made in a boardroom on another continent. This is what loss of control looks like in practice — quiet, invisible in headline statistics, but accumulating steadily in deferred generation, and decisions made by constraint rather than engineering judgment.
In yet another situation, a serious conversation was had about acquiring a used turbine—not to run it, but to dismantle it for parts. These are not isolated stories. They are becoming normal. And when this becomes normal, systems begin to change. Maintenance becomes reactive. Decisions become constrained. Performance gradually declines.
From the above examples, the real risk for Africa is arguably losing grip of what we already have. For a long time, the conversation around Africa’s power sector has focused on building more capacity to plug the huge electrification deficit. That remains important. But a more immediate risk is now emerging. It is not just about how much power we can build. It is about whether we can keep what we already have running—reliably and on our own terms.
Because when the ability to repair, maintain, and optimize your own infrastructure depends on supply chains you do not control, you are not fully in control of your power system. That is the shift that is now underway. In other words, The deeper risk and most immediate constraint is that even where capacity is installed, the ability to sustain it—reliably, predictably, and at scale—is increasingly tied to supply chains that are being reprioritized elsewhere.
There is, however, an important counterpoint which paradoxically is that the same problem creates the solution for Africa. The same force driving this shift—the AI revolution—also offers a way to respond to it.
AI is beginning to change how engineering work is done. Tasks that used to take days—analyzing faults, designing parts, planning repairs—can now be done in hours, sometimes minutes, using historical data and intelligent systems. Parts that are no longer available from OEMs can be redesigned and manufactured locally. Maintenance can be predicted instead of guessed. Repairs can be optimized instead of improvised.
The response does not have to wait for a continental policy framework. At Emerald Industrial, a structured AI-in-Industry program is already applying intelligent tools to real maintenance problems. In one recent technical session, what stood out was how quickly a complex problem could be broken down. Instead of waiting weeks for back-and-forth analysis, the team had a set of clear options almost immediately. The site resident engineer still made the final call—but the path to that decision was compressed dramatically.
This is deliberate capability-building, not technology demonstration — aimed at shortening the distance between a field problem and a qualified decision. When you can analyze a fault faster, redesign a component locally, and predict a failure before it grounds an asset, you are less dependent on a support ecosystem that is increasingly looking the other way. That is the model Africa needs to scale. That changes things. Not overnight. But steadily—and in a way that begins to shift capability closer to where the problem exists.
So, what now for Africa? What is the practical path forward? If Africa is to avoid being left behind, the response cannot be abstract. It must be deliberate and practical. The starting point is clear: we must invest in the ability to sustain our own power systems. That means focusing on four things.
First, localize the supply chain. Critical spare parts should not depend entirely on overseas production lines with five-year backlogs. Africa must begin to build its own capacity to manufacture, refurbish, and stock high-demand components—starting with the most common failure points across existing fleets. Industry user-groups and associations must lead the way with thought-leadership and collaboration with governments and policy leaders on how to do this speedily and at scale.
Second, build technical talent at scale. The next generation of engineers and technicians must not only operate turbines—they must be able to diagnose, redesign, and improve them using modern tools, including AI. This is no longer optional. The National Power Training Institute of Nigeria (NAPTIN) is at the forefront of this talent development imperative. Its mandate and ambition must be expanded multiple times by collaborating with universities, polytechnics, technical institutions in the country and globally.
According to NAPTIN, Nigeria currently lacks about 50% of the skilled workforce needed to run the power sector efficiently, or about 20,000 engineers and technicians urgently needed today.
Third, develop local repair and maintenance ecosystems. Instead of fragmented, reactive interventions, there needs to be a coordinated network of workshops, specialists, and digital platforms that can respond quickly and consistently across markets. Local service providers like Emerald Industrial must actively cooperate with each other, and with foreign ISPs and repair shops on structured, ambitious, and aggressive programs for localizing capabilities.
Fourth, use AI intentionally, not passively. This is not about adopting technology for its own sake. It is about using AI to shorten repair cycles, improve decision-making, and reduce dependence on external expertise. Herein lies a fantastic opportunity for collaboration between the Ministry of Digital Economy and the Ministry of Power to unlock this latent capability to enable the leapfrog.
None of these replace the need for new capacity and the required capital funding. But these ensure that what already exists does not quietly degrade and ensure that future new investments are prioritized and optimized. The is where institutions such as The Infrastructure Corporation of Nigeria (InfraCorp) become critical. The next phase of infrastructure investment cannot focus only on building new assets. It must also support the systems that keep those assets running—local manufacturing, technical platforms for local repairs and maintenance, and capability development. In simple terms, it is not just about financing power plants. It is about financing the ability to keep them alive.
My take-away message is that we are faced with some clear choices and the possibilities for Africa are very hopeful if we act smartly, boldly, and decisively. The global power industry is moving quickly, driven by the demands of the AI economy. Africa now faces an obvious choice: continue to depend on a supply-chain and capability systems that are increasingly focused elsewhere; or begin to build the capability to stand on its own. In the AI age, power is no longer defined by how much you can generate. It is defined by how much you can maintain, control, and manage. And that is a shift Africa must respond to—now.
The utopic vision of a Nigeria with an enabled power infrastructure is no longer a pipe dream but an attainable reality – as seen by the advent of the power infrastructure development for the Data Center industries. We need not allow ourselves to be relegated to the back of the line for support to our aging power infrastructure, saddled with the cross of impractical lead times and unsustainable traditional MRO support.
There is a growing energy infrastructure enablement platform for the refurbishment, repair and ultimately rebirth of aging power equipment fleet, being powered by AI and its application in Reverse Engineering and manufacturing – that is set to wake Nigerian industries up from its long dreary nightmare of insufficient power to the hope of a future re-energized. It is time to take back our power, literally.
_*Dr. Lazarus Angbazo is a leader in Africa’s infrastructure and energy sector in his capacity as Managing Director/CEO of InfraCorp, Nigeria’s infrastructure investment platform, which is working at the intersection of capital mobilization for infrastructure and industrial capability. Through his involvement with InfraCorp, he is engaged in efforts to unlock private institutional investment into infrastructure and strengthen the long-term sustainability of critical assets. He also serves as Non-Executive Chairman of Emerald Industrial Co., where he brings practical operating insight from supporting power, Oil & Gas, and industrial assets across Nigeria. L.angbazo@InfracorpNigeria.com
Leave a comment