The Grid Is Becoming the New Foundation of Industry

Last Updated 57 mins ago by Kenya Engineer

For two centuries, industrial development has followed energy. Factories grew around coalfields, refineries around ports and pipelines, and energy-intensive industries around reliable sources of fuel. The geography of industry was therefore shaped partly by where energy could be extracted, transported and consumed economically.

That relationship is changing. The next phase of industrialisation is increasingly being shaped not simply by access to energy resources, but by access to large quantities of reliable, affordable and increasingly low-carbon electricity. Artificial intelligence and data centres have made the shift especially visible, but they are only part of a much broader transformation involving electric vehicles, industrial electrification, cooling, heat pumps, automation and increasingly electricity-dependent manufacturing.

A Reuters analysis published in September 2026 described this emerging competition as a new industrial race centred on electricity. The underlying engineering question is more fundamental: can countries build power systems quickly enough to support the industries they want to attract?

The answer will depend on considerably more than building new generating plants.

Electricity is becoming industrial infrastructure

The International Energy Agency’s Electricity 2026 outlook describes a rapidly changing electricity system. Global electricity demand is expected to grow substantially faster than overall energy demand through 2030, driven by industrial activity, electrification, cooling, electric vehicles, data centres and other new loads. The IEA says meeting this growth will require annual grid investment to increase by about 50% from today’s roughly US$400 billion by 2030.

The distinction between energy and electricity is important.

An economy can have substantial energy resources and still struggle to provide the electricity required by modern industry. Solar and wind resources may be abundant. Gas may be available. Hydropower or geothermal resources may be significant. But without transmission capacity, substations, distribution networks, transformers, storage, system controls and sufficient investment, those resources do not automatically become productive electricity.

This is becoming one of the central infrastructure questions of the decade.

The IEA estimates that more than 2,500 GW of renewable generation, storage and large electricity-load projects are currently sitting in grid connection queues around the world. At the same time, planning, permitting and constructing major grid infrastructure can take five to 15 years, compared with roughly one to five years for many renewable projects and one to three years for data centres.

The result is an increasingly familiar paradox. A country can have electricity projects ready to build, industries ready to invest and data centres ready to deploy, yet still be unable to connect them quickly enough.

The bottleneck is no longer necessarily generation. It is the network between generation and demand.

AI has made the problem impossible to ignore

Few technologies illustrate the changing relationship between industry and electricity as clearly as artificial intelligence.

AI systems require data centres filled with increasingly powerful processors. Those processors require electricity, but also cooling, power conditioning, backup systems, networking and increasingly sophisticated electrical infrastructure.

The IEA estimates that electricity consumption from data centres worldwide will roughly double from about 485 TWh in 2025 to around 950 TWh by 2030. Electricity consumption from AI-focused data centres is growing even faster.

The engineering challenge is not only the amount of electricity involved. It is the concentration and behaviour of the load.

An AI data centre can represent an enormous electrical load concentrated at a single location. AI workloads can also create rapid changes in power demand. The IEA notes that AI server power density has increased sharply and that an individual advanced server rack could by 2027 have peak power demand equivalent to around 65 households. It also expects AI workloads to increase the importance of energy storage because of their rapid power fluctuations.

This creates an unusual situation in which the digital economy is increasingly dependent on very physical engineering assets.

A new AI model may exist in software, but running it requires transformers, switchgear, substations, generators, batteries, cooling systems, cables, transmission capacity and a reliable electricity supply.

The computing revolution is therefore becoming an electrical engineering problem.

The data centre is becoming a grid problem

The conventional model of electricity planning was relatively straightforward. Utilities forecast demand, build generation and networks to meet it, and connect customers as required.

Large data centres complicate that model.

They can arrive with very large projected loads and short commercial development timelines. Grid infrastructure, meanwhile, has long planning and construction cycles.

The IEA identifies this mismatch between rapidly developing data centres and slower-moving electricity infrastructure as a potential source of system misalignment. Data centres can trigger requirements for new generation and network investment, while their actual load can develop progressively and may initially be uncertain.

This is already changing how data-centre developers think about power.

Where grid connections are delayed, developers in some markets are considering onsite generation, batteries and other forms of dedicated supply. The IEA notes that onsite gas generation is emerging as one response to grid constraints, although it also points out that this does not remove the underlying need to expand and strengthen the grid.

In other words, the data centre cannot simply be treated as another building that happens to consume electricity.

Its location, electrical load profile and connection requirements increasingly have to be considered alongside the development of generation and transmission infrastructure.

That is a major change in infrastructure planning.

The grid itself is becoming the strategic asset

The generation side of the electricity equation is attracting much of the attention. Solar and wind continue to expand rapidly, while nuclear, hydro, geothermal and gas remain important in different markets.

But electricity only becomes useful when it can be moved to where and when it is needed. This makes the grid increasingly strategic.

The IEA says grid investment has lagged behind investment in generation and that more than 2,500 GW of projects are currently waiting for grid connections. It also identifies technologies such as dynamic line rating, advanced power-flow control, reconductoring, voltage uprating, battery storage and flexible connection arrangements as ways of unlocking additional capacity from existing networks.

This is significant because building more lines is not always the only answer.

Digital monitoring can provide system operators with better visibility of actual network conditions. Dynamic line rating can allow transmission capacity to reflect real-time environmental conditions rather than relying entirely on conservative static ratings. Advanced power-flow control can improve the utilisation of existing networks. Battery storage can shift demand and supply across time.

The future grid is therefore likely to be not simply larger, but more observable, more controllable and more flexible.

That has implications all the way down to distribution networks.

Resilience is becoming as important as capacity

There is another dimension to the electricity race that is sometimes overlooked.

A power system can have enough installed generation and still be unreliable.

The IEA’s 2026 electricity outlook highlights growing risks from ageing infrastructure, extreme weather, cyber threats and other vulnerabilities. Recent large-scale outages around the world have demonstrated how quickly failures in critical electricity infrastructure can spread into wider economic and social disruption.

This changes the engineering definition of a good power system.

It is no longer sufficient to ask whether a grid can deliver enough megawatts under normal conditions. Engineers and system planners increasingly have to ask what happens when a major transmission line fails, when a transformer is unavailable, when extreme weather damages infrastructure, when a cyberattack affects control systems or when a sudden change in generation or demand destabilises the system.

Resilience therefore needs to be designed into the system rather than treated solely as a response after an outage.

The IEA’s work on energy-system resilience argues that resilience planning is most effective when incorporated during system planning, supported by risk assessment, physical protection, monitoring and the ability to restore service rapidly after major disruptions.

For increasingly digital power systems, the boundary between electrical resilience and digital resilience is also becoming less distinct.

The more sensors, communications networks, remote controls and automated systems a grid contains, the more important cybersecurity and secure system architecture become.

Kenya is already encountering the same engineering question

Kenya provides an interesting illustration because its electricity challenge is not simply about generating more power.

The country has made substantial progress in expanding electricity access and renewable generation, but demand is also increasing. Kenya’s peak demand reached 2,439.06 MW in December 2025, according to the Energy and Petroleum Regulatory Authority, an increase of about 151 MW from the previous peak.

At the same time, Kenya continues to extend and reinforce its electricity network.

In September 2026, Kenya Power announced that it was nearing completion of a KSh1.01 billion project comprising a new 66/11 kV Lodwar substation and a 90-kilometre 66 kV line between Lokichar and Lodwar. The project is intended to connect Lodwar and surrounding areas to the national grid, replacing reliance on diesel generators and creating additional capacity for future demand.

On the coast, Kenya Power is investing approximately KSh765 million in two new substations in Kwale and Kilifi to strengthen the quality and reliability of electricity supply.

These projects illustrate an important point. The electricity race does not look the same everywhere.

In one location, the engineering priority may be connecting a previously isolated town to the national grid. In another, it may be strengthening a substation to support industrial and commercial growth. Elsewhere, it may involve adding transmission capacity for new renewable generation or upgrading distribution networks to accommodate electric mobility and other emerging loads.

All of these are parts of the same electricity infrastructure challenge.

Kenya’s data-centre question brings the issue closer to home

Kenya is also beginning to confront the other side of the equation: what happens when very large new electricity loads arrive?

Kenya Engineer has been following this issue through its coverage of Africa’s data-centre expansion and the changing engineering requirements created by AI.

Our recent feature, “What AI Is Doing to Data-Centre Engineering in Africa,” examined how AI is changing data-centre design, including the increasing importance of power, cooling, fibre connectivity and physical infrastructure.

An earlier Kenya Engineer feature on Africa’s data-centre evolution similarly examined the growing relationship between AI, data centres and new energy requirements.

The issue has become particularly tangible in Kenya.

A proposed large-scale AI data-centre development in Mombasa has already raised questions about power supply, while the broader development of digital infrastructure is creating a need to consider electricity, marine infrastructure, fibre, cooling and water together rather than as separate engineering disciplines. Kenya Engineer examined that intersection in its recent feature on the proposed US$1.5 billion Mombasa AI data centre.

The lesson is not that data centres should simply be supplied at any cost.

It is that major new electricity loads need to be incorporated into infrastructure planning early enough for generation, transmission, distribution and storage to develop alongside them.

That is precisely the type of planning challenge now emerging in larger electricity markets around the world.

The answer may not be more generation alone

There is a temptation whenever electricity demand rises to frame the answer simply as building more power plants.

More generation will certainly be necessary in many markets. But the emerging electricity system is more complicated.

Solar and wind output varies with weather. Electric vehicles can create concentrated charging demand. Industrial loads can be large and inflexible. Data centres require high availability. Batteries can shift electricity through time but require appropriate charging and dispatch strategies. Transmission networks have physical limits. Distribution networks can become constrained even when national generation capacity appears adequate.

The result is a growing need for flexibility.

The IEA identifies batteries, demand-side participation, flexible connections and other technologies as important tools for managing increasingly variable supply and concentrated demand.

This is also where AI could become part of the solution to the problem it is helping create.

AI and digital technologies can be used to monitor grids, identify abnormal behaviour in transformers and other equipment, optimise network operation and improve maintenance. The IEA notes that AI-enabled monitoring and digital grid technologies could help optimise existing capacity and reduce unexpected equipment failures.

The possibility is therefore emerging of a more intelligent electricity system in which generation, storage, demand and network assets respond dynamically to conditions.

That does not eliminate the need for physical infrastructure.

It makes the physical infrastructure more valuable.

The race is ultimately about how quickly infrastructure can be built

The industrial advantage created by electricity will not come simply from having the cheapest solar panels or the largest installed generation capacity.

It will come from the ability to convert energy resources into usable electricity, and then deliver that electricity reliably to factories, transport systems, homes, data centres and other productive activities.

That requires generation, transmission, distribution, storage, transformers, switchgear, control systems, communications, skilled engineers and financing.

It also requires institutions capable of planning these systems faster than the demand for them develops.

This is where the timing problem becomes particularly important.

The IEA estimates that major grid infrastructure can take five to 15 years to plan and construct. Yet some of the industries creating the new demand — particularly data centres — can be developed much faster.

A country that waits for electricity demand to materialise before beginning the network expansion may find itself permanently chasing its own industrial ambitions.

The opposite risk also exists. Building large infrastructure without a sufficiently credible understanding of future demand can leave consumers carrying the cost of underused assets.

The engineering challenge is therefore one of coordination as much as construction.

Africa has an unusual opportunity

For Africa, the electricity race presents a particular paradox.

Many African countries remain constrained by electricity access, reliability and affordability. At the same time, the continent has substantial solar, wind, hydro and geothermal resources and, in several markets, large amounts of undeveloped renewable potential.

The opportunity is to use this resource base not simply to increase electricity access, but to support industrialisation.

That could mean renewable-powered manufacturing, data centres, electric mobility, green hydrogen, mineral processing, cold chains, irrigation, digital services and other electricity-intensive activities.

But the resource itself is not enough.

A solar resource in a remote location does not automatically power an industrial park. A wind farm without adequate transmission capacity cannot deliver its full economic value. A data centre cannot operate reliably simply because a country has sufficient annual generation. And a modern city cannot become resilient merely by adding more megawatts to the national system.

The engineering is in the connections.

A new measure of industrial competitiveness

For much of the industrial era, countries competed over access to fuel, raw materials, ports and markets.

Those factors will remain important. But electricity is becoming a more visible measure of whether an economy can support the industries now emerging.

Can a factory obtain sufficient power when it needs it?

Can a data centre connect to the grid without waiting a decade?

Can renewable energy be moved from areas of high resource availability to centres of demand?

Can batteries and flexible loads help manage peaks?

Can the grid withstand extreme weather and equipment failures?

Can utilities detect faults before they become outages?

Can transmission and distribution networks expand quickly enough to keep pace with electrification?

And perhaps most importantly, can all of this be done at a cost that allows industry to remain competitive?

These are engineering questions, but they are also economic questions.

The old industrial map was drawn around coal, oil and gas. The emerging one is increasingly being drawn around electricity.

The countries that benefit will not necessarily be those with the largest energy resources. They will be those able to turn available energy into reliable electricity, move it efficiently, manage it intelligently and build the infrastructure quickly enough to meet demand.

For engineers, that makes the coming decade particularly consequential.

The industrial race may indeed be a race to build electricity.

But the deeper race is to build the power systems capable of making electricity useful.