Utilities are becoming the second market of artificial intelligence

This article is part of In conversation about sustainable finance & emission reduction systems, a new series by Diego Balverde. You’re reading volume 26 of the Breaking news series. Here is volume 25

The largest financial opportunity created by artificial intelligence is beginning to move beyond chip manufacturers and major technology companies toward a less spectacular but potentially more durable layer of the global economy: utilities, electricity networks, power producers, storage developers, transformer manufacturers, data-centre operators and the entire infrastructure required to supply firm electricity to a digital expansion growing much faster than the physical capacity of the energy system, which is why markets are beginning to understand that the next major financing cycle of artificial intelligence will not be decided only in technology equities but also in bond markets, project finance, private credit, infrastructure funds and long-duration corporate debt.

A data centre is not simply a building filled with processors. It is a twenty-year power contract, a new substation, transmission infrastructure, generation capacity, transformers, cooling systems, BESS, backup power, land, water, fibre, permits and millions or billions of dollars of capital that must be deployed before the first server generates revenue.

This reality is changing the perception of utilities because for years they were treated as defensive, regulated and relatively slow businesses, while some are now positioned in front of demand growth combining artificial intelligence, industrial electrification, electric vehicles, logistics centres, cooling, advanced manufacturing and grid expansion. The strategic asset is no longer only the company developing the model but also the company capable of supplying five hundred, one thousand or several thousand megawatts of firm electricity for decades without destabilising the territory in which that facility is located.

This turns electricity into a financial platform because every new major load can justify investment in generation, transmission, distribution, storage and digital control and every one of those assets possesses cash flows, contracts, guarantees, tariffs or services capable of being transformed into financeable instruments. The next wave of bonds indirectly linked to artificial intelligence may therefore never carry the words AI in their prospectus. They may appear as utility debt, hybrid bonds, grid infrastructure, project bonds, storage finance, data-centre debt, digital-infrastructure securitisation or credit for new energy installations. The investor looking only at technology companies risks seeing the surface while patient capital can capture the system allowing those companies to operate.

Europe needs to mobilise enormous volumes of capital to modernise grids, connect new generation, reinforce interconnections and finance operators whose borrowing capacity does not always expand at the same speed as investment requirements, forcing innovation through hybrid bonds, securitisation of regulated revenues, co-investment platforms, institutional debt and structures capable of attracting private capital without excessively damaging utility balance sheets.

The United States faces a similar tension because large-load growth is arriving on networks developed under assumptions of much more stable demand and every new campus requiring hundreds of megawatts can force utilities to reconsider transmission, generation capacity, tariffs, interconnection and who ultimately pays.

The financial question is decisive because not every required investment should automatically appear on household electricity bills. If a data centre creates an extraordinary infrastructure requirement, the financial contract should allocate a significant portion of that cost to the beneficiary of the capacity and prevent households and small businesses from indiscriminately subsidising the growth of an exceptionally profitable industry. Otherwise artificial intelligence can become politically an electricity problem even while remaining economically an opportunity.

The correct model needs to distinguish infrastructure benefiting the entire system from infrastructure constructed specifically for major consumers, establish demand guarantees, avoid stranded assets if projects are cancelled and structure contracts long enough to support competitive financing.

At this point bonds and credit become extraordinarily important because they allow the temporal cost of assets operating for several decades to be distributed correctly. A transmission line, substation or power plant cannot be financed as though it were a digital application with a three-year cycle. These assets require patient capital, long-duration debt and predictable cash flows.

This opens space for BlackRock, CPP Investments, Goldman Sachs, Brookfield, Macquarie, Apollo, KKR, BlackstoneInfrastructure and other major asset managers, insurers, pension funds and credit platforms capable of financing infrastructure, while Galdana Ventures and other technology-oriented investors can participate in software, automation, energy optimisation, grid intelligence and technologies capable of increasing the productivity of physical capital.

BalGreen’s opportunity is not to present itself as an intermediary for global institutions without confirmed mandates but to construct assets sufficiently robust for that universe of capital to analyse. That requires moving from a commercial presentation to an investment architecture.

If BalGreen proposes a BESS, it should demonstrate the weakness being addressed, which peak is reduced, how much real arbitrage exists, which grid service is provided, expected degradation, guarantees, O&M requirements, fuel or electricity savings and the cash flow capable of supporting debt. If it proposes port modernisation, it should demonstrate how many waiting hours disappear, how much fuel is avoided, how utilisation increases, which electrical capacity is released and what new revenues appear. If it intervenes in a utility or industrial park, it should quantify network losses, interruptions, demand, congestion, backup costs and additional capacity unlocked.

DOIX must operate as an evidence layer rather than digital decoration, measuring the system before intervention and verifying it afterward so that savings and operating improvements can enter a credit structure. The equation is straightforward: DOIX identifies the weakness, BalGreen designs the operating package, implementation reduces the loss, DOIX verifies the gain and financial markets can purchase the structured yield derived from that improvement.

This logic becomes even more important when recognising that the energy infrastructure required for artificial intelligence faces several simultaneous forms of risk. Technology risk exists because improvements in chip efficiency can alter future consumption projections. Concentration risk exists because some data centres depend heavily on a small number of major customers. Construction risk exists because transformers, turbines, transmission lines and substations carry long lead times. Regulatory risk exists because governments may change interconnection conditions or require larger contributions from major users.

Energy risk exists because gas prices, hydrological conditions and renewable variability can affect costs. Financial risk exists because projects financed with excessive leverage can suffer if interest rates remain elevated. Social risk exists because communities may oppose projects consuming large volumes of power and water while generating limited visible local benefits. Systemic risk can even appear if private credit, banks, funds and technology companies finance one another through an expansion based on overly optimistic expectations for AI growth.

The answer is not to stop investment but to make financing more disciplined. Every major load should demonstrate contracted demand, every infrastructure asset should identify who pays, every utility should understand balance-sheet impacts and every lender should model scenarios in which technology growth is slower than expected.

The same discipline reveals extraordinary opportunities. A data centre accepting partial demand flexibility may require less infrastructure than a completely inflexible facility. A BESS integrated with the campus can reduce peaks and provide grid services. A microgrid can combine generation, storage and network access to reduce exposure. Long-term contracts with nuclear, hydroelectric, renewable or dispatchable generation can stabilise costs. A hybrid portfolio can be more robust than reliance on one technology.

Artificial intelligence and electricity can even begin optimising one another because selected computing processes can shift toward hours or regions where more electricity is available, turning part of digital demand into a flexible resource. This radically changes infrastructure economics because a relatively small reduction in demand during the most critical hours can avoid enormous investments designed only to cover exceptional peaks. The smartest utility will not be the one that builds indiscriminately but the one combining new infrastructure with flexibility and efficiency.

This also redefines employment. If capital moves toward networks, generation and storage, the economy needs electricians, engineers, installers, BESS technicians, substation operators, high-voltage specialists, transformer manufacturers, cooling experts, cybersecurity professionals, maintenance teams, civil construction workers, industrial programmers, energy analysts and infrastructure-finance specialists. AI automates tasks at one end of the system while creating pressure to build physical infrastructure at the other.

The real labour challenge will be the speed at which people can be trained to execute these projects. A country can possess enough capital to build a network and discover it lacks engineers. It can buy batteries and lack commissioning technicians. It can order transformers and lack local maintenance capability. It can attract data centres and lack specialists capable of operating critical systems.

Training therefore becomes infrastructure itself and BalGreen Academy can be integrated into every BalGreen project as an operating requirement, creating programmes directly linked to BESS, grids, ports, industrial energy, maintenance, EMS, SCADA, safety and MRV. The objective should not be generic courses but training the people who will operate the financed assets. This reduces execution risk, creates local employment and improves investment credibility.

Ports can also capture this new economy because they possess characteristics difficult to replicate: large industrial land areas, high-power connections, logistics access, fibre, infrastructure, industrial customers and the ability to receive oversized equipment. BalGreen Ports can develop zones where energy, storage, data centres, BESS assembly, transformer logistics and generation integrate within the same ecosystem.

A port with a robust microgrid can power shore power, cranes, cold storage, truck charging and digital activity. A BESS can reduce peaks and provide resilience. Local generation can reduce exposure. A data centre can become an anchor customer supporting grid improvements that then benefit other users. Waste heat can even be used in selected industrial applications where nearby demand exists. The port stops being merely transportation infrastructure and becomes an energy and digital platform. The financial advantage lies in diversified revenues because the same infrastructure can serve multiple customers and sectors rather than depend on a single source of return.

The same philosophy applies to industrial parks and cities. Instead of building disconnected energy projects, a platform can measure demand and combine storage, generation, charging, efficiency, digitalisation and finance. DOIX can maintain visibility over operating flows and provide continuous evidence. BalGreen can intervene on identified losses and capital can finance portfolios rather than isolated installations. Institutional investors generally find diversified portfolios of assets with verifiable outcomes more attractive than a succession of small projects without consistent data.

This creates another major opportunity: aggregation. Thousands of industrial-efficiency improvements or smaller BESS installations may individually be too small for BlackRock, CPP Investments, Brookfield, Macquarie or major credit funds, but a platform standardising, verifying and aggregating them can reach the ticket sizes institutional capital requires. DOIX can provide a common methodology, BalGreen can implement repeatable operating packages and outcomes can be structured into vehicles where investors purchase a portfolio of savings and resilience. The real financial product then becomes not the battery, panel, substation or software separately but the aggregated yield of a more efficient system.

This concept can transform climate finance because it moves the discussion away from which technology appears greener toward how much measurable value each intervention generates. A global investor does not need to fall in love with a battery chemistry. It needs an asset whose cash flow it can understand. It does not need to believe in a savings promise. It needs data proving the savings. It does not need an environmental campaign. It needs governance, contracts, safety, return and an exit path.

The most sophisticated global funds are searching precisely for measurable, verifiable and credible assets capable of surviving political and technology cycles. If BalGreen can build that layer between operations and capital, its business stops depending on selling projects one by one and can evolve toward creating replicable financial platforms. That is real scale.

Artificial-intelligence expansion accelerates this transformation because it dramatically increases demand for the infrastructure BalGreen aims to optimise: electricity, storage, ports, grids, data centres, industrial energy and digital systems. The risk is that investment arrives before efficiency and governments respond by building capacity to cover losses that could have been reduced first.

Before installing a new power plant, the system should ask how much demand can become flexible. Before expanding a substation, it should determine how much peak load BESS can remove. Before building another terminal, it should assess how much additional throughput efficiency can release. Before financing new generation, it should measure how much power is currently being lost. This sequence can save billions and is precisely where a BalGreen-DOIX architecture can differentiate itself. Measure first. Design second. Execute third. Verify and finance the result. The coming market will be less tolerant of narratives without evidence because the amount of required capital is too large and the risks too important.

Utilities are becoming the second market of artificial intelligence because the next phase of the digital boom requires far more electricity, grid infrastructure, storage and finance than the initial phase and those physical assets can generate cash flows for far longer than a particular generation of chips. Technology changes in years while a transmission line may operate for half a century. A server can become obsolete quickly while a substation continues serving multiple technology cycles. That difference explains why patient capital can look beyond equity-market excitement and finance the physical arteries of the digital economy.

The next opportunity is not simply to build more data centres but to build the ecosystem that allows them to operate without destroying margins, grids or social legitimacy. Utilities capable of selling firm capacity, flexibility, storage and data will stop being merely sellers of kilowatt hours and begin becoming platforms for digital infrastructure. Grid operators capable of anticipating where demand will appear can capture value before congestion emerges. Transformer and power-system manufacturers will gain strategic importance because the digital economy cannot grow without them.

Bond markets will play a central role because the long useful lives of electrical infrastructure fit naturally with long-term debt and because a significant share of expansion can be structured through instruments backed by stable contractual cash flows. Private credit can fill gaps where traditional banks do not want to carry construction risk, but it will need to avoid financing projects without sufficient contracted demand. Insurers can become major sources of capital because of the duration of their liabilities. Pension funds can find in grids and energy assets a combination of stability and inflation protection. Infrastructure managers can build portfolios combining data centres, energy, storage and transmission.

Goldman Sachs can play an important role as a reference point for investment banking, capital markets and structuring around a transformation connecting technology, utilities and finance. BlackRock can approach the opportunity through infrastructure and private capital. CPP Investments can do so through a long-duration institutional perspective. Brookfield, Macquarie, KKR, Apollo and other large investors can capture different risk layers. Galdana Ventures and technology-focused firms can seek the software and automation layers increasing the productivity of those physical assets. Value will not be concentrated in one point of the system. It will be distributed from the electron to the algorithm.

BalGreen’s challenge is to build a position precisely in the middle of that chain, where physical efficiency, data and capital become one proposition. It should not limit itself to selling energy, software or consulting services separately. It should identify operating losses, design solutions, integrate equipment, automate measurement, demonstrate improvement and financially structure the result.

That architecture can be applied to a utility, a port, a refinery, an industrial park, a grid, a data centre, a city or a transport fleet. The specific technology changes, but the logic remains the same. Measure the weakness. Design the package. Execute. Verify. Finance the yield. That sequence allows efficiency to become a financial asset and, if repeated at scale, can evolve into a platform large enough to attract global institutional capital.

Utilities are becoming the second market of artificial intelligence because electrical infrastructure has moved from being the invisible background of the digital economy to becoming one of its main bottlenecks and one of its largest opportunities. Whoever controls firm megawatts, interconnection, storage, flexibility, data and finance will occupy a strategic position comparable to whoever controls chips or models. The coming decade will not reward only those developing the best artificial intelligence but those capable of ensuring that intelligence has electricity, grid access, cooling, capital and continuity.

Which utilities will know how to convert demand growth into profitable investment without transferring unfair costs to consumers, which governments will establish rules requiring major data centres to finance the infrastructure they need, which banks will distinguish projects backed by genuine demand from speculation, which global funds will prefer bonds, grids, storage and energy assets rather than chasing technology valuations alone, which regions will train the workforce required to construct the new industrial electricity system, which ports will become energy and computing hubs and how much value can BalGreen capture if DOIX measures the gaps, BalGreen converts those gaps into operating packages, execution reduces losses, DOIX verifies the results and institutional capital transforms that proven improvement into a new generation of bonds, credit and financeable infrastructure?

illuminem Voices is a democratic space presenting the opinions of leading Sustainability Thought Leaders, their views do not necessarily represent those of illuminem.

The world needs sustainability knowledge. At illuminem, no interest group or shareholder can influence our work. Thank you for supporting our mission to make high-quality and independent sustainability information free for all. Every contribution helps. Thank you for donating today.