Global Electricity Undersupply - Executive Summary

As mentioned previously, when identifying topics for reports, we look for where inevitability will collide with complexity to generate investor misunderstanding and therefore, exploitable opportunity. We believe this pursuit is particularly valuable at this time of excessive AI enthusiasm, when seismic CAPEX expectations are baked into market assumptions despite ever-mounting evidence that AI is likely to disappoint on its path to ROI. This is why we’ve written reports this year on “Defense Innovation” and “The Medical Innovation Inflection”—two themes we believe can help hedge AI risk as stakeholders realize accelerating growth near regardless of AI CAPEX’s trajectory. And it’s why we’ve selected “Global Electricity Undersupply” for this report.

Of late, data center electricity demand has dominated the conversation about electricity supply challenges. Obviously, that’s for good reason. As we wrote in our previous report on “The AI Trade”, electricity undersupply is a key threat to the AI buildout and core to why we expect hyperscalers will fall short in building out the data center capacity required to scale their models enough to inspire adequate enterprise spending. In 2025, US electricity generated grew by 2.8%, with data centers accounting for 55% of that growth, according to the EIA. Annual growth between 2% and 3% is expected through the end of the decade, with data center electricity demand more than doubling from 31 GW in 2025 to 66 GW in 2027, according to Goldman Sachs calculations. Just since publishing our report on “The AI Trade” in September, Morgan Stanley released new, even more dire estimates about the electricity-supply shortfall to come, projecting a 30 to 40% gap between US capacity and sales projections, equal to the US being “short power by six New York Cities”.

To quote the FT: “Electricity is to AI what coal was to the Industrial Revolution…The US’s failure to expand the electricity grid risks strangling the AI boom.” Yet, the fixation on data center electricity supply neglects that data centers are only a small part of the global electricity supply challenge. Data centers currently account for only 1.5% of global electricity consumption. Even considering current AI buildout expectations, that share of consumption is projected to reach just 3 to 4% by 2030. And nearly half of data-center-driven growth (49%) is expected to come from the US, which currently accounts for roughly 40% of the world’s total data center electricity consumption, according to the IEA.

Meanwhile, electricity demand is growing faster outside of the US than it’s growing in the US. Last year saw global electricity demand increase 3%. The IEA expects demand to only accelerate over the next two years, hitting 3.6% demand growth in 2026 and 3.8% in 2027. India is now underway with its “National Electricity Plan 2023 to 2032”, which will see $100 billion spent to expand its grid by roughly 30%. China’s State Grid is expected to spend roughly $560 billion to upgrade the nation’s electrical grid between 2026 and 2030, according to the state-run Xinhua news agency. And last year, Goldman estimated that Europe needs “$3.5 trillion worth of investment in the power sector…over the next 10 years to reduce the risk of a European power crisis.” Last year, electricity system investment globally totaled $1.5 trillion, almost 50% higher than investment in fossil fuel supply.

The megatrends driving seismic electricity investment extend far beyond the AI buildout. Electrification is proliferating across transportation, buildings, and manufacturing. Geopolitical instability is intensifying energy independence ambitions across regions, pushing nations towards clean electricity dependence over fossil-fuel dependence. A warming planet is increasing cooling needs (electricity demand growth spiked to 4.4% in 2024 in large part due to record global temperatures). And the emerging world continues to urbanize, and urban areas account for a disproportionate percentage of global electricity consumption—roughly 58% of the world’s population lives in urban areas and those areas account for 75% of primary energy consumption.

Yes, a significant pullback in the AI buildout would lower electricity-demand growth expectations and therefore, drag on the performance of stakeholders in that demand growth. But again, that’s largely a US story. Across the globe, we see mounting evidence of an electrification inflection point, with enabling technologies maturing at the same time geopolitical and climate pressure points are intensifying. The price of solar power has fallen by more than 80% over the past decade and a half. Battery grid-storage costs have fallen by three times in just the past three years. The nuclear power renaissance is underway, with roughly 200 reactors now either under construction or planned, which could increase the total number of operational reactors globally by close to 50% over the next decade. Meanwhile, on the demand side, EVs are expected to account for 30% of total worldwide car sales in 2026, up 5% versus 2025 and 100% versus 2022, according to Our World in Data. Electricity demand for space cooling is projected to nearly double over the next decade, from 2,900 terawatt-hours in 2025 to more than 4,500 terawatt-hours by 2035. And at 3.6%, industrial manufacturing’s electricity demand growth is outpacing global demand growth as factories electrify amidst accelerating reshoring.

Undoubtedly, there have been challenges in translating the electrification theme into alpha. Persistent inflation has dragged on performance given the rate sensitivity of key stakeholders, whether utilities or solar and wind leaders. Meanwhile, a cross-continent political backlash alongside pervasive “greenwashing” by fund managers compromised sentiment towards sustainable investing, dragging on clean energy ETF inflows, which remained largely negative from 2022 through 2025 (chart on the left). Then, of course, there’s China’s domination of the innovation frontier, and the fact that Chinese stocks do not return value to international shareholders near regardless of growth rates (chart on the right).

We view that sentiment depression as an opportunity. This is especially true with the Ukraine and Iran wars likely to continue to elevate the national security imperative of electrification. And it’s also true with a record-setting El Niño brewing in the Pacific, threatening extreme heat and devastating natural disasters, which could intensify the climate imperative of electrification.

But even before considering those bigger-picture factors, we believe investors have always boiled the electrification theme down too much. They’ve neglected the breadth of beneficiaries, fixating on the energy generators (solar, wind, and nuclear), the electricity deliverers (utilities), and the most enticing consumer-facing innovations (EVs most of all). That narrow focus hasn’t been helped by excessive inflow concentration into US equities. Electrification is a global theme generating growth opportunities across regions in a multitude of sectors. US equities now account for roughly 60% of the world’s total stock market capitalization, up more than five percentage points in just the past five years. That US centrism has not only dragged on non-US stock performance, but driven a reduction in the number of analysts covering non-US equities, likely resulting in exploitable neglect of compelling growth opportunities, including electrification stakeholders.

Source: Acadian

Since launching Sage Road over a year ago, we’ve advocated for diversification via greater exposure to non-US equities. The MSCI EM, Stoxx 600, and MSCI Japan have outperformed the S&P 500 by roughly five percentage points combined over that time period, as of writing this. We’ve also advocated for replacing AI trade exposure with greater exposure to neglected sectoral beneficiaries of AI and themes beyond AI. Swapping AI picks-and-shovels exposure with healthcare exposure has been one of our key convictions recently and since publishing our “AI Trade” report in September, the SOX index has fallen more than 10% while the XLV has gained nearly 11%. With our concern about the fragility of the AI trade only intensifying, we believe now more than ever, it’s essential to look for non-US and non-AI-trade opportunity. And we believe electrification is a key thematic lens to look through for that opportunity, employing a broad perspective across the electricity generators and deliverers, equipment providers, consumer goods innovators, software enablers, materials stakeholders, and second-order beneficiaries.

So, we dug deep into both the supply and demand sides of the electrification equation. We studied the innovation frontier and the risk factors that could hold back progress. We believe it’s more important now than ever to understand the global electricity supply/demand imbalance. We expect it will prove a persistent contributor to stagflationary pressure across the globe. It will also continue to propel the public backlash against AI data centers as consumers increasingly confront rising electricity bills—again, especially in the US. Obviously, with a global trend on the scale of electrification, holistic encapsulation is an impossibility. What follows is a playbook for identifying growth opportunities we believe can mitigate risk and generate alpha in an investment environment increasingly vulnerable to an AI sentiment rupture.