IMF Says AI Needs More Power as IEA Sees Data Center Demand Surpassing Japan by 2030

The narrative around AI and energy is shifting from a cyclical shock to a structural, multi-decade reordering. This is not a temporary spike in costs, but the foundation for a new, persistent dynamic where inflation and growth are inextricably linked. The evidence points to a fundamental reconfiguration of global demand and supply, with implications that will outlast any single business cycle.

At the core is an explosion in electricity demand. The International Energy Agency projects that electricity demand from data centres worldwide is set to more than double by 2030, with the demand specifically from AI-optimised data centres projected to more than quadruple by 2030. This isn't just growth; it's a transformation of the power sector's trajectory. In advanced economies, data centres are projected to drive more than 20% of the growth in electricity demand over the next decade, putting the sector back on a growth footing after years of stagnation. The scale is staggering: by 2030, AI-driven power needs could consume as much electricity as the entire country of Japan today.

This surge in energy demand is the physical manifestation of a broader economic shift. The IMF analysis reveals that AI has the potential to raise the average pace of annual global economic growth. The implication is clear: the very engine of future growth is also the primary driver of a new, sustained inflationary pressure. This creates a dual dynamic where productivity gains from AI are counterbalanced by rising costs for the energy that powers it. The result is a new equilibrium where higher energy prices become a structural feature of a more productive global economy.

The bottom line is that this is a framework, not a forecast. The IEA and IMF analyses highlight significant uncertainties-on adoption rates, efficiency gains, and supply responses. Yet the central thesis holds: AI is establishing a persistent, multi-decade demand for electricity that will reshape energy markets and, by extension, the cost of doing business and the trajectory of inflation. The coming decade will be defined by managing this new, intertwined relationship between growth and energy costs.

The Infrastructure Investment Paradigm

The structural shift in energy demand is actively reshaping where and how capital is deployed. The market is no longer just reacting to AI's growth; it is proactively building the physical backbone to support it, creating a new investment paradigm. This is a fundamental reallocation of resources, where data centre operators are no longer seeking power-they are 'following the power' and approaching renewable energy generators to build new facilities directly on existing renewable sites. This strategic siting is a clear signal that the era of building massive data centres in energy-poor locations is ending. The new logic is one of co-location and efficiency, where proximity to generation minimizes grid strain and cost.

This physical convergence is unlocking a vast new opportunity set across the entire infrastructure value chain. The exponential growth in AI-driven data processing is a direct catalyst for investment in electricity and fibre networks, as well as the data centres themselves. The scale is immense, with US data centre electricity demand alone projected to more than double by 2028. This creates a multi-decade pipeline for capital, from new power generation and transmission projects to the dense fibre-optic backbones required to connect them. For investors, this is the essence of a structural opportunity: a durable, long-term demand for critical assets that are now being redefined as part of a digital-physical ecosystem.

Furthermore, this infrastructure boom is being amplified by a parallel geopolitical and industrial policy trend. As highlighted in IFM's analysis, governments are focusing on energy independence and boosting advanced manufacturing capabilities. This push for onshoring vital industries-be it semiconductor fabrication or next-generation battery production-will inevitably require the same robust infrastructure of power, data, and logistics. The result is a dual engine for investment: one driven by the digital economy's insatiable appetite for compute, and another by the physical need to secure supply chains and energy sources. This convergence is redefining the infrastructure asset class from a collection of static utilities into a dynamic, growth-oriented portfolio of assets that are central to both economic productivity and national security.

Financial and Policy Implications

Translating this structural shift into financial reality requires a new lens on valuation and risk. The core challenge is clear: countries must accelerate new investments in electricity generation and grids to meet the data centre demand surge. The IEA report explicitly warns that countries need to accelerate new investments in electricity generation and grids to meet data centre demand. This isn't a future possibility; it's an immediate capital allocation imperative. For financial markets, this means the cost of capital for power and grid projects may be re-rated, as the risk of supply bottlenecks and price volatility rises. Conversely, it creates a powerful tailwind for companies and funds that can execute on this build-out.

The investment thesis is now twofold. First, there is the direct opportunity to build the physical assets-power plants, transmission lines, and data centres themselves. Second, and more subtly, is the potential for AI to enhance the financial and operational performance of these very assets. Implementation of AI is expected to power improvements in efficiency, safety, and revenue optimisation across infrastructure networks. This creates a virtuous cycle: the growth in demand justifies the massive capital expenditure, while AI-driven operational gains can improve the returns on that capital over the asset's multi-decade life.

Yet this complex, interdependent system introduces a new class of systemic risks. The IEA notes that cyberattacks on energy utilities have tripled in the past four years and become more sophisticated because of AI. This is a critical vulnerability that extends beyond individual companies to the stability of entire energy and digital networks. A holistic risk management framework is therefore not optional; it is foundational. Investors must look beyond traditional credit and market risks to assess cybersecurity resilience, supply chain concentration for critical minerals, and the potential for cascading failures in a tightly coupled system.

The bottom line is that the financial and policy landscape is being redefined. Valuation models for energy and infrastructure assets must now embed a multi-decade demand curve driven by AI, while also accounting for the heightened operational and systemic risks. Policy, in turn, must evolve to de-risk this transition-through streamlined permitting, strategic stockpiling of critical minerals, and robust, forward-looking cybersecurity standards. The companies and nations that navigate this dual imperative of massive investment and sophisticated risk management will be best positioned to capture the returns of the new, energy-intensive digital economy.

Catalysts and Watchpoints

The structural thesis hinges on a few critical, observable signals. The near-term catalyst is the pace of new renewable energy capacity additions versus data center demand growth. If the build-out of solar and wind power cannot keep up with the electricity demand from data centres worldwide is set to more than double by 2030, grid strain will intensify, leading to higher wholesale power prices and localized bottlenecks. This mismatch is the primary inflationary pressure valve. Investors should monitor regional grid operators' capacity reports and renewable project permitting timelines for early warnings of supply-demand imbalances.

Government policy announcements are the next major watchpoint. The scale and timing of public investment in energy security and AI infrastructure will determine whether the private sector can execute at the required speed. The IEA report notes that countries need to accelerate new investments in electricity generation and grids. Watch for concrete fiscal commitments, streamlined permitting for power lines and data centres, and industrial policies aimed at boosting advanced manufacturing. These signals will gauge the political will to de-risk the massive capital expenditure required.

Finally, operational efficiency gains in data centres are a crucial counterweight. The thesis assumes that AI will not only drive demand but also improve the efficiency of the assets that consume it. As noted, implementation of AI is expected to power relevant improvements in financial and operational performance of infrastructure assets. Failure to achieve these gains-through better cooling, server utilization, or predictive maintenance-would amplify the inflationary and capital intensity pressures over the decade. Monitor industry reports on data centre power usage effectiveness (PUE) trends for evidence of this critical offset.

The bottom line is that the multi-decade framework is being tested in real time. The convergence of physical infrastructure build-out, policy support, and technological efficiency will validate the investment opportunity. Any deviation in these three areas will challenge the projected equilibrium.