AI Computing Power Drives $9 Trillion Energy Investment Super Cycle

Overview

This report explores the profound impact of explosive AI computing power growth on energy security globally, especially in the Asia-Pacific region. Morgan Stanley believes that frequent energy shocks make energy and economic security crucial in the AI era. It is estimated that over $5 trillion in energy security investment demand will kickstart a 'super cycle', not only ensuring stability in AI, food, and technology supply chains but also unlocking asset value of up to $9 trillion. The report recommends investors focus closely on reliable energy assets, diversified energy source infrastructure, and bottleneck areas such as grids and metals.

Core views

The core logic of the report revolves around 'AI computing power demand reshaping energy supply chains', proposing five main investment themes and three market 'surprises': Regarding investment themes: First, 'reliable energy assets' capable of providing stable power for AI are core, including nuclear power, renewable energy, energy storage, as well as local oil & gas production and refining. Secondly, diversification of energy sources is critical; tankers, LNG import/export facilities, and midstream pipeline assets will benefit. Thirdly, nuclear power, fertilizers, and grid deployment will accelerate, and data centers can cross-subsidize residential electricity through tiered electricity pricing. Fourthly, oil & gas exploration is heating up, and the return of coal power will double coal gasification capacity, benefiting oil services and coal equipment manufacturers. Finally, power infrastructure construction will significantly drive demand for metals such as copper, aluminum, uranium, and nickel. Regarding market 'surprises': First, average annual energy investment in Asia will double by 2030. Second, coal returns to the stage to guarantee AI energy needs; Asian data center power demand's compound annual growth rate (CAGR) from 2023-2030 is projected to reach 24%, far exceeding US/Europe, with Asian coal-related investment reaching $318 billion. Third, energy procurement diversifies towards the Americas; US shale gas will be exported to Asia on a large scale for the first time in 2027, and new global LNG export capacity of over 1.2 billion tonnes/year will reshape market patterns. Additionally, the report emphasizes that insufficient grid capacity has become a key bottleneck requiring approximately $1 trillion in investment; meanwhile, accompanied by high penetration of renewable energy, it is estimated that Energy Storage Systems (ESS) will add approximately 2700GWh by 2030. The refining cycle has also become tighter due to capacity growth lagging behind consumption growth by 30-40%.

Analysis framework

The institution's analysis approach starts from the underlying variable of 'power demand surge brought about by explosive AI computing power', employing a 'supply-demand framework' and 'industry chain transmission' logic for deduction. First, by quantifying the CAGR of Asian data center power demand (24%), the urgency of energy shortage is established; second, breaking the traditional 'new energy only theory', pragmatically pointing out that under energy security demands, coal, oil & gas and other traditional 'reliable energy' sources and nuclear power will return and be revalued; finally, extending downstream and upstream along the industry chain to identify capital expenditure gaps in the grid (transmission bottleneck), storage (regulation bottleneck), and metals such as copper/uranium (material bottleneck), thereby constructing an investment framework covering multi-asset classes called a 'super cycle'.

Methodology notes

• Industry / Industrial Analysis FrameworkSupply-demand framework

  Supply-Demand Mismatch Analysis of Energy and Computing Power Needs

  The report derives the massive capital expenditure gap by comparing the explosive growth of power demand in AI data centers (e.g., 24% CAGR in Asia) with the status quo of insufficient investment in energy infrastructure over the past decade, using this as the core driver for energy asset revaluation and the start of a super cycle.

• Industry / Industrial Analysis FrameworkUpstream-Midstream-Downstream Industry Chain Transmission

  Industry Chain Transmission from Computing Power to Energy and Metal Materials

  Analysis is not limited to a single industry, but unfolds layer by layer along the chain of 'AI Computing Power -> Power Demand -> Generation/Grid/Storage Infrastructure -> Upstream Metals/Uranium/Fossil Fuels', deconstructing beneficiary links to help investors see how terminal demand translates into orders for upstream resources and midstream equipment.

Key data

• Value Unlocked by Energy Investment$9 TrillionTotal value expected to be unlocked from >$5 Trillion investment demand
• Asia Data Center Power Demand CAGR~24%Projected CAGR for 2023-2030, exceeding US/Europe
• Asia Coal Investment Scale$318 BillionExpected investment amount driven by power demand and coal-to-gas conversion
• Global New LNG Export Capacity>1.2 Billion Tonnes/yearNew capacity by 2030, driving US shale gas exports to Asia
• Grid Investment Demand$1 TrillionInvestment scale required to solve grid capacity bottlenecks
• New Energy Storage System (ESS) Scale2700GWhEstimated new installed capacity by 2030

Impact & implications

The report believes that the power supply chain has already undergone a 20-50% valuation re-rating over the past two years, but this is just the beginning; the re-rating story will extend to all 'reliable energy' sectors. Strong corporate balance sheets and up to $4 trillion in potential debt financing capability provide sufficient ammunition for this capital expenditure cycle lasting several years. For the market, this means investment logic must shift from simply 'green transition' to prioritizing both 'energy security and AI computing power assurance', with traditional fossil fuels, nuclear power, and critical metals receiving long-term structural tailwinds.