35-50% of planned AI data centres are behind schedule

“This gap will have significant consequences for energy markets, climate progress, and AI development,” says lead author Dr Antif Ansar, Programme Director of the Oxford Programme on the Sustainable Future of Capital Intensive Industries.

A key finding of the paper is that markets appear to be pricing in the announcement curve, not the delivery curve. Commodity and equipment markets (copper, transformers, electrical steel) are responding to lofty buildout claims that may materialise years late or not at all, risking short‑term overpricing followed by oversupply.

“Media and markets are expecting a rapid, exponential ramp-up of Al infrastructure. The evidence tells a different story: a rollout that is delayed, lumpy, and constrained by physical realities,” says co-author Dr Stephen Lezak. “Our analysis suggests a significant share of announced capacity is inflated or postponed - ‘bragawatts’, not real assets."

“If enterprise AI adoption does not scale as expected, the result could be stranded assets or sharply lower utilization,” warns co-author Dr Guy Wolf, Global Head of Market Analytics at Marex. 

The white paper details how execution challenges are reshaping energy markets by raising local power costs and driving increased natural gas usage. While power grids struggle to catch up, operators are increasingly turning to on-site gas turbines, fuel cells, and battery systems. By 2028, off-grid generation for data centres could account for a substantial uptick in natural gas usage not captured in official power sector statistics.

The authors conclude by warning that the economic case for the scale of buildout remains unproven: “The winners will be those who balance ambition with realism – investing in the future of AI without losing sight of the practical constraints of the present.”

Download the white paper "The Great AI Infrastructure Buildout: Impact on Power and Commodity Markets."

The views expressed in this white paper represent those of the authors and do not necessarily represent those of the Oxford Smith School or any other institution or funder. The paper is intended to promote discussion and to provide public access to results emerging from our research. It may have been submitted for publication in academic journals. It has been reviewed by at least one internal referee before publication. 

This report forms part of an ongoing research programme on the sustainable future of capital intensive industries and commodities funded by Marex, a diversified global financial services platform.

The Smith School of Enterprise and the Environment at the University of Oxford equips enterprise to achieve net zero emissions and the sustainable development goals, through world-leading research, teaching and partnerships.

Marex Group pie (NASDAQ: MRX) provides market access, infrastructure services and essential liquidity to clients across global commodity and financial markets. The Group provides comprehensive breadth and depth of coverage across four services: Clearing, Agency and Execution, Market Making and Hedging and Investment Solutions. It has a leading franchise in many major metals, energy and agricultural products, with access to more than 60 exchanges. Marex has over 3,400 active clients, including some of the largest commodity producers, consumers and traders, banks, hedge funds and asset managers. With more than 50 offices worldwide, the Group has over 3000 employees across Europe, Asia and the Americas. For more information visit www.marex.com.