UBS's latest report points to a shift in the focus of AI investment from model development and computing capacity to infrastructure. The stability of power supplies, securing data center capacity and access to critical raw materials are cited as factors that will determine the next phase of growth.
UBS's report points out that the logic of AI investment is shifting away from whether demand will materialize to whether the supply side can keep pace with expansion. Where will the power come from? Can data centers be built as scheduled? And will companies controlling critical raw materials gain pricing power? These are the issues under discussion.
Catherine Gordon, an energy and resources specialist at UBS, said that long-term power purchase agreements (PPAs) further underscore the importance of power supplies and that infrastructure-related assets will become a focus of the market. Big Tech companies are signaling to the market that companies able to secure stable energy through contracted supply will hold the upper hand.
The sustainability of AI capital expenditure depends on data center delivery times and the supply chain's ability to scale, while stronger funding capacity is seen as a factor determining execution feasibility.
UBS analyses control of assets that are bottlenecks, including grid expansion, data center construction and critical materials supporting domestic reshoring, as a focus of investment decisions. Supply arrangements for metals such as rare earths and tungsten could affect the stability of long-term investments.