Lenders and CFOs Tighten AI Financing

DBS Group Holdings CEO Tan Su Shan said markets are reassessing the risks associated with financing artificial-intelligence-related capital expenditure. She described investors as seeking higher returns and paying closer attention to deal structures as they commit capital to the AI ecosystem.

Tan characterized this shift as a form of discipline in how money flows into AI projects. Investors are focusing on whether long-term returns from AI investments will be sufficient to service obligations to debt, bond and equity holders.

DBS has arranged approximately S$25 billion in AI-related deals this year, often acting as lead manager on syndicated loans. The bank has retained only a fraction of this exposure on its own balance sheet, distributing much of the risk among participating lenders.

The financing that DBS is arranging is concentrated in hardware, including semiconductors, memory chips and data centres. This reflects significant capital needs in the physical infrastructure that underpins AI applications.

As part of the repricing of risk, investors are asking for higher returns on AI-related financings. They are also paying more attention to transaction structures, aiming to ensure that projects can generate sufficient cash flows over time.

Tan noted that this environment is prompting more rigorous scrutiny of how AI capital expenditure is funded. The emphasis is on aligning financing terms with credible expectations of long-term return on investment.

Corporate finance chiefs are also tightening control of AI spending. One of their main concerns is the cost of tokens and usage-based pricing models, which have led to AI bills coming in higher than expected.

In response, CFOs are setting clearer budgets for AI and managing them more centrally. They are working to contain unexpected cost escalation tied to AI workloads and usage patterns.

Finance leaders are increasingly centralizing oversight of AI initiatives rather than allowing spending to be dispersed across business units. This centralization is intended to improve visibility and control over aggregate AI costs.

CFOs are also demanding measurable return on investment from AI projects. This means closer evaluation of business cases and tighter controls to ensure that AI deployments deliver quantifiable financial benefits.

The combination of stricter investor demands in capital markets and tighter corporate budgeting suggests a more cautious environment for AI-related capex. Lenders are rebalancing risk and return expectations, while CFOs are curbing unchecked spending.

This convergence is likely to influence the pace and structure of future AI infrastructure projects, especially in hardware-intensive areas like semiconductors and data centres that require substantial upfront financing.