Vanguard warns of AI capex risks; Bank of America (BAC) tightens underwriting and risk models

Headline: Vanguard warning prompts banks to tighten underwriting and risk models for AI-driven capex

Vanguard flags unseen risks from AI spending

Vanguard’s senior economist Shaan Raithatha warns that a rapid surge in capital expenditure tied to artificial intelligence is creating concentrated, specialised asset buildups that may embed unseen financial and operational risks. The brief cautions that heavy spending on data centres and specialised hardware can lead to asset misallocation, stranded capacity and heightened supply-chain concentration.

Bank credit and risk management take centre stage

For large lenders such as Bank of America, the note heightens urgency to revise credit underwriting and portfolio stress-testing for borrowers funding AI infrastructure. Banks are increasingly financing data-centre builds, bespoke semiconductor procurement and related services, and the potential for rapid technological obsolescence means loans secured by niche assets could lose value faster than traditional collateral models assume. Bank of America is therefore likely to adapt its lending criteria, incorporate scenario analysis for model and architecture shifts, and tighten covenants that govern use of proceeds and asset maintenance.

Operational and systemic exposures concern corporate banking desks

The AI capex wave also shifts non-credit risks that matter to corporate banking and transaction services. Elevated energy consumption, complex global semiconductor supply chains and concentration around a small set of suppliers or regions amplify operational and reputational exposures for banks that provide treasury, custody and payments services to affected clients. Bank of America’s risk teams are likely to extend due diligence beyond financial metrics to include supplier diversification, cyber resilience and environmental impact when assessing relationships with heavy AI spenders.

Implications for disclosure and supervisory oversight

Vanguard’s warning underscores calls for clearer corporate reporting on AI capital plans and timelines. Banks, as intermediaries and fiduciary partners, face pressure to demand transparency from corporate clients and to feed consolidated intelligence to regulators. Supervisors may respond by monitoring concentration in lending books, encouraging sector-specific stress tests and revising capital guidance where systemic dependence on a narrow set of suppliers or locations emerges.

Environmental, fiduciary and investor stewardship angles

The note also highlights energy use and ESG considerations that affect banks’ stewardship roles and pension-fund clients. For custodial and asset-management services, including those offered by major banks, there is growing demand for scenario analysis that accounts for both productivity upside and longer-term fragility associated with AI capex — prompting a wider reassessment of governance, engagement and portfolio resilience.