AI enters its political phase: implications for US credit

US politics is likely to make AI-related US credit a more selective, issuer-specific opportunity rather than a broad sector allocation.

• The November US midterms could reveal which AI concerns resonate with voters, as bipartisan tensions grow around infrastructure costs, data centres and state-level regulation despite broad support for competing with China.
• If national security and technological leadership prevail, bottlenecks in power, chips, buildings and skilled labour could prolong the buildout and support issuance from hyperscalers, utilities, data-centre developers and related financings.
• An infrastructure backlash would raise permitting, completion and refinancing risks, potentially pulling capital expenditure and debt issuance forward while increasing the dispersion between completed assets and speculative projects.
• Labour shortages could initially raise wages, construction costs and inflation, while later enterprise adoption might displace workers and weaken consumption, banks and lower-quality credits; the rates impact could be weaker front-end employment alongside a higher long-term neutral rate.
• AI financing already depends on public bonds, securitised debt, private capital and hybrid structures, with Goldman Sachs tracking $578 billion of global AI-related debt issuance through mid-September 2026; elevated risk-free yields and choppier deal receptions show that capital is not unlimited.

Takeaway: Investors should favour well-capitalised hyperscalers and data-centre assets with contracted cash flows, attractive valuations and visible funding, while demanding compensation for supply indigestion, refinancing, weaker earnings quality and political or infrastructure constraints.