Wall Street is openly doubting AI’s debt-fueled buildout. Projected AI infrastructure spending through 2030 far exceeds expected revenue. Over $4 trillion of that spending could face write-downs by 2032. More than 80% of specialized AI cloud providers may fail by 2030. Global AI-related debt issuance fell from $113 billion in June to $23 billion in September 2026. Banks are tightening credit to avoid concentration risk. A $18 billion Oracle data center loan package slid into stressed territory. Nvidia-backed Firmus shelved a $30 billion IPO. CoreWeave’s stock and debt sold off sharply before rebounding. One third of AI capex is now debt-funded. Companies are adopting open-weight models to cut compute costs. AI adoption may take a decade or two to reshape operations. Anthropic turned free cash flow positive in Q3.
What changed
AI infrastructure spending was largely cash-funded with limited debt concerns.
What it unlocks
Wall Street is openly questioning whether AI spending can generate sufficient returns.
- $7 trillion projected AI infrastructure spend through 2030
- $3.6 trillion projected AI-related revenue through 2030
- $4 trillion AI infrastructure spend at risk of write-downs by 2032
- 80% of specialized AI cloud providers may cease operating independently by 2030
Sources