Goldman Sachs warns that surging AI infrastructure spending and government debt are driving up the cost of capital, creating a potential earnings bubble that threatens tech valuations.
Nvidia fell over 3% and led chipmakers down sharply after AI safety concerns, and if AI capex growth slows, chip suppliers have no comparable fallback to hyperscalers.
Goldman Sachs' chief strategist downgraded the near-term outlook on stocks and warned of a potential AI-driven earnings bubble, with concentration risk and rising capital costs pressuring the index.
The Philadelphia Semiconductor Index fell nearly 6% after calls to slow AI development, illustrating chipmakers' vulnerability to any slowdown in AI infrastructure spending.
Microsoft is highlighted as a hyperscaler planning to triple data center capacity, with the ability to harvest returns if AI progress slows, but it faces risks from rising capital costs and an AI earnings bubble.
Meta is cited as a hyperscaler funding AI infrastructure, and while it can pause capex if needed, the broader earnings-bubble risk and capital competition create uncertainty.
Alphabet is mentioned among hyperscalers that rose despite AI slowdown calls, benefiting from optionality to reduce capex, but remains exposed to earnings-bubble concerns and cost-of-capital pressure.
Goldman Sachs strategist Peter Oppenheimer warns that an AI-driven earnings bubble may be forming as record infrastructure spending competes with government borrowing for limited capital. While tech balance sheets remain robust, rising interest rates and the high cost of capital threaten to pressure equity valuations if profit growth slows.
Oppenheimer argues that tech valuations are not currently at extreme levels, but earnings could collapse if the high cost of capital and slowing AI demand undermine the profitability of massive infrastructure investments.