Goldman Sachs forecast a 7 percentage point average decline in Big Tech return on equity over the next year, warning that AI infrastructure spending is outpacing revenue generation.1
The June 12 report centers on Microsoft, Alphabet, Meta, Amazon, and Apple — the five largest AI infrastructure investors.1 These companies dominate S&P 500 weighting and collectively set the tone for market earnings expectations.
AI hardware and infrastructure investment cycles typically lag revenue generation by 18 to 24 months.1 That structural gap means capital expenditure compresses margins before monetization begins to offset the cost. The bigger the upfront spending, the longer the drag.
Return on equity measures how efficiently a company converts shareholder equity into profit. A 7-point compression across five of the most profitable businesses in the index would represent a notable deterioration in capital efficiency — one that current valuations may not reflect.
The S&P 500 returned 9% year-to-date on the back of earnings growth.1 Market pricing assumes continued profitability at the largest tech firms. Goldman's forecast challenges that assumption directly.
The counterargument rests on timing. AI proponents argue monetization will outpace prior infrastructure cycles. Cloud computing in the 2010s took years to turn margin-accretive; the case is that AI services — inference, APIs, enterprise tools — will close that gap faster.
Confirming the Goldman thesis requires tracking quarterly ROE across all five companies over four consecutive quarters. A verified average decline of 5 to 9 percentage points would validate the projection.1
The implications reach beyond individual earnings. ROE compression at this scale would ripple through index returns, analyst estimates, and institutional capital allocation. AI infrastructure spending shows no sign of slowing — making the revenue side of that equation the critical variable to watch over the next 12 months.