Goldman Sachs: AI Investment Crowding-Out Effect Exists but Is Limited
Nashnova编辑部
Goldman estimates U.S. AI investment at roughly $600 billion this year — it is crowding out other activity, but far less than the market assumes. Both the bulls and the bears may be mispricing.
$600 billion into AI — how big is that, really?
Goldman economists Jessica Rindels and David Mericle peg this year's AI investment at roughly $600 billion, or about 2% of U.S. GDP.
Drill down: that is 10% of business fixed investment and 15% of equipment spending. This means → roughly one in every six to seven dollars of corporate equipment spend now flows to AI.
In plain terms = the number is large, but it is nowhere near "swallowing the whole economy."
Where exactly is the squeeze happening?
Path one: IT budgets reshuffled internally. Firms paying new AI-compute bills cut other software and tech spending — but this is mostly money moving from one pocket to another, with limited net GDP impact.
Path two: construction labor and equipment pulled toward data centers. Data-center projects carry gross margins more than double those of non-tech construction, so workers and machinery migrate to the higher-paying sites. Other building projects get squeezed.
Path three: borrowing costs tick up slightly. Hyperscalers — Google, Amazon, Microsoft and peers — have issued large volumes of debt, nudging overall corporate borrowing costs higher. Goldman's estimate: roughly 0.05 percentage points, suppressing about $10 billion in non-AI investment. A mild effect.
If the crowd-out is small, what is the market worried about?
Rindels and Mericle are blunt: "Media reports and market commentary often claim that AI is contributing enormously to GDP growth while also crowding out other activity significantly — our analysis suggests both claims are overstated."
This means → it is not just the bears overpricing crowd-out risk; the bulls are also overpricing AI's near-term economic boost.
This reflects a deeper signal: the market's current AI pricing — whether the "AI changes everything" narrative or the "AI bubble crushes the economy" fear — may be off on both ends.
Content is for reference only, not financial advice.