Bridgewater: AI Infrastructure Trade Largely Priced In, Minimal Positioning Shift Toward Application Plays
nashnova research
Bridgewater, one of the world's largest hedge funds, has cut its AI infrastructure position to near zero. Co-CIO Greg Jensen says most upside is already priced in — signaling a broader institutional pivot from the "picks and shovels" layer toward application and disruption plays.
Why does Bridgewater call AI infra "no longer a great trade"?
Co-CIO Greg Jensen told *The Information*: "Two years ago we thought this was a great trade, but most of the expectation is now priced in."
This means → the run-up in Nvidia, data-center REITs, and other "AI picks-and-shovels" assets has already absorbed most of the demand story.
Bridgewater's current AI-infra position is very small; the fund is now focused on disruption and application-adoption trades.
What does Bridgewater's model show?
The fund has modeled global data-center build-out and its supply-chain impact through 2028, and is already building a 2029 model.
Jensen sees the market as slightly underestimating 2028 build scale — but the gap is no longer large.
In plain terms = even in the most bullish scenario, data centers get built a bit more than the market expects — but that sliver of upside no longer justifies a big bet.
What risks remain in the infra trade?
Jensen flags that AI "shovel" assets — chips and other infrastructure — face financing challenges and construction delays alongside demand upside.
This means → even if demand holds, capital availability and execution risk could derail the timeline.
His outlook also assumes no major disruptions — trade friction or policy shifts would be additional headwinds.
Where is the money going?
Bridgewater explicitly says it is more interested in disruption and application-adoption trades.
In plain terms = the "who builds the infrastructure" question is priced; the next question is "who turns that infrastructure into revenue?" — and that is where Bridgewater is looking.
This reflects a broader institutional consensus shift: the AI investment battleground is moving from building the stack to monetizing the stack.
市场有风险，内容仅供研究参考，不构成投资建议。