Bridgewater: Most AI infrastructure trades are already priced in, only holding a "very small position"

Bridgewater Associates, one of the world's largest hedge funds, is seeing its enthusiasm for AI infrastructure investment themes cool. Bridgewater's co-chief investment officer Greg Jensen stated that most of the upside potential for AI infrastructure-related trades has already been priced in by the market, and the fund is now turning its attention to other investment opportunities.

According to technology media outlet The Information, Jensen candidly said, "Two years ago this was a great trade, but now most of it is already priced in." He revealed that Bridgewater currently holds "a very small position" in the direction of AI infrastructure and is shifting its investment focus toward transactions related to AI-driven "disruption and applications." This statement marks a significant strategic adjustment in the AI investment logic of this major hedge fund.

For the market, Bridgewater’s change in stance carries important reference value. AI infrastructure concept stocks—especially those in the chip and data center supply chains—previously benefited from expectations of a surge in demand for computing power, resulting in rising valuations. Bridgewater’s signal of reducing its position means that this "picks and shovels" investment logic, at current valuation levels, can no longer provide sufficient risk-reward.

Modeling Through 2028: Limited Upside and Persistent Risks

Jensen revealed that Bridgewater has systematically modeled the scale of global data center construction and its impact on various supply chain segments, with the time horizon extending to 2028, and has already started building forecast models for 2029.

He pointed out that demand growth for AI "picks and shovels" assets—typically referring to chips and other infrastructure—faces potential upside but is also confronted by practical risks such as financing challenges and construction delays.

Jensen remains cautiously optimistic about demand prospects for 2028, but chose his words carefully: "We still think the market may be slightly underestimating the actual construction scale for 2028, but the gap is not large, and this view is based on the assumption that there won’t be significant disruptions."

This means that even if there is some upside potential, the extent is quite limited and highly dependent on smooth progress at the macro and execution levels.

According to reports, Bridgewater's portfolio adjustment reflects its stage-based judgment regarding the AI investment cycle. Jensen clearly stated that the fund is currently "more interested in AI disruption and application trades", meaning a shift from betting on large-scale AI infrastructure build-out to focusing on opportunities for industry disruption brought about by the adoption and implementation of AI technology.

This shift in logic carries a certain forward-looking significance in the investment community: as the excess returns from infrastructure investment are gradually digested by the market, capital often starts seeking the next narrative that has not yet been fully priced in—that is, which industries will be impacted by the widespread adoption of AI applications, or which companies will be the first to benefit from embracing AI.

Analysis suggests that Bridgewater has not fully exited AI-related investments, but has chosen to rebalance its portfolio structure to cope with the reality of declining risk-return for infrastructure themes under the current valuation environment.