CICC: How to Hedge Against AI Risk?

CICC Research

We remain optimistic about the return prospects for AI assets and recommend using asset allocation to cope with recent market volatility—essentially "buying insurance" for AI assets. Beyond defensive assets such as cash, bonds, and high-dividend stocks, the dollar-depreciation trade can also help hedge AI asset risk, and increasing gold allocations may be an effective way to express that trade.

Against the backdrop of a restructuring international monetary order, we believe "de-dollarization" remains the broad trend, and the dollar faces long-term depreciation pressure ("Is De-dollarization Over?"). However, since 2026 the dollar's trajectory has been temporarily strong; aside from inflation pressure triggered by geopolitical conflicts, the AI rally may be providing key support. On one hand, cross-border transactions in the AI supply chain—advanced chips, cloud services, and computing power—are often settled in dollars, so industrial expansion may temporarily increase dollar demand, creating a so-called "AI dollar" effect. On the other hand, AI investment strengthens the growth momentum of the relevant economies, especially the United States, where nearly 30% of the increase in real GDP in 2026 is expected to come from AI. The resilience of economic growth creates conditions for the Fed to raise rates, and rising Treasury yields push the dollar higher. But for economies that benefit less from AI, tighter external financing conditions and capital-outflow pressure brought by a stronger dollar will instead drag on their economies and market performance.

Under the influence of these mechanisms, the dollar has become the dividing line in the divergent performance of global stock markets amid the AI revolution: since 2026, the stock markets of the US, Japan, and Korea—which are deeply involved in the AI supply chain—have been positively correlated with the dollar, while the Indian stock market and the Hang Seng Index, which benefit relatively little from AI, have been negatively correlated with the dollar.

Chart 1: Since 2026, the Nikkei 225 has been positively correlated with the dollar

Source: Wind, CICC Research Department

Chart 2: Since 2026, Korea's KOSPI has been positively correlated with the dollar

Chart 5: Since 2026, correlation between the dollar index and major asset moves: the dollar has become the "dividing line" in global stock market divergence

Note: Data for various assets are as of October 8.

Source: Wind, CICC Research Department

Although the AI revolution is still in its early stages, the market has begun to worry about slowing growth at leading players and whether future returns can cover current large-scale investment. The market expects the year-on-year growth rate of capital expenditure by the US AI "Magnificent Seven" to fall from about 90% in 2026 to 36% in 2027, and uncertainty over investment returns has also intensified AI asset volatility.

Chart 6: The market expects AI capital expenditure to keep growing in 2027, but at a markedly slower pace

Source: Bloomberg, CICC Research Department

Against this backdrop, increasing allocations to defensive assets such as cash, Chinese bonds, and high-dividend stocks is reasonable. Beyond that, based on the analysis above, a cooling of the AI trade may weaken the US's relative growth advantage and dollar demand, and the dollar-depreciation trade is also expected to hedge the adjustment pressure on AI assets.

Another advantage of the dollar-depreciation trade is that its profit scenarios are not limited to a weakening AI rally: if easing US-Iran tensions push oil prices down, or if US inflation improves markedly and prompts the Fed to shift toward rate cuts, AI assets would benefit from improved liquidity and the dollar-depreciation trade would benefit as well—and AI stocks could also rise in tandem. Therefore, compared with simply reducing AI holdings, positioning for the dollar-depreciation trade can both retain exposure to AI upside and provide portfolio protection across multiple scenarios.

The main risk to this hedging logic is a dollar liquidity crisis. If a sharp AI selloff triggers concentrated global asset selling and investors rush to cash out and obtain dollar cash, the dollar could strengthen and move inversely to stocks, rendering the hedge ineffective. But the current AI adjustment mainly reflects a decline in risk appetite and asset reallocation driven by concerns over investment returns, and has not yet evolved into a disorderly liquidity shock; the dollar and tech stocks are more likely to be positively correlated. During the recent adjustment in US tech stocks, the dollar index also fell back from its October 14 high, consistent with this judgment.

The above analysis focuses mainly on the short-term relationship between exchange rates and stock markets; over a longer horizon, even if the AI revolution succeeds, stock gains may not necessarily be accompanied by dollar appreciation, because the "reverse Balassa effect" brings real exchange rate depreciation pressure (see "The 'Reverse Balassa Effect' Under the AI Shock" for details).

Gold is an effective tool for positioning for the dollar-depreciation trade, and we recommend increasing allocations

The most direct way to position for the dollar-depreciation trade is to short the dollar, but foreign exchange trading faces many restrictions, and increasing gold allocations may be a more convenient way to express the strategy. In theory, gold, as an alternative asset to the dollar system, has a fundamental negative correlation with the dollar, but because gold prices are also affected by multiple factors such as interest rates, inflation, and safe-haven demand, this relationship has not always been stable historically. But entering 2026, the negative correlation between gold and the dollar has strengthened markedly, with the correlation rising to -81.2%. The dollar and gold have become two sides of the same coin, making higher gold allocations an effective tool for the dollar-depreciation trade.

Chart 7: Before 2026, the correlation between gold and the dollar was not stable

Source: Wind, World Gold Council, CICC Research Department

Chart 8: Since 2026, the negative correlation between gold and the dollar has strengthened markedly, with a correlation coefficient as high as -81.2%

Source: Wind, World Gold Council, CICC Research Department

By contrast, the impact of oil prices and US Treasury yields on gold is more of a temporary disturbance: sharp oil price swings during the year did not consistently dominate gold price moves, and rising US Treasury yields did not always suppress gold performance.

Chart 9: Oil prices fell in May-June but did not benefit gold; oil prices surged in August but did not suppress gold

Source: Wind, World Gold Council, CICC Research Department

Chart 10: After July, US Treasury yields accelerated upward, but this did not prevent gold from rising in August

Source: Wind, World Gold Council, CICC Research Department

Besides gold, US Treasuries are also a potential option for positioning for a weaker dollar, but the win rate may be lower than for gold: if dollar depreciation stems from falling inflation and Fed rate cuts, both gold and US Treasuries may benefit; if dollar depreciation stems from rising US debt risk or eroding confidence in the dollar, gold may benefit from demand for alternative assets, while US Treasuries may fall due to rising term premiums.

Historical review also supports this judgment. We reviewed all 91 dollar depreciation episodes lasting at least 14 days since 2000; gold rose in 84 of them, a win rate of 92.3%, with a median return of 5.46% over those periods; over the same periods, 2-year and 10-year US Treasury yields fell in 68.1% and 64.8% of cases, respectively.

Chart 11: During dollar index declines since 2000, gold's win rate has been higher than that of other major asset classes

Source: Wind, CICC Research Department

Source of this article: CICC Insight Risk Disclosure and Disclaimer

Markets carry risk, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. Investors bear responsibility for their own investment decisions based on this article.