Apollo's chief economist: if the AI boom subsides, there is a fear of a sharp drop in the US dollar.

Apollo Global Management's chief economist Torsten Slok issued a warning on Monday: If the selling pressure on artificial intelligence-related stocks intensifies, the recent upward trend of the US dollar will face significant risks. Slok pointed out that in the past 12 months, international investors have been actively investing in the AI theme, leading to record-breaking rolling net inflows into the US stock market. However, most of these investors have not hedged their dollar exposure against exchange rate risks. If AI disappoints, leading to a retreat in capital inflows, it will pose significant downside risks to the US dollar, Slok wrote in a report released on Monday. This means that the US dollar is actually "hiding a dependence on AI trading". Implicit dependency of the US dollar on AI In his report, Slok pointed out that in the past 12 months, international investors have been actively investing in the AI theme, leading to record-breaking rolling net inflows into the US stock market. However, the vast majority of these investors have not hedged against foreign exchange risks. This detail constitutes the core mechanism of the US dollar's "implicit dependence": overseas investors need to exchange their local currency for US dollars before purchasing US stocks, providing important support for the US dollar. At the same time, compared to other major economies, the US maintains high interest rates, making the cost of purchasing hedging instruments for exchange rate risks too high, making the US dollar more susceptible to stock market fluctuations. Slok warned, "Therefore, if AI disappoints, leading to a decline in capital inflows, it will pose significant downside risks to the US dollar." This means that the US dollar is actually "hiding a dependence on AI trading". Deutsche Bank echoes: US dollar evolving from "safe haven asset" to "technology high beta" Analysts at Deutsche Bank's senior financial market analyst Malika Sachdeva echoed Slok's assessment. In a report to clients, she pointed out that global capital is flowing into US technology stocks at an unprecedented scale, fundamentally changing the risk attributes of the US dollar - it is gradually transitioning from a traditional safe haven asset to a "technology high beta" asset highly correlated with the NASDAQ 100 index. For a long time, the US's "twin deficits" (with a current account deficit of around $1.12 trillion in 2025) have been mainly offset by foreign official institutions and long-term invested funds purchasing US government bonds, which have a clear countercyclical feature. During economic recessions or risk asset retractions, funds tend to flow into US bonds and the US dollar simultaneously, forming a natural buffer mechanism. However, this buffering mechanism is eroding. Sachdeva warned, "If the funding mode shifts towards a more cyclical type of stock funds driven by retail investors, the US dollar exchange rate will become riskier, and its reliance on the super bullish market of AI-related stocks will deepen." Once AI profit expectations cool down and there is a deep correction in US stocks, foreign investors may simultaneously sell stocks and actively hedge their US dollar exposure, weakening the US dollar's safe haven support as a result. Record foreign capital inflows and unhedged risk exposure Foreign capital inflows into US stocks in 2026 have indeed reached unprecedented levels. Bank of America's strategy team cited EPFR data stating that for the week ending June 17, US stock funds saw a net inflow of $119.2 billion, setting a historical record. On an annualized basis, US stock funds are expected to attract a total of $739 billion in 2026, also setting a record. AI is the core drive of this flood of funds. Investors from Europe and Asia, among other regions, believe that their own markets lack AI-leading companies with scale and depth, so they can only invest in US stocks to ride the wave of technology. Funds are highly concentrated in AI infrastructure development and cloud computing-related assets. CMoney analysis pointed out that most overseas investors have not hedged against exchange rate risks - they must actually purchase US dollars while buying US stocks, further boosting the US dollar index. This means that the strength of the US dollar increasingly relies on capital inflows into the stock market driven by the AI theme, rather than just traditional interest rate or trade factors. However, this capital structure also implies significant potential risks. Once foreign capital loses confidence in the AI theme and starts selling stocks, the US dollar may also decline accordingly. Unlike traditional US bond financing, funds driven by retail investors in stocks are more cyclical, making the US dollar exchange rate riskier and its dependence on AI-related market movements deeper. Signals of cooling AI trading are increasing Slok's warning is not without basis. In recent times, there have been several signals of cooling in AI trading: AI concept stocks have experienced significant retractions from their highs. As of July 10, many AI concept stocks that had doubled in price year-to-date have retraced by over 20% from their highs, with some individual stocks experiencing retracement exceeding 40%. Tech stocks have faced continuous selling pressure. In early July, due to the news of Meta's plan to sell idle AI computing power, Micron fell by 10.57%, SanDisk fell by 10.62%, and the Philadelphia Semiconductor Index fell by 6.27%. The Technology Innovation 50 Index intraday fell by over 5%. Funds are starting to focus on more realistic operational indicators, such as whether orders are continuously growing, and if revenue can profits, rather than purely chasing the AI concept. A previous report from Goldman Sachs revealed that hedge funds were selling tech stocks at a record scale ahead of the Russell index rebalancing, with the total exposure and net exposure of the "Big Seven Tech" falling to the lowest levels of the year. Bank of America Securities believes that the biggest risk in the market is a sustained and significant rise in oil prices, which would compel the Fed to maintain a hawkish stance, pushing real interest rates further upwards and potentially triggering a deeper correction in AI stocks that are overvalued. Outlook: The fragility of the US dollar's "AI premium" Since the beginning of the year, the US dollar index has risen by approximately 1.4%. This increase has been largely driven by the influx of foreign capital driven by AI and expectations of Fed rate hikes. However, Slok's warning reveals a structural risk in the US dollar's trend that has been widely overlooked: once AI trading fades, the US dollar may lose its core support. As warned by Deutsche Bank analysts, the US dollar is losing its stability provided by the counter-cyclical demand for US Treasury bonds. With geopolitical rifts weakening international investors' long-term willingness to hold US debt, the US Treasury may have to pay higher term premiums in the future to complete bond auctions. For investors, this means that the trajectory of the US dollar is becoming more and more like a "technology stock" - closely tied to the rise and fall of the AI concept. When the AI hype recedes, the "AI premium" of the US dollar may also evaporate. The "double-edged sword" of rate hikes Traders still expect the Fed to possibly raise rates as early as September. According to Kalshi data as of July 13, traders estimate a 54% probability of the Fed raising rates by the end of this year. Federal funds futures are pricing in a over 60% probability of a rate hike in September. Rate hike expectations themselves support the US dollar - because other major economies have weaker resilience and lower willingness to tighten monetary policy. But rate hike expectations are also a "double-edged sword": if rate hike expectations are driven by rising oil prices and inflation concerns, they could trigger a sell-off in AI stocks, weakening the US dollar through the transmission chain of "AI foreign capital inflow US dollar". Bank of America Securities points out that if a sustained rise in oil prices compels the Fed to maintain a hawkish position, it will trigger a deeper sell-off in overvalued AI stocks. This is precisely the core scenario that Slok warned about: a chain reaction of AI sell-offs slowing foreign capital inflows weakening US dollar. Market divergence: Is the AI narrative reaching a tipping point? Despite Slok's warning drawing attention, there is still disagreement in the market about the extent of this risk. Optimists believe that the core drivers of the US dollar's strength are still the Fed's policy stance and the resilience of the US economy, rather than capital flows driven by AI. As long as US economic data does not significantly weaken, the US dollar is unlikely to decline. The net long position in the US dollar has reached $34.3 billion, the highest level since January 2025, but this may indicate that the US dollar is overbought rather than about to reverse. Pessimists, represented by Deutsche Bank, believe that the US dollar is losing its stability provided by the counter-cyclical demand for US Treasury bonds. Once AI profit expectations cool down and there is a deep adjustment in US stocks, foreign investors may simultaneously sell stocks and actively hedge their US dollar exposure, weakening the US dollar's safe haven support. US bonds may not automatically attract sufficient overseas buyers as they have in the past. CMoney analysis points out that global capital is focusing on a single theme, concentrating bets on a small number of US AI companies. This highly concentrated capital structure means that once the AI hype cools down, funds may quickly shift, triggering synchronous oscillations in the forex and stock markets. Slok's warning reveals a structural fragility in the current US dollar trend that is easily overlooked: record foreign capital inflows to AI stocks, combined with these investors generally abandoning exchange rate hedges, make the US dollar dependent on "implicit reliance on AI trading". With AI concept stocks having retraced significantly from their highs and market questioning the ROI of AI capital expenditures, this structural risk is worth close attention. The US CPI data and Fed Chair Powell's testimony to Congress to be released this week will be crucial in testing this risk. If inflation exceeds expectations and triggers a rise in rate hike expectations, it could have complex effects on AI stocks and the US dollar - this is precisely the "most dangerous scenario" that Slok warns about.