The AI bubble is facing a "trial"! Bank of America provides a "trade guide" for the midterm elections: warning that U.S. stocks may encounter a 10% correction, with three major scenarios outlining investment logic.

Global bond yields have soared to a 20-year high, becoming the biggest threat to the AI capital expenditure boom, while the upcoming U.S. midterm elections may serve as a critical variable that ignites market volatility. Michael Hartnett, Chief Investment Strategist at Bank of America Securities, delivered a stern warning in the latest edition of the "Flow Show" weekly report: if the Democrats sweep both the House and Senate in the midterm elections, U.S. stocks could fall more than 10%, the dollar could weaken, and bond yields could decline, posing a risk of a bubble burst in AI. The Underestimated Tail Risk: The Probability of a Democratic Sweep Rises to 50% Hartnett categorizes a "Democratic sweep" as one of the biggest tail risks in the current market, which investors are largely not pricing in. This optimism is partly due to two major misjudgments: first, the belief that the Trump administration primarily governed via executive orders, making changes in Congress less impactful; second, that the Senate electoral map still favors the Republican Party. Data from Polymarket indicates that the probability of the Democrats sweeping both chambers has risen to 50%, significantly higher than the 10% chance for the Republicans. This sharp increase in probability is closely tied to President Trump's approval ratings, which currently hover between 35% and 40%, well below the historical average of 53% two months before midterm elections. According to Bank of America's August fund manager survey, 47% of respondents expect a "Republican-controlled Senate and Democratic-controlled House," 23% anticipate a Democratic sweep, while only 9% foresee the Republicans retaining control of both chambers. Hartnett admits that the midterm elections will not be a "regime change" election like those in 1980 with Thatcher/Reagan or 2016 with Trump, as it will not fundamentally change the upward trajectory of U.S. government spending. However, the structural differences in election outcomes could have significant impacts on asset prices. Three Scenarios, Three Distinct Market Outcomes Hartnett outlines three potential scenarios for the midterm elections, each with drastically different implications for asset prices: Scenario One: Democratic Sweep (Probability about 50%) A Nuclear Reaction in Risk Aversion If the Democrats simultaneously take both chambers, it will trigger the most severe risk aversion response. Hartnett warns that U.S. stocks could face declines of over 10%, the dollar could weaken, and bond yields could fall. More critically, the AI bubble would be at risk of bursting, with policies potentially shifting from "populist capitalism" to "populist socialism," reversing the trends of low taxes and deregulation, thereby directly challenging the core logic of AI capital expenditures. International stock markets might perform better, with European markets outpacing Asian markets. Bank of America's preferred hedge strategy in this scenario is to short the financial sector and the dollar. Scenario Two: Republican Sweep (Probability about 10%) A Green Light for Risk Appetite If the Republicans unexpectedly retain both chambers, it would serve as a "green light" for risk appetite, clearing policy hurdles for AI transactions. Hartnett believes this could drive the AI sector to continue strengthening, potentially evolving into a "bubble-like" market by 2027, supported by a return to the theme of American exceptionalism, which would bolster the dollar. Scenario Three: Divided Congress (The Most Likely Probability) Gridlock = Goldilocks This is the most likely outcome at present the Republican Party controlling the Senate and the Democratic Party controlling the House. Hartnett categorizes it as a risk-moderate scenario referred to as "gridlock = goldilocks." In this scenario, both parties counterbalance each other, leading neither to radical policy shifts nor to a complete stifling of AI capital expenditures, which is the "just right" result for the market. Texas Governor's Election: A Barometer for AI Policy In August, Hartnett particularly emphasized that the Texas gubernatorial election is the most critical variable in this midterm election. He defines this election as a referendum on "the cost of living and AI data centers." Texas currently has 335 data centers, with an additional 247 in the planning stage. The current Republican Governor Greg Abbott recently announced a temporary halt to data center expansion, reflecting concerns that rising energy costs and grid pressures are translating into electoral unease. If Democratic challenger Gina Hinojosa wins and the Democrats secure the Senate, AI investments and risk assets might face repricing. Bank of America points out that if the Republicans were to lose in Texas, it would send a signal that voters are prioritizing the affordability of living costs over tax cuts, directly undermining the policy foundation of the current AI bull market. The "Early Warning" from the Bond Market: Yields are the True Anchor for AI Trading Hartnett asserts that the most critical market signal in recent times is not the better-than-expected employment data but the complete breakdown of the global bond market. The yield on 10-year U.S. Treasuries has risen to 4.81%, nearing levels seen during the 2008 financial crisis; the yield on 30-year U.S. Treasuries has reached 5.31%, the highest since 2007. Meanwhile, the yield on Japan's 10-year government bonds has breached 3.0% for the first time since 1996; Germany's 10-year bond yield has risen to 3.38%, a new high since 2011. The Bloomberg global bond yield index has soared to its highest level since 2007. Hartnett distills this phenomenon into a core judgment: "bonds dominate the bubble." He believes that long-end yields rather than the stock narrative are the true anchor for current AI trading. Until global 30-year yields fall below 5%, those engaged in AI infrastructure construction and investment will continue to lag behind the users of AI technology.