When the capital expenditures of AI giants transform from "money printers" into "money shredders," a dramatic shift concerning liquidity and asset pricing may be brewing. With nine months remaining until the US midterm elections, Michael Hartnett, a leading analyst at Bank of America, elevated his market warnings to a new level in his latest "The Flow Show" report published on February 15. Compared to his alert three days prior, Hartnett's current perspective is more incisive and specific. He points out that as expectations for AI capital expenditure (Capex) are revised further upward, the "AI disruption trade" is spreading at a startling pace from the technology sector into traditional service industries. For investors, the true signal for a market shift may hinge on a single action: when the tech giants "capitulate" and begin cutting their expenditures.

**$740 Billion: From "Printing" to "Shredding" Money** The most impactful new information in Hartnett's report concerns the repricing of AI capital expenditures. While the market was digesting a $670 billion expenditure forecast just three days ago, Hartnett now indicates that the projected capital expenditure for hyperscale cloud providers in 2026 has surged to $740 billion. This astronomical figure is not only staggering but also perilous. This frenetic level of investment will have extreme financial consequences: "This could push the free cash flow of the Mag 7 toward zero, or even into negative territory." To sustain capital expenditures of this magnitude, tech giants may be forced to embark on a massive bond issuance spree. This implies that technology growth stocks, once possessing pristine balance sheets, are undergoing a comprehensive "creditization." Hartnett believes the market narrative is rapidly shifting from "AI-awe" to "AI-poor." Against this backdrop, Hartnett provides a very clear trading signal: "The most obvious catalyst to reverse this situation would be an announcement by an AI hyperscaler of a capital expenditure cut." Should this occur, it would directly trigger a violent rotation out of technology giants and into "Main Street" assets.

**"Wildfire" Spread of AI Disruption Effects** It is a mistake to assume AI's impact is confined within technology stocks. The latest research indicates the disruptive effect is rapidly spreading into traditional service sectors. Hartnett terms this phenomenon "wildfire AI disruption." He chronologically listed the spread of the AI impact: "(Last) Monday was insurance brokers, Tuesday was wealth advisors, Wednesday was real estate services, Thursday was logistics..." Notably, the first sector disrupted by AI – Indian tech stocks (e.g., Infosys, TCS) in Q1 2025 – has yet to see any significant buying support. This suggests that once a stock is labeled an "AI victim" by the market, its price recovery could be a distant prospect.

**Political Countdown: February 24th** Hartnett reiterates that political factors are exacerbating this rotation. He notes that while Donald Trump's approval rating on Wall Street has hit a record high, his approval on Main Street (among the general populace) has hit a new low (with dissatisfaction over inflation as high as 36.4%). Hartnett explicitly states that the State of the Union address on February 24th will be a critical juncture. "If the anticipated 'Trump bump' fails to materialize by then, expect the administration to pivot towards more aggressive 'affordability' policies to win the midterm elections." To appease voters anxious about AI, these policies could involve suppressing energy, healthcare, and housing costs, potentially financed through mechanisms like yield curve control (YCC) or direct payments (universal basic income). This would further benefit small-cap stocks rather than the lofty tech giants. Based on positioning and price performance, Hartnett believes a strategy of "long Main Street, short Wall Street" is proving effective. Since the rate cut on October 29th, asset performance has been highly divergent: Winners (Inflation/Main Street Assets): Silver (+56%), South Korea's KOSPI (+34%), Brazil's Bovespa (+30%), Energy (+20%). Losers (Bubble/Wall Street Assets): Mag 7 (-8%), Cryptocurrencies (-41%), Software Sector (-30%).

**Japanese Yen: From Safe-Haven to "Long Bull" Signal** Regarding asset prices, Hartnett has identified a historically significant signal change. The correlation between the Japanese Yen and the Nikkei stock index has turned positive for the first time since 2005. Simply put, when the Yen strengthens, Japanese stocks also rise. Hartnett argues: "Nothing says 'long bull market' more than a currency appreciating alongside its equity market." This phenomenon has previously occurred in Japan (1982-1990), Germany (1985-1995), and China (2000-2008). Although this is a long-term positive for Japanese equities, the strong Yen is intensifying the pain of unwinding positions in cryptocurrencies, silver, and software stocks in the short term. Hartnett specifically warns about a currency "red line": "Japan will not tolerate a disorderly surge in the Yen (i.e., the USD/JPY rate falling below 145). This would crush Japanese exporters, hit global liquidity, and has historically coincided with global deleveraging."

**Fund Flows: Sell Signal Still Flashing** Despite $46.3 billion flowing into global equities this week, suggesting continued buying interest: Equity inflows: $46.3 billion Bond inflows: $25.4 billion Cash inflows: $14.5 billion Gold inflows: $3.4 billion ("no panic selling") Cryptocurrency inflows: $100 million, with selling pressure also abating after Bitcoin's ~50% decline from its October highs last year, accompanied by leverage unwinding. However, BofA's Bull & Bear Indicator reading dipped slightly from 9.6 to 9.4, remaining firmly in "Sell" territory. Concurrently, Hartnett reminds us that the "sell signal" for risk assets (active since December 17th) remains in effect. Hartnett maintains that the adjustment in risk assets is not over. He suggests the signal initiated in December will only truly be lifted when panic-driven cash accumulation begins, technology stock positioning decreases, and the indicator falls back to around 8. Key markers for this would include: Cash allocations rising significantly from the historical low of 3.2% to 3.8% or higher; Bond allocations recovering from a net 35% underweight to a net 25% underweight or less; Technology stock allocation decreasing from a net 17% overweight to neutral; Consumer staples allocation recovering from a net 30% underweight to a net 10% underweight or less.

**Reviewing 50 Years of "Great Rotations": Major Events Ignite Shifts, Leadership Changes, Next Phase Points to Emerging Markets and Small Caps** Hartnett uses the framework of "great rotations of the past 50 years" to explain the current environment: major political, geopolitical, and financial events often change asset leadership – 1971 End of Bretton Woods: New leaders were gold and real assets (up 417% from 1971-1980), laggards were bonds and financial assets (up only 67%). 1980 Reagan/Thatcher/Volcker Shock: Inflation peaked (14.8% in March 1980), the 10-year Treasury yield fell from 16% to 6% by 1985, making bonds the leader. 1989 Fall of Berlin Wall, Globalization & Deflation: The US stock market's relative position in global assets hit a 75-year low; commodities lagged in the 90s, with copper being the only asset with negative annualized returns that decade. 2001 9/11 & China's WTO Accession: Laggards were the US Dollar and tech stocks; leadership rotated to emerging markets/commodities and resource/financial sectors. 2009 Post-Financial Crisis QE and Buybacks: US stocks became leaders again, with private equity and growth stocks rising (the share of tech/telecom/healthcare in the ACWI index rose from 24% in 2008 to 44% in 2020; finance/energy/materials fell from 44% to 20%). 2020 Pandemic and Monetary/Fiscal Expansion: US government spending increased +56%, nominal GDP grew >50%; leaders included the "Magnificent Seven," etc.; laggards were bonds (the US 30-year Treasury fell 50% from 2020-2023).

Looking ahead, Hartnett judges the next structural leaders will be: emerging markets and small-cap stocks. His supporting rationale includes: From US Large-Cap Growth to Small-Cap Value: A shift from services to manufacturing, from asset-light to asset-heavy models, and rising costs from the AI arms race; he notes that hyperscalers have issued $170 billion in debt over the past 5 months, significantly higher than the ~$30 billion annual pace from 2020-2024, and widening spreads are already creating pressure. From US to Emerging Markets: He describes a "New World Order = New World Bull Market" for global rebalancing, mentioning a new "Anything But Dollars (ABD)" trade; he also emphasizes that asset allocations to China and India remain low, despite both being among the world's top four economies, and points out that Chinese bank stocks have quietly reached 8-year highs.