Global markets are entering the most pivotal week of the summer, with two things now abundantly clear: trading is no longer confined to the artificial intelligence computing theme. A significant rotation has emerged, led by gold and precious metals, a healthcare sector revitalized by mRNA technology entering a new era in cancer treatment, and traditional defensive value plays such as energy, telecom, and banking stocks that have long lagged tech. This broadening trend has broken the previous pattern of AI-exclusive dominance, creating a healthier and more diversified bull market. Meanwhile, the recent violent swings in long-dated US Treasuries have proven that catalysts can fuel downside moves just as easily as they drive asset prices higher.

This is the backdrop entering late August. On Wednesday, NVIDIA will report earnings—the closest thing to a quarterly referendum on the AI trade. Then on Friday, Federal Reserve Chair Kevin Warsh will deliver his inaugural Jackson Hole keynote speech, all while markets debate the Treasury's efforts to curb selling in long-dated government debt. The publicly announced topic of Warsh's speech is payment innovation, which sounds unlikely to move markets significantly. Yet he could touch on other matters, and there is always the risk of markets overinterpreting a single comment.

The latest Bank of America Global Fund Manager Survey paints a picture not of caution or pessimism, but of a striking contradiction: institutional risk perception has risen sharply while actual positioning remains extremely adventurous and AI allocations stay crowded. A record 56% of respondents bet on a US "no-landing" scenario over the next twelve months, a net 37% expect double-digit earnings growth, and 72% believe the Fed won't hike before the midterm elections. Correspondingly, global equity net overweight has surged to 56%, the highest since November 2021, cash levels have fallen to 3.5%—the sixth lowest historically—and bond net underweight has expanded to 39%.

During the survey period, another BofA research report highlighted attractive opportunities in value stocks, biotech, regional banks, select credit, and commodities, amid the deleveraging storm in AI themes and forced unwinding of crowded positions, as well as inflation risks challenging traditional portfolios. The report also noted that international small-cap value stocks have become more attractive relative to US large-cap growth, and that Japanese corporate profitability has climbed to record levels. The bank also views listed private equity managers as a contrarian opportunity and prefers high-quality high-yield bonds over investment-grade debt.

Wall Street institutions, including Bank of America, are not turning bearish on the AI theme. Rather, they are advocating a clear asset allocation upgrade: shifting from highly concentrated AI/US large-cap growth trades toward a portfolio that retains structural AI longs while adding low-correlation, high-cash-flow, low-valuation assets. As AI transitions from a scarcity narrative to a trillion-dollar capex realization phase, the determinant of excess returns shifts from "whether you have AI exposure" to "whether valuation, free cash flow, ROIC, and crowding are aligned."

This week is far more than just an NVIDIA earnings week; it is a dual stress test of "corporate earnings anchor plus global discount rate anchor." NVIDIA will answer whether token computing demand can translate into sustained inference revenue, profit, and AI capital returns. Warsh, meanwhile, could reset the valuation benchmark for all risk assets by influencing policy rate expectations, term premium, and dollar liquidity.

The correlation between AI and non-AI sectors has fallen to roughly negative 0.6, indicating this is more a capital redistribution than an outright exodus. However, with hedging insufficient and the first batch of systematic fast-money quant selling sitting just 4% below current market prices, a "strong NVIDIA earnings report plus stable long-duration bonds" could turn the rotation into a healthy broadening bull market. Conversely, an unfavorable outcome could escalate a localized correction into systemic de-risking.

The Token Throne Faces Its Quarterly Referendum: The AI trade is not entering a bear market trajectory but rather facing a judgment on AI investment returns. Before this pivotal week, it's worth assessing where markets stand. Despite some tech rebounds, the rotation out of AI continues. Over the past five trading days, the best performers included "quality stocks" with historically strong cash flows, a basket tracking so-called "AI victims," and value assets sensitive to inflation data or with relative advantages in a stagflationary environment. Semiconductor stocks and AI data center infrastructure concepts have fallen sharply.

Last week, inflation and rate concerns drove the rotation—quality fundamental assets, stagflation beneficiaries, and commodities outperformed while AI lagged. One chart encapsulates the broader picture: Goldman Sachs' broad AI stock basket's 40-day correlation with the S&P 500 excluding AI components has turned sharply negative for the first time, now around negative 0.6. This suggests capital exiting AI hasn't left the equity market entirely but is fueling net inflows across other sectors.

However, this doesn't mean the AI narrative has broken. Infrastructure-type tech companies at the core of AI computing still contribute roughly half of S&P 500 earnings growth. Meanwhile, median S&P 500 companies saw profits grow 14% last quarter, indicating a broadening earnings base under a soft-landing trajectory. Reports that Anthropic's annualized revenue run rate has exceeded $65 billion, with OpenAI around $40 billion, offer some comfort to investors questioning AI end-demand. But tougher questions remain: who will be the ultimate winners, whether capex can sustain strong returns, and what AI computing resources are truly worth.

Currently, a widely-watched token cost metric—the Silicon Data AI LLM spending index—has fallen nearly 50% from its May peak. It has become a hot topic on trading desks and social media discussions about reducing inference costs. Notably, Ornn market data does not show significant declines in large AI labs' token prices in August. Regardless of the actual situation, the index's prominence itself reflects prevailing market anxiety. Meanwhile, CDS spreads and credit spreads for hyperscale cloud vendors continue to climb, heightening concerns about AI investment returns and whether massive capex could trigger credit defaults.

The AI narrative still faces many unresolved questions—usage, scaling laws, pricing, actual profits, and industry leadership. AI infrastructure still contributes about half of S&P 500 earnings growth, and Anthropic and OpenAI's annualized revenue run rates of over $65 billion and $40 billion respectively prove end-demand is real. But the token cost index nearly halving from its May peak and widening hyperscaler credit spreads mean the valuation focus has shifted from "who owns the most GPUs" to "who can convert computing power into cash flow."

What NVIDIA's earnings and outlook truly need to validate is not just demand for the next-generation Vera-Rubin AI cluster, but whether inference computing power, revenue growth, pricing power, customer capital returns, and the AI financing structure of credit and securitized debt can collectively support the next major leg of the AI trade's rally.

Warsh Holds the Global Discount Rate Switch: 5.30% Long Bonds and $90 Oil Squeeze Tech Stocks. Next is the slower-burning but more consequential concern of rates and Treasury yields. Last week, the 30-year Treasury yield touched 5.30%. The Treasury then announced it would at least double its long-duration bond buyback program, which briefly pulled yields lower, but by Friday's close, the 30-year yield had nearly recovered all of its post-intervention decline. The MOVE index, measuring rate volatility, is near 70—well below the March peak triggered by the US-Iran war but far from winter's calm. The current yield-driven tension is slowly building, not yet boiling over, and the new Fed chair will speak against this backdrop.

Geopolitical risk runs through all these variables. Brent crude has held above $90 per barrel, up over 50% this year. This supports energy stocks while reinforcing inflation anxiety and term premium concerns, prolonging bond market tension on long-end yields. Options skew is rising and credit spreads remain elevated—some market indicators show investors remain cautious about risk-taking.

August positioning tells its own story. In the mid-month week, hedge funds bought US stocks every trading day, following three consecutive weeks of buying that was the most aggressive since March 2020. However, momentum is fading. Goldman Sachs prime brokerage data shows global stocks experienced the fastest net selling in two months last week. S&P 500 call option volumes have normalized from July's frenzy; current buying resembles prudent re-entry rather than the breathless one-way melt-up chasing of the past two years.

The 4 million call options fueling the rally chase quickly normalized—latest daily volumes indicate no appetite for risk-taking at any cost. Markets enter this week with solid underlying support, but an unfavorable combination of outcomes could suddenly open a "trapdoor" selling trigger. Approximately 96% of S&P 500 constituents are in their buyback window, with over $1 trillion in repurchase authorizations providing steady demand. Trend-following funds (CTAs) hold roughly $140 billion in global equities, with baseline demand turning modestly positive.

However, the risk structure is not perfectly symmetrical. If markets suffer a serious drawdown within a month, these systematic strategies could trigger short-term violent selling exceeding $150 billion globally, with the first batch of sell triggers sitting just 4% below current market benchmark levels. This week's major events will determine which side of the risk structure prevails.

Ultimately, fund flows are still providing support, options gamma is stabilizing, and the rotation out of pure AI trades—particularly toward value stocks and healthcare with low valuations and long-term underperformance versus tech—appears quite healthy. More defensive allocations suggest risk appetite is cooling somewhat; the steepening skew curve indicates investors are adding hedges against rate volatility and AI intrinsic value uncertainty. If markets remain in summer doldrums, this pattern can persist. However, a potential negative catalyst would land on a market where hedge fund or traditional asset manager hedging levels remain near yearly lows, and various buffers have yet to be truly tested.

Support remains in place until the trapdoor suddenly opens. This week may reveal whether it can bear the weight. Corporate buyback authorizations exceed $1 trillion, trend strategies hold about $140 billion in global equities, and options gamma is stabilizing, providing short-term buffers. But global stocks have turned to net selling, and July's call option frenzy has clearly cooled. If this week's two major events deliver a friendly combination, capital could return to the AI supply chain and healthily spread to strong cash-flow quality stocks, energy, commodities, and stagflation beneficiaries, driving a breadth-expanding bull market. If earnings and rates both disappoint, a 4% index decline could trigger systematic selling exceeding $150 billion—the market floor may look solid, but the real danger lies in the untested trapdoor beneath it.

Bank of America strategists are urging investors to step outside the crowded AI trade. They believe that amid the deleveraging storm in AI themes, increasingly demanding earnings expectations, forced unwinding of extreme positions, and inflation risks challenging traditional portfolios, value stocks, biotech, regional banks, select credit, and commodities all offer attractive opportunities. As the AI bull market enters a "high valuation, high crowding, high capital consumption" phase, BofA advocates shifting marginal capital from the most expensive AI beta toward "cheaper earnings growth, real cash flows, and inflation-resistant assets"—a rebalancing from a single tech theme to market-wide earnings and high-quality cash flow sectors, not an end to the AI bull market.

For example, billionaire and legendary hedge fund manager Bill Ackman's Pershing Square has pivoted to a "buy high-quality fundamental stocks at mispriced/crash prices" framework during the AI deleveraging and crowding unwinding—initiating positions in digital payment and card network giants Visa and Mastercard, along with four other companies with low concentration and historically strong cash flow quality.