Market sentiment has shifted once again, with AI application and consumption sectors now outperforming AI infrastructure hardware names. This particular ETF could serve as a key barometer for the current bull market's trajectory.

It's not all negative: with yields on the rise, bonds are poised to act as an effective buffer against sudden downside surprises in equities. Let's revisit a chapter from the aftermath of the September 11 terrorist attacks, when stocks endured a brutal first week of trading.

Wall Street has returned from its summer recess, and so far, the market's steadfast bet on equities is being validated. The S&P 500 is barely holding its upward channel; last week saw a brief, modest pullback that approached but did not break through the upper boundary of the May-July trading range. On several occasions, the index had opportunities for a deeper correction, but these never materialized, with adjustments remaining hidden beneath the surface of the tape. The resilience of the broad market index reflects institutional investors' full-throttle commitment to equities. Even amid widespread warnings about typically weak seasonal performance following the summer months, institutions remain unmoved. Or perhaps they simply harbor the expectation that even if September and October bring turbulence, the fourth quarter tends to repair the damage.

Data from Goldman Sachs, State Street, the National Association of Active Investment Managers, and Bank of America all indicate that asset allocators have completed their positioning for the autumn rally, as measured by equity exposure and risk appetite metrics. The courage/fear ratio tracked by Rutherold Group has climbed to roughly an 18-year high. This indicator compares two portfolios: the "courage" basket (small caps, emerging markets, commodities, S&P 500 cyclical stocks) against the "fear" basket (dollar, gold, S&P low-volatility stocks, 10-year Treasuries).

Barclays equity strategist Venu Krishna noted last week that retail investors' aggressive sentiment has cooled: "The recent weeks have seen a pullback in retail participation, meaning this wave of FOMO buying is being driven primarily by institutional investors, rather than individual traders." Institutional money chasing earnings growth is the strongest driving logic of this bull market. But the risks are equally real: AI capital expenditure could be overextending future profits, and large enterprises face the hidden danger of "inflated earnings." The CBOE Volatility Index (VIX) sits below 15—a level that's difficult to sustain long-term but is reasonable for now, and it's prompting many quantitative models to maintain elevated risk exposure.

Additionally, the two core market concerns haven't actually materialized. Last week's string of positive economic data, including a strong non-farm payrolls report, has alleviated fears that macroeconomic weakness would lead the Federal Reserve to make a policy mistake. There have been rational concerns about the sustainability and pace of the AI investment supercycle, but the earnings season has delivered no substantive negative signals. AI "demand-side" companies have raised capital expenditure guidance, and AI "suppliers" have lifted earnings forecasts—as seen in last week's reports from Dell and Broadcom.

Rising yields aren't all bad? Of course, the resilience of the economy and the continued willingness to spend on AI are two factors intensifying investors' biggest worry: rising bond yields. The 10-year Treasury yield's steady march toward 4.8% is better viewed as a normalization shock—rates returning to pre-global-financial-crisis levels, and bonds and stocks reverting to their pre-2000 relationship, where rising yields correlate negatively with stock prices. As I mentioned last week, the absolute level of yields itself doesn't hinder a strong stock market. But this journey of the 10-year Treasury from under 1% to 4.7% feels entirely different from seeing 4.7% on the way down from 8% 25 years ago. Today's 4.7% yield means most bonds issued in recent years trade below their par value, making investors wary of fixed income. Yet paradoxically, it's precisely now that bonds, thanks to their coupon income, are regaining their value as a hedging buffer.

Jim Reid, head of global macro and thematic research at Deutsche Bank, says: "Over the medium term, it's increasingly difficult for government bonds to deliver outright negative returns. Despite the persistent bearish headlines, bonds are finally behaving like bonds again." The last time yields and stocks moved largely in opposite directions over the short term was the 1990s. This doesn't mean the 60/40 stock-bond balanced portfolio has lost its value. The Vanguard Balanced Index Fund (60/40 allocation) delivered a 14.8% annualized total return from March 1990 to March 2000 (the tech bubble peak), capturing more than 69% of the S&P 500's annualized return over the same period. Over the past decade, bonds were viewed as a hedge against stock declines, but the opportunity cost of balanced funds rose significantly, delivering just 61% of the S&P 500's return, primarily because starting bond yields were far lower than today.

From a trading perspective, bonds may continue to weaken. With ten days until the Fed's next policy meeting, market pricing for a rate hike versus a hold is nearly split; this week's inflation data is seen as the key swing factor. Oil prices are again experiencing sharp volatility; Japan may sell Treasuries to defend the yen; and global fiscal imbalances add psychological pressure. But when bonds' coupon yields are sufficiently high, they can cushion against sudden market shocks if the situation turns more turbulent than institutions' fully-loaded positioning anticipates.

Market thermometer: This indicator synthesizes multiple data points while observing both investor rhetoric and actual trading behavior.

Market notes: The September 11, 2001 terrorist attacks devastated America's financial center, and while the market impact was far less severe than the humanitarian, geopolitical, and cultural trauma, it was still significant. After the longest NYSE closure since the Great Depression, the S&P 500 plummeted 11% during the week of September 17. Prior to that, the index had already fallen 28% over 18 months due to the tech bubble's collapse. Following that brutal first week, I published a cover story titled "Now Is the Time to Buy Stocks." That call was well-timed, partly due to luck and partly to contrarian thinking. Over the subsequent ten weeks, the S&P 500 rallied 24%, marking one of the strongest bear-market rebounds in history. In hindsight, that bounce somewhat prolonged the 2000-2003 bear market, delaying the market's full cleansing: risks such as tech overinvestment, declining corporate profitability, and accounting scandals didn't fully erupt until late 2002; the final market bottom was 18% lower than the post-9/11 trough.

Regarding the endless stream of ETF products and portfolio construction, Morningstar's Jeffrey Ptak offers many pragmatic insights. His views tend to be contrarian, maintaining skepticism toward overly complex fund structures and flashy return promises. He cites the Defiance 2x Long OKLO Daily Target Leveraged ETF, where investors' capital has experienced an average annualized loss of 98.5%. He also flags a newly filed ETF with the SEC that aims to capture pre-IPO startup valuations through layers of derivatives: "The fund advisor plans to allocate approximately 80% of the portfolio to the pre-IPO space, primarily through swap agreements linked to perpetual futures."

Closing commentary: As mentioned above, at least at the corporate level, the AI capital expenditure boom shows no signs of abating. But the way the AI trade is unfolding is evolving in unpredictable directions. Since June 30, Nvidia has outperformed the semiconductor complex by 35 percentage points; in the 11 months prior, Nvidia consistently lagged. The software sector has clawed back 70% of the losses from last October's to April's "SaaS crash." Following Nvidia's acquisition of AI model development and distribution platform Hugging Face, and Meta's latest open-source AI product progress, market winds have shifted in the short term: AI consumer application sectors are taking the lead while AI hardware infrastructure names take a back seat. That said, memory chip stocks are showing early signs of breaking free from a months-long downtrend; some AI-related industrial stocks relying on years of backlogged orders may have already priced in their upside.

The major indices remain glued to historic highs, primarily thanks to a rebound in the mega-cap tech platform stocks. The iShares Nasdaq Top 30 ETF (ticker: QTOP) captures this shift well, offering a broader view than the narrow "Magnificent Seven." Its holdings include not just the seven tech giants but also core semiconductor names and other non-tech leaders that are heavy AI users. QTOP's relative strength peak against the broader market occurred in May-June, around the SpaceX IPO anticipation; momentum then collapsed, the index pulled back significantly, and it partially recovered during earnings season. Yet its relative performance against the broader market remains nearly 5% below the three-month-ago high. If this ETF fails to notch new relative highs in the near term, it could signal that the engine driving this AI-fueled bull market is beginning to lose steam.