Subscribe first, and you will receive Market Memo from Mike Santoli in your inbox as soon as possible.

The difference between restorative recuperation and unhealthy stagnation lies in time. It is not always easy to distinguish between resilience and fatigue.

The S&P 500 index has been trading sideways for several months and last week fell back to the level it first reached on June 1st. Over the past three months, the index has almost consistently remained within the range covered by that strong two-day rebound from August 3rd to 4th.

Meanwhile, the AI trading sector, which had previously led the gains – represented by iShares and AI Innovation ETF – once fell by 35 percentage points compared to the equal-weight S&P 500 index. Subsequently, it managed to recover half of that loss, but then gave back most of its rebound.

Putting aside the fact that this process has left everyone with plenty of time and room for overinterpretation, this continuous oscillation and impact on positions have also brought some positive effects from a tactical perspective.

It created short-term consolidation conditions in semiconductors and other once-momentum-driven AI hardware concept stocks. At the same time, cyclical sectors in the market, ranging from discretionary consumer goods to industrial stocks to bank stocks, also experienced double-digit pullbacks due to higher bond yields and persistently high oil prices.

Stock investors clearly prefer quality and defensiveness, while the bond market has begun to allocate growth capital more strictly, raising the financing thresholds for those marginal AI data center borrowers.

In this process, some familiar cross-market relationships have also been disrupted: the yield on U.S. Treasury bonds is no longer highly correlated with oil prices. Although crude oil prices remain above $100 per barrel, the yields have clearly fallen from their high levels. The original negative correlation between the S&P 500 index and U.S. Treasury bond yields has also weakened, as buying activity in the bond market seems to be more related to the setbacks in the construction process of AI, which has also dragged down large technology companies.

Goldman Sachs' trading desk noted over the weekend that there was "trader fatigue" in the market, partly due to the momentum factor stock basket experiencing at least 30 single-day price movements of 5% or more in 2026, a frequency higher than in any other full year, including years of crisis such as 2000, 2009, and 2020. Goldman Sachs stated, "We are not sure how many more iterations of these thematic narratives we can withstand."

Of course, confusion and pressure may ultimately also reset positions and attitudes, leading to positive outcomes. The comprehensive indicator of sentiment and positions tracked by John Kolovos of Macro Risk Advisors is approaching an "excellent entry point" suitable for a tactical rebound, he stated. By most standards, hedge funds have significantly reduced their leverage and exposure to tech stocks. Deutsche Bank noted that the positions held by institutional investors in the tech sector are not much higher than they were during the Iran war, and "the last two tech stock rebounds have both coincided with the release of strong financial reports."

Against this backdrop, cloud platforms with extremely large market values such as Alphabet, Amazon, and Microsoft will announce their financial results this week. Meanwhile, the IPO filing of Anthropic shows that its operational indicators are good, and it already has $100 billion in cash before the funds raised through IPO are credited to its account, which may alleviate some of the financing anxieties of AI. In addition, Blackstone providing financing for the capital expenditures of SpaceX and Broadcom arranging $63 billion to fund the chips for Anthropic also help to ease these concerns.

AI The transaction needs some breathing room; it's not yet the moment of a "either/or" choice, but we're not far from that point.

If there is one disturbing exception among the evidence of investors withdrawing and a decrease in risk appetite, it is that retail investors are still continuously buying into those tech stocks whose charts look the worst.

Citadel Securities Head of Equity and Equity Derivatives Strategy Scott Rubner wrote on Friday, following a 3% drop in tech stocks on Thursday: "Although several sell-side institutions commented that retail investors had given up, our platform data shows the opposite." In other words, Citadel as a major market maker, the flow of retail orders it processes indicates that retail investors are aggressively buying at bargain prices in leveraged semiconductors ETF, Micron, Meta, and other stocks that suffered heavy losses. This is consistent with the data from Robinhood; according to a report by Sherwood Media under that company, Robinhood traders flocked to buy on Friday when Bloom Energy fell 11%, setting a new record for the second-highest single-day net buying volume in the stock's history.

This aggressive buying behavior during significant weakening periods is indeed impressive to some extent and even garners admiration, but it usually does not coincide with important market lows. Data from Citadel shows that the market bottoms at the end of March and July occurred precisely during periods when retail investors were panicking and selling net for one or two weeks in a row.

This is not to say that retail investors' stubborn buying at lower levels will not lead to some rebound in the severely hit sectors, but it is likely to limit the intensity and sustainability of that rebound.

Market Thermometer

The latest reading from John Kolovos of Macro Risk Advisors states: 'Emotions and breadth have finally provided us with some aspects that are worth liking; perhaps an excellent entry point is approaching, but the macro environment must cooperate as well. Our research indicates that when financial conditions remain tight, improvements in momentum often lead to disappointing returns in history, and expose the market to serious vulnerabilities to left-tail risks or pullbacks.'

Street Observation

Sometimes, to assess the weight of evidence, one has to wait a long time to see all of it. I admit that the anxiety-driven fluctuations and valuation compression of the past few months have created conditions for an unexpected upward movement and a round of relief in prices.

But from a broader perspective, I still want to remain vigilant about the changing atmospheric environment.

As I pointed out in the weeks following the launch of this newsletter, I believe we are experiencing the tremors of the latter phase of a cycle: the market leadership is narrowing, investors are unwilling to pay higher valuations for rapid profit growth, the yield curve is flattening, and the credit market has become more stingy.

I have suggested that the market may be telling us that companies are achieving “excess profits” through a cycle of upfront capital expenditures. The Iran war and its macroeconomic disruptions are likely to have prevented the AI theme from evolving into a full-blown bubble, confining the frenzy only to the semiconductor and related sectors. However, this does not mean that this cycle will necessarily inflate into a broader bubble.

I also believe that even though stocks have performed well, bonds can now provide a useful income buffer to hedge against the risks associated with concentrated stock holdings.

At this time last year, the outperformance of cyclical sectors, favorable fiscal conditions, and a moderate credit environment all indicated that there was no need to worry about economic growth in 2026. So, now, with defensive stocks leading the gains, weakened fiscal momentum, and unstable corporate credit, what are they whispering in investors' ears?

I also believe in the law of mean reversion, even though it only manifests intermittently. In an article on his blog “A Wealth of Common Sense”, Ritholtz Wealth Management’s Ben Carlson discussed in detail the outstanding returns of the S&P 500 over the past 15 years, as well as what this may imply for the future.

The article with the Dario Perkins tag, TS Lombard, has been widely shared in the past few days for good reasons. It is a clear and thought-provoking experiment that raises the alarm about how the prosperity of AI may currently be breeding its own severe collapse.

Closing Market Observation

The process of funds shifting towards pure, traditional defensive stocks is somewhat too mechanical, making it difficult to determine that this is purely a "signal" at the macro level, as the rotation of such factors seems almost scripted, which in turn is less credible.

However, it is indeed the case that the valuations of consumer staples and technology stocks are converging. This may seem somewhat unusual to today's generation of investors, as there was a time when consumer staples enjoyed a valuation premium due to their stability and enduring brands, while technology stocks were seen as being more cyclical and volatile.

Procter & Gamble has long been a benchmark in the consumer goods industry, and its symbol of management quality is just like Pampers is to diapers. To some extent, Google has also almost become a synonym for "searching online."

Whether you see this cross of price-earnings ratios as a sign that world order is returning to normal, or as a strong buying signal for the Alphabet stock price, depends on what kind of investor you are.

Subscribe to Market Memo of Mike Santoli here.