Executive Summary:
The AI "Prove It" Market: Mega-cap tech sold off as unprecedented AI infrastructure spending – projected to top $720 billion this year – drove free cash flow negative across major hyperscalers. The market's focus has ruthlessly shifted from top-line cloud growth and TAM promises to near-term returns on invested capital.
Multiple Compression and Seasonal Tailwinds: The S&P 500's forward P/E has compressed to 18.7x as investors price in peak earnings growth. However, clearing the "peak event risk" of the mid-terms, Fed policy, and Treasury issuance sets a constructive stage for a seasonal year-end rally.
The Volatility Wedge & Defensive Rotation: While headline index moves appeared tame, internal market divergence reached historic levels. The equal-weighted S&P 500 beat the cap-weighted index by 6% in October (the widest spread since 2009) as capital rotated out of momentum trades and into defensive sectors, pushing dividend and low-volatility stocks into highly expensive valuation percentiles.
Full Commentary:
The AI theme – and virtually any investment tied to it – got downright scary heading into the Halloween weekend. Once again, the weekly results for the major averages (S&P 500 -0.57% and the Dow -0.85%) masked the carnage taking place in the largest and most critical segment of the stock market: Tech and AI. The Nasdaq Composite experienced a deeper slide of -2.0%, but it was the downside price action among the major AI players that was truly unnerving.
What started with Meta two weeks ago (so-so results overwhelmed by increased spending and capital-raising guidance) continued with Alphabet reporting on Thursday, resulting in its largest single-day drop in history (-13% at its lows). The trend carried into Friday, with Intel crashing (-30%) and Amazon following suit (down over -12% at the lows of its after-hours session) following a disappointing quarter. When the dust settled, the Nasdaq had briefly slid into correction territory during Friday’s weak trading session and is currently testing massive support at its 200-day moving average.
It was the ex-AI components of the equity market that kept the broader indices in check. The equal-weighted S&P 500 beat the cap-weighted index by 6% in October – the widest monthly spread since April 2009. A rotation into defensives and interest-rate-sensitive sectors proved to be a port in the storm during a volatile month for bonds, where yields initially spiked on Fed tightening and inflation concerns, only to drop as faltering growth expectations took hold. The 2-year Treasury yield started the month at 4.89%, peaked at 4.92%, and ultimately fell 30 basis points to 4.59%. The 10-year Treasury followed a similar arc, beginning the month at 5.29%, peaking at 5.31%, and declining 36 basis points in the back half of the month to close at 4.93%.
I’ve been attempting to describe the erratic price action for months, and I must admit I’m running out of ways to articulate the stark difference between heightened volatility under the surface and the eerie calm on the surface. Goldman Sachs recently published data that contextualizes this massive wedge, illustrating just how difficult it has been for investors to manage risk along the path.
Comparing the monthly returns of the S&P 500 against Goldman’s proprietary momentum basket, the divergence is staggering. For the better part of the last four months, the S&P 500 has traded in a tight, roughly 4% range. Meanwhile, the momentum basket – which encapsulates the bulk of the AI trade – has violently oscillated within a 60% range.
Why is this happening? It is never just one variable. Global central bank tightening, the largest infrastructure investment cycle in history (AI), a prolonged U.S./Iran conflict carrying deeper-than-expected impacts, the mid-term election cycle, hyperscaler capex crowding out liquidity, and peak earnings growth all play a role. But the primary driver right now boils down to earnings, financing needs, and liquidity.
The Shifting Mindset: From "Belief" to "Prove It"
Peeling back the onion on hyperscaler earnings over the past two weeks reveals a clear psychological shift: investors have transitioned from blindly hoping for AI returns to demanding tangible proof.
To understand the sheer scale of the expectations currently priced into AI, look no further than SpaceX’s recent S-1 filing, which cited an astonishing $28 trillion Total Addressable Market (TAM). To contextualize how nonsensical that figure is, the annual GDP of the entire U.S. economy is roughly $33 trillion. Or consider the data highlighted by Barclays strategist Alex Altman on the Odd Lots podcast: the combined, inflation-adjusted cumulative revenue for Apple and Microsoft from 1990 to today is $9.8 trillion. Two of the most successful businesses in modern history generated roughly one-third of SpaceX's claimed TAM over 35 years.
This is not a knock on SpaceX, but a necessary reality check for investors regarding the astronomical numbers being thrown around the AI space. There are limits to this growth, and those limits are beginning to surface in hyperscaler earnings.
The recent earnings reports highlighted a worrisome new milestone: tech’s biggest companies are now spending more cash on data centers than they generate from their core operations.
Alphabet (Google): Reported underwhelming ad revenue last Wednesday, raising fears that generative AI is cannibalizing its core search engine. More concerningly, this marked Alphabet's second consecutive quarter of negative free cash flow as investor anxiety mounts over its massive capex plans for 2027.
Amazon: Delivered great top-line numbers, including a massive 39% year-over-year surge in AWS sales to $46 billion. However, shares traded down sharply because AI-related capex topped $75 billion – more than double last year's spend. Despite generating $168.3 billion in operating cash over the past year, heavy investments in data centers drove Amazon’s free cash flow down to -$42.4 billion.
Meta & Microsoft: Meta shares fell 7% two weeks ago after revealing a cash burn so severe it required a $25 billion equity raise to fund infrastructure. Microsoft similarly announced it was tapping the bond market for $25 billion.
Across Amazon, Microsoft, Google, and Meta, AI infrastructure spending is on track to exceed $720 billion this year alone – roughly triple what they spent just two years ago. Crucially, all of them plan to spend even more next year.
Just as hyperscalers turned free cash flow negative and looked to capital markets to fund blowout capex commitments, a hawkish Fed drove market interest rates up by over 100 basis points. The resulting leverage has unsettled investors who are, for the first time, fretting that these mega-projects may not pay off in the near- or intermediate-term.
To be fair, the AI demand is real. Google Cloud revenue grew 103% year-over-year, with its backlog surging 256% to $561 billion on roughly $120 billion of Annual Recurring Revenue (ARR). Amazon’s AWS backlog grew 189% to $578 billion. These are startup-like growth rates yielding a combined $1.1 trillion of contracted cloud revenue between just two companies, proving that these massive capex investments are generating returns.
However, a $1.1 trillion backlog is still just top-line revenue. In the current environment, markets are ruthlessly focused on cash. Alphabet’s $561 billion backlog coincided with its second consecutive quarter of negative free cash flow, a year where it has halted stock buybacks, and issued over $41 billion in new debt and equity.
It is not about demand. It is not about revenues. The pressing questions are the return on capital and the price investors paid for it. For every major player except Microsoft, free cash flow has turned negative. This past quarter proved that investors are increasingly keeping an eye on this single metric, and they are no longer willing to underwrite the AI expansion without seeing the cash to back it up.
Multiple Compression and Peak Event Risk
Looking beyond AI and the hyperscalers' enormous investment commitments, the broader U.S. stock market has undergone a meaningful de-rating in 2026. The S&P 500 is up nearly 11% on the year, yet its forward P/E has fallen to 18.7x. This is below its 5- and 10-year averages of 19.8x and 19.1x, respectively, and represents a 16% haircut from the 22.2x multiple it commanded at the start of the year.
There is little debate that S&P earnings have been superb. Third-quarter results are tracking toward +30% year-over-year growth, marking the third consecutive quarter of +25% EPS expansion. Estimates for Q4 project another +25% growth rate, putting calendar 2026 results somewhere north of +30%.
However, it is looking increasingly likely that earnings growth has peaked. This rate-of-change delta is what I believe the market is sniffing out, driving the multiple compression we’ve seen this year alongside other steady headwinds. Looking ahead to 2027, consensus estimates have EPS growth decelerating to +15% – still solid, but a noticeable step down nonetheless.
Let’s leave next year for next year and pivot to how things are stacking up heading into year-end. A valid argument can be made that by the end of this week, we will be past the point of "peak event risk." Setting aside the unpredictable path of the U.S./Iran conflict, by Friday we will know the outcome of the mid-terms, the market will have a firmer handle on the Fed’s policy path, the heart of Q3 earnings season will be behind us, and we will have digested the Treasury’s QRA (Quarterly Refunding Announcement) issuance schedule.
Removing these uncertainties serves as a clearing event for markets. Add the fact that we are entering the most constructive part of the seasonal calendar for equities – having just cleansed a significant amount of froth in both sentiment and positioning – and the deck appears stacked in favor of the bulls. A year-end rally seems more likely than the market simply limping to the finish line. That said, this is still not a setup where investors should be extending themselves too far out on the risk curve.
Speaking of market-moving events, this week is loaded: Apple reports tonight, AMD follows on Tuesday, alongside the midterms (with Kalshi currently pricing 65% odds of a Democratic sweep). Wednesday brings the QRA details, culminating in Friday’s payroll data following last month's soft +29K print.
One note on Apple (AAPL): it enters earnings riding an all-time high P/E multiple, both on an absolute basis (36x forward P/E) and relative to the broader market (a 17x premium to the S&P). The one mega-cap still trading like a safe haven is also the most expensive it has ever been.
This brings me to the defensive corners of the equity market where capital has recently rotated. Just as momentum and growth are hitting oversold territory (the 30th and 40th percentiles, respectively), defensive segments are screening rich. The dividend yield factor has reached its 90th percentile, and the valuation factor sits in its 80th percentile on a 10-year lookback. Nothing says "defensive" quite like paying a growth multiple for low-volatility stocks. Yes, that is sarcasm, but it is a fitting sentiment for a market that has become relentlessly frustrating.
The Breakdown in Bullion
I am going to close out this week’s missive with some thoughts on gold, which recently slipped below $4,000 per ounce, setting a new low for the year at the close of last week. The prior YTD low of $3,976 could not hold—a discouraging development from a technical standpoint. Given this breakdown, it is reasonable to expect gold could slip to $3,500 per ounce before this correction has fully run its course (which would be a further 7% decline from Friday’s close).
We continue to hold firm with our gold allocation in client portfolios. While it is clearly not working this year, I still want it as a hedge in a mad world awash in debt and fraught with geopolitical risks. However, I recognize that gold is currently behaving in a bizarre and unpredictable pattern, with traditional correlations breaking down. Recently, neither dollar weakness nor lower market rates have managed to spark anything but weakness in bullion.
It is telling that on Friday, gold underperformed Bitcoin (-2.5% to $76,813)—and crypto at least had a fundamental reason to slide following the $6.3 billion Binance hack that exposed deep vulnerabilities in the crypto financial infrastructure. Recently, gold has acted more like a risk asset, an entirely new dynamic likely driven by short-term positioning and factor-based algorithmic trading. Over the long term, I expect the fundamental drivers of gold will eventually reassert themselves, reminding investors why it has historically proven to be one of the few true diversifying assets.
The articles and opinions in "Capital Market Musings and Commentary" are for general information only, and not intended to provide specific investment advice. Performance, dividends and other figures have been obtained from sources believed reliable but have not been audited and cannot be guaranteed. Past performance does not ensure future results. Investing inherently contains risk including loss of principle. Advisory services offered through Casilio Leitch Investments, an SEC registered investment advisor. Copyright © 2026 Casilio Leitch Investments. All Rights Reserved.