WEEKLY MARKET & ECONOMIC COMMENTARY
Week Ending Oct 23rd, 2026
Market Recap
Two of the four major indices ended the week on Oct 23rd up, and two closed the week down. The DJIA +0.41% and the Russell 200 +0.31%; the Standard & Poor's 500 index slipped 0.5% this week, led by a slide in communication services amid worries over corporate spending on artificial intelligence.
The S&P 500 ended the week at 7,623.68. With just one week remaining in the month, the market benchmark is down 0.4% for October but up 11% this year.
Worries ramped up over increasing capital expenditures as more companies released quarterly results, especially in relation to artificial intelligence.
Facebook parent Meta Platforms (META) reported Q3 results above views and raised the bottom end of its guidance range for 2026 capital expenditures. Tesla (TSLA) posted better-than-expected Q3 earnings but elevated operating expenses for AI projects "dragged meaningfully" on the electric vehicle maker's operating profit margin, Truist Securities said in a note to clients.
The week's drop came on declines in just four of the S&P 500's 11 sectors. Communication services tumbled 6.2% and industrials fell 1.9%. Technology and materials also edged lower.
Treasury yields fell last week as investors sought safe-haven assets amid AI-spending fears that reverberated across credit markets, sending the 10-year US Treasury yield below 5%. Internal forecasts from Meta Platforms disclosed last week said it needs to raise hundreds of billions of dollars to fund its AI ambitions, increasing concerns over the supply of debt needed to fund the AI infrastructure buildout. Debt supply concerns helped push global credit spreads to their widest in 17 months. Meanwhile, escalating geopolitical tensions, particularly in the Middle East, sent oil prices higher for the week, with Brent crude settling on Friday at its highest level since April. Economic data released last week showed US business activity expanded in October based on S&P Global’s preliminary US Composite PMI. While growth slowed from September, both services and manufacturing activity remained robust. However, costs also rose sharply, particularly as the impact of higher oil prices affected a wide variety of goods and services.
By Sector
AT&T (T) posted the largest percentage drop in communication services, down -11%. The telecommunications company raised its full-year earnings outlook following stronger-than-expected third-quarter results, but its postpaid phone subscriber growth missed projections. Meta also got hit hard, sliding -9.4% on the week. The company's Q3 adjusted earnings per share and revenue topped analysts' expectations but the stock was weighed down by concerns about its increasing capital expenditures related to AI.
Comfort Systems USA (FIX) had the largest percentage drop in the industrials sector as well as the overall S&P 500, falling 19%. The company reported Q3 earnings and revenue above year-earlier results and analysts' mean estimates, but it announced plans for Chief Financial Officer Bill George to retire. Gainers were led by the energy and health care sectors, up 1.55% and 1.45%, respectively. Real estate, utilities, financials, consumer discretionary and consumer staples also edged higher.
The energy sector's advance came as crude futures rose amid geopolitical tensions. SLB (SLB) was the best performer in the energy sector, climbing +7.2%. The oilfield services company reported better-than-expected third-quarter results as a surge in North America revenue and momentum in its data center business offset weakness in its Middle East operations.
Intuitive Surgical (ISRG) was among the top gainers in health care, rising +6%. The company reported Q3 adjusted earnings and revenue above analysts' consensus views. Meanwhile, Intuitive Surgical adjusted its global da Vinci procedure growth target range for 2026 and reported a $450 million settlement charge that outweighed its Q3 beat.
This week, Verizon Communications (VZ), Visa (V), Coca-Cola (KO), Boeing (BA), Google parent Alphabet (GOOG, GOOGL), Microsoft (MSFT), GE Vernova (GEV), Amazon.com (AMZN), Eli Lilly (LLY), Mastercard (MA), Intel (INTC), Merck (MRK), Caterpillar (CAT), ExxonMobil (XOM), Chevron (CVX) and AbbVie (ABBV) are among companies releasing quarterly results.
Economic data will include September personal consumption expenditures and the advance estimate of Q3 gross domestic product. There will also be a two-day meeting of the Federal Reserve's Federal
Open Market Committee that concludes Wednesday.
The $800 Billion Question
Big Tech’s AI spending faces its toughest judge this week: free cash flow
This is the busiest week of the earnings season. Roughly a third of the S&P 500 reports, including three of the world’s largest companies (Microsoft and Alphabet on Wednesday after the close, Amazon on Thursday). This will tell us whether the most expensive corporate building program in modern history is starting to pay for itself. Apple follows next Monday. Wednesday’s reports will be announced just hours after the FOMC rate decision, so there is an unusual amount drama being over a few days.
For most of the past three years, investors rewarded these companies simply for spending on artificial intelligence. That changed last week. Meta Platforms beat revenue expectations and issued a stronger-than-expected outlook, and its stock still fell about -7% in a single session. The problem was not earnings. It was cash flow.
That shift, from judging AI spending by its promise to judging it by the cash it consumes, is the most important thing to understand heading into the week.
Three Terms That Matter This Week
Capital expenditures (capex) are the dollars a company spends on long-lived physical assets. For these companies, that means data centers, AI chips, networking gear, and the power equipment to run it all.
Free cash flow is the cash a business generates from its operations, minus its capex. It is what is actually left over for owners: money available for dividends, buybacks, paying down debt, or simply keeping a cushion. If earnings are what a company reports, free cash flow is what it can actually spend.
Free cash flow yield is free cash flow divided by the company’s market value. It works like an earnings yield but uses cash instead of accounting profit. A 2% free cash flow yield means that for every $100 of stock you own, the business generates $2 of spendable cash per year.
If earnings are what a company reports, free cash flow is what it can actually spend.
Why do earnings and cash tell different stories right now? Accounting. When a company buys $100 billion of servers, it does not record $100 billion of expense that year. It spreads the cost over the equipment’s useful life. Using six years for illustration, that is roughly $17 billion of expense per year, while the full $100 billion leaves the bank account up front. In a normal year the gap is small. In a year when spending nearly doubles, it is enormous. Industry-wide tech infrastructure capex is expected to reach about $790 billion in 2026, up from roughly $404 billion in 2025, according to S&P Global’s Visible Alpha. That is why reported earnings can look healthy while free cash flow shrinks or turns negative.
Meta Fired the Warning Shot
Meta’s third-quarter report was, on the surface, a good one. Revenue grew 26% to $64.6 billion, and fourth-quarter revenue guidance of $73 to $76 billion came in above Wall Street’s estimate. But capex for the quarter reached $34.9 billion, and free cash flow swung to negative $2.1 billion, compared with positive $10.6 billion a year earlier. In other words, Meta’s advertising business, one of the most profitable businesses ever built, no longer covered its AI spending on its own.
Management also raised the floor of its 2026 capex range to $135 to $145 billion. With $85.8 billion spent through September, that implies roughly $49 to $59 billion in the fourth quarter alone, more than in any prior quarter. Meta announced a new class of nonvoting shares, in part to give itself another way to finance growth, and the cost of insuring Meta’s debt against default (its five-year credit default swap spread) hit an all-time high. The stock closed at $648.85 on October 22, down 7% on the day.
On adjusted earnings, Meta trades at roughly 22 times its current run rate, which is not demanding. On free cash flow, there is currently nothing to value. The market chose to look at the cash, and that sets the standard for this week.
Where the Big Three Stand
On a price-to-earnings basis, the three companies reporting this week look remarkably similar, all trading at roughly 23 to 25 times expected earnings. On cash, they look very different.
Microsoft is funding a record build largely from operating cash flow, with room to spare. It guided to more than $50 billion of capex this quarter, up from roughly $41 billion last quarter, against Azure cloud growth of about 45%, and management has committed to positive free cash flow for fiscal 2027.
Alphabet spent more than it generated in the second quarter: $44.9 billion of capex against $39.1 billion of operating cash flow, leaving about $5.9 billion of negative free cash flow. Its forward multiple has already compressed from about 28 times at the end of 2025 to about 23 times today, so some of that concern is in the price.
Amazon is in the most capital-intensive position of the three. Capex is running at about 108% of operating cash flow on a plan of roughly $220 billion this year, leaving trailing free cash flow slightly negative.
Here is the comparison that matters most to us. Microsoft, the strongest cash generator of the group, offers a free cash flow yield of about 1.7%, below its five-year average of about 2.3%. The 10-year Treasury pays about 5%. An investor buying Microsoft today is accepting roughly one-third of the cash yield of a risk-free government bond, in exchange for the expectation that free cash flow roughly triples over time. That may well happen. But it is a bet on future returns from spending, not a reflection of cash the business is producing today.
Investors are accepting about one-third of the cash yield of a Treasury bond, on the expectation that free cash flow roughly triples.
From Cash to Credit: The Financing Shift
Until recently, these companies paid for everything out of their own pockets. That is no longer true. FactSet estimates that new borrowing went from about 9% of hyperscaler capital spending in fiscal 2024 to about 32% on a trailing basis by mid-2026. Estimates for total AI-related debt issuance this year range from roughly $450 to $570 billion.
The cost of that borrowing is rising. Bonds from AI-related issuers have recently traded at spreads (the extra yield over Treasuries) of around 1.15 percentage points, versus about 0.78 for the broader investment-grade market, according to Goldman Sachs and ICE BofA data reported by Reuters. Amazon paid extra yield on the longest bonds in its $25 billion July offering, and a bond tied to a Meta data center in Texas was expected to yield about 7.5%. Some obligations do not show up as debt at all: Alphabet disclosed $85.2 billion of future payments on data center leases that had not yet begun as of June 30.
The stress is most visible at the edges. The cost of insuring CoreWeave’s debt, one of the so-called neoclouds that rent AI computing capacity, reached a record on October 15. Meanwhile, the semiconductor index finished last week about 26% below its peak from last summer, and Thursday’s AI-driven selloff gave the Nasdaq its worst session since June.
This matters to income investors, too. AI-related debt now makes up roughly 10% of the largest long-term investment-grade corporate bond funds, up from about 7% a few years ago. Many investors who believe their bond allocation is insulated from the AI trade own more of it than they realize.
Three Tests for This Week
Is revenue keeping pace with spending? Cloud growth is where AI demand shows up first. If Azure, Google Cloud, and AWS accelerate faster than capex grows, the spending is earning its keep. If capex rises again while growth merely holds steady, the market will apply the same discount it applied to Meta.
What do they say about 2027? Meta declined to give a 2027 capex figure but said aggressive investment will continue. A large step-up in 2027 spending plans, without a matching revenue story, would push free cash flow estimates lower and stretch valuations.
How is it being paid for? Compare operating cash flow to capex, and watch new borrowing and lease commitments. A company funding its build from cash deserves a different multiple than one funding it with debt.
What It Means for Investors
First, earnings multiples alone can mislead in a capital spending boom. Three companies trading at nearly identical price-to-earnings ratios have very different cash profiles, and over time investors are paid in cash, not accounting profit.
Second, owning a broad market index means owning this bet without any discrimination between the companies funding AI from strength and those funding it with borrowed money. The index does not ask whether each dollar of capex is earning a return. We do.
Third, expect volatility in both directions. With sentiment already bruised, strong results could spark a sharp relief rally, and weak ones a sharp decline. Neither a single quarter nor a single day’s price move should drive a long-term decision. The questions that matter are the three above, and the answers will play out over several quarters, not one.
Artificial intelligence may well be the defining investment theme of this decade. That does not make every dollar spent on it a good investment, and it does not make every stock tied to it a good price. This week, the market starts telling the difference.
The Week Ahead
This is shaping up to be one of the most consequential weeks of the quarter, with a packed calendar of mega-cap earnings, central bank rate decisions and the first look at 3rd quarter GDP. For equities, investors will focus on a key question: whether strong corporate earnings can offset growing concerns about escalating AI infrastructure spending and the potential for higher energy prices to keep inflation elevated. Microsoft, Alphabet and Amazon will provide a critical test of whether the AI investment cycle continues to justify the extraordinary capital commitments. Monetary policy will provide the other major catalyst, with the Federal Reserve widely expected to leave the Fed Funds target rate unchanged. Fed Funds futures pricing reflects an only 10% probability of a 25-bps hike for Wednesday’s meeting, but an 80% chance for higher rates at their next and final meeting for the year in December. Thursday offers our first look at 3rd Quarter GDP, along with September’s PCE data; being that the latter is “The Fed’s preferred inflation gauge”, the FOMC committee members will certainly have access to it a day early to help inform and guide their decision and statement. Internationally, we’ll hear from a number of central banks: The Reserve Bank of New Zealand and Bank of Canada are widely expected to raise rates on Wednesday, while the European Central Bank and Bank of Japan are expected to hold rates steady following their recent tightening. With roughly one-third of S&P 500 companies reporting during the week, the combination of earnings announcements, forward guidance surrounding cap ex spending, and central bank announcements will offer investors plenty of news to digest alongside whatever geopolitical surprises may be in store.
Important Disclosures
This commentary is provided by Optimized Capital for informational and educational purposes only and does not constitute investment, tax, or legal advice, or a recommendation to buy or sell any security. Opinions expressed are as of the date of publication and are subject to change without notice. Information has been obtained from sources believed to be reliable, but its accuracy and completeness are not guaranteed. Valuation figures are approximate and may change materially following the earnings reports discussed. Past performance is not indicative of future results. All investing involves risk, including the possible loss of principal. Please consult your advisor regarding your individual circumstances.
