US Stock Market Daily Review — Friday, July 24, 2026

Published by @dailyanalysts · Market close data as of 4:00pm ET / 22:01 UTC

HEADLINE VIEW: The market just proved it can no longer tell the difference between "good" AI capex and "bad" AI capex — Intel delivered its best quarter in 15 years and still got sold 7.9%, while the Nasdaq lagged a flat-to-green S&P/Dow as oil's sharp reversal (Brent -4% to $96.78) offset the chip-complex rout. STANCE: CAUTIOUSLY BEARISH ON MEGA-CAP TECH, CONSTRUCTIVE ELSEWHERE (Medium-High Conviction) — the broadening-out trade (industrials, staples, real estate, software) is real and intact, but the next 5 trading days (FOMC Wednesday, then Microsoft/Meta/Amazon/Apple earnings) are the highest-stakes stretch of the summer.

1. Market Snapshot

IndexCloseChange% Change
S&P 5007,413.30+3.74+0.07%
Dow Jones Industrial Average51,947.52+235.87+0.46%
Nasdaq Composite24,980.34-157.35-0.63%
Russell 2000 (IWM proxy)~2,912 (IWM $291.17)-0.92-0.32%

Source: Reuters preliminary close data

Sectors — Best & Worst (SPDR sector ETFs)

Best%Worst%
Real Estate (XLRE)+2.22%Technology (XLK)-1.44%
Materials (XLB)+1.93%Small Caps (IWM)-0.32%
Consumer Staples (XLP)+1.11%

Technology was the only red S&P sector today, dragged down entirely by chips (Intel -7.9%, Micron, SanDisk -10%+, Marvell -7%). Real estate led on Digital Realty Trust's raised FFO guidance. This is a genuine "market broadening" signal, not just a tech story — 10 of 11 sectors were green.

VIX — What It Means

The Cboe Volatility Index closed at 17.52, down 6.3% from Thursday's 18.70 close (which itself had spiked 12.4% on the GOOGL/TSLA rout). In plain terms: the acute fear from Thursday's crash is fading, but 17.5 is still elevated versus the low-teens "calm" readings seen earlier this year — investors are pricing real uncertainty into next week, not panic today.

Treasury Yields

The 10-year yield eased slightly to 4.68% from Thursday's cycle-high near 4.70-4.71% (highest since January 2025), per Trading Economics. The 30-year is closing in on 5.2%, with some strategists warning a run to 6% would be "hugely negative" for equity multiples. Translation for average investors: mortgage and auto-loan rates are moving up in lockstep with the 10-year right now — this is a real cost-of-capital problem, not just a bond-market abstraction.

The One Technical Level That Matters

S&P 500 7,400 / SPY $740 — flagged this week by SpotGamma/Barchart options positioning as the negative-gamma flip point. Today's close of 7,413.30 sits right on top of that line. A confirmed break back below 7,400 means dealer hedging flows start amplifying moves instead of dampening them, opening a fast path to ~7,260 (June's closing low). A reclaim of the 50-day moving average (~7,470) on real volume would be the all-clear signal. The S&P closed below its 50-day average Thursday for the first time in weeks — that's the line to watch Monday.

2. Story Behind the Numbers

The catalyst: Two forces fought each other today. Oil's sharp reversal — Brent crude fell nearly 4% to $96.78 and WTI fell 3% to $89.31 (CNBC) — on reports that Pakistan, backed by China, is pushing to restart U.S.-Iran talks, gave the Dow and S&P room to recover. But chip stocks got hammered on AI-capex-sustainability fears ahead of next week's Microsoft/Meta/Amazon reports, pulling the Nasdaq the other way.

Narrative strengthened: The market is no longer giving anyone — hyperscaler or hardware maker — a pass on capital spending. Moody's Ratings put a number on it this week: the six hyperscalers it tracks (Microsoft, Amazon, Alphabet, Meta, Oracle, CoreWeave) are set to spend $785 billion on AI capex in 2026, rising to roughly $1 trillion in 2027 — with direct debt already at ~$460 billion and off-balance-sheet lease commitments ballooning to $1.2 trillion. This is now a credit story, not just a sentiment story.

Narrative challenged: The clean "sell hyperscalers, buy chips" rotation that worked yesterday broke down today. Intel beat on every metric that matters — revenue of $16.1 billion (its fastest growth in 15 years, +25% YoY), EPS of $0.42 (double the $0.21 estimate), data center revenue +59% YoY, 18A node yields running 25% above target (CNBC) — and still closed down 7.9% because it didn't announce a new foundry customer. If a genuine beat-and-raise with real proof points gets sold this hard, the market isn't discriminating between good capex (monetized, proven) and bad capex (GOOGL/TSLA-style, unproven, FCF-negative) — it's punishing capex exposure itself, indiscriminately. That's a flow/positioning story, not a fundamentals one, and it's the most important mispricing in the market right now.

What most investors are overlooking: the bond market. Oracle's 5-year credit default swap is trading at a multi-year high and has become, in Barclays' words, "a proxy for AI debt fears" (CNBC). Meta is reportedly financing its $12 billion Texas data center at a higher rate than prior deals. Credit spreads are the leading indicator here — they will crack before the equity multiples do, and right now they're already widening.

Real-world implication: New 10-12.5% tariffs on 60 trading partners took effect today (replacing an expired temporary 10% global tariff), oil is still up ~10% for the week even after today's pullback, and jobless claims just hit a cycle low of 187,000 — a combination that hands the Fed cover to hike, not cut. Fed funds futures price a 33.7% chance of a hike at next Wednesday's meeting (not zero), and core PCE data due Thursday is expected to print above 3%. Mortgage rates are already moving with the 10-year in real time.

3. Company Spotlight

Winners

Losers

Most Surprising Mover

Intel. A stock that popped 4% in after-hours trading Thursday and was still up 3% in Friday's premarket round-tripped into a 7.9% loss by the close. That is the clearest single data point today that the market has stopped distinguishing between capex that's paying off (Intel: record data center growth, 18A yields ahead of plan) and capex that isn't (Alphabet: negative FCF, Tesla: stagnant margins). The signal: expect Microsoft, Meta and Amazon to face the identical bar next week — beat the headline number and you'll still get sold unless free cash flow visibly improves.

4. What To Do Now

1. ACTIONABLE FOR MONDAY (Short-Term Traders): Trim or hedge single-name concentration in Microsoft, Meta and Amazon ahead of their Wednesday/Thursday earnings, via QQQ puts or reduced position sizing.
Rationale: GOOGL and TSLA already got the "beat headline, sold on capex/FCF" treatment this week, and Intel just proved even a clean beat with real execution proof isn't safe. With the FOMC decision Wednesday at 2pm ET (33.7% priced odds of a surprise hike) and core PCE Thursday expected above 3%, this is the most event-dense 48-hour window of the summer. VIX at 17.5 is cheap insurance relative to what's coming.
Entry: Hedge initiated now through Tuesday · Target: Protect against a 5-8% single-day gap similar to GOOGL/TSLA this week · Invalidation: Cover the hedge if MSFT (Wed) beats with flat-to-improving free cash flow · Timeframe: 1-3 days · Conviction: SPECULATIVE (one clear signal — repeated pattern — but unconfirmed for these specific names)
2. CONTRARIAN MOVE (Long-Term Investors): LONG INTEL (INTC)
Rationale: The sell-off is sentiment and positioning (stock was +172% YTD, no new foundry customer announced), not fundamentals — record data center revenue (+59% YoY), EPS beat by 2x, 18A yields running 25% above target, and a Q3 guide above Street consensus. Bank of America ($160 PT) and Goldman Sachs ($150 PT) targets imply 60-70% upside even while several banks keep neutral/hold ratings. The market conflated "capex-heavy" with "unproven ROI" — Intel already has the proof, unlike its hyperscaler peers.
Entry: $90-96 (closed today at $92.32) · Target: $115 (conservative blend of Wells Fargo $120 / Bernstein $110, below BofA/Goldman) · Invalidation: Weekly close below $85 (JPMorgan's bear case) or announcement of a dilutive capital raise · Timeframe: 1-3 months · Conviction: HIGH CONVICTION (fundamental beat + multiple independent analyst target hikes + record data center growth = 2+ signals)
3. DEFENSIVE POSITION (Long-Term Investors): Utilities (XLU)
Rationale: Utilities were green today (+0.2%) even as tech wobbled, benefit structurally from real (not speculative) AI-driven power demand, and offer ballast heading into the seasonally weak August-September stretch (S&P 500 historically down 0.4-0.8% in midterm-election-year Aug/Sept, per Stock Trader's Almanac) plus oil-driven inflation risk that could force the Fed's hand.
Entry: $45.80-46.50 (closed today at $46.29) · Target: $49 · Invalidation: Weekly close below $44.50 · Timeframe: 1-3 months · Conviction: WATCH (thematically supported, no confirmed trigger yet)

Carryover Suggestion Status (from July 23 AI Open-Source Repricing trade)

GOOGL (entry $305-320, target $380/$420+, invalidation daily close <$290) — closed today at $319.74 (+0.65%), stabilizing after Thursday's capex-driven crash. Still open, not invalidated. NVDA (entry $195-210, target $250, invalidation <$185) — closed at $206.84 (-0.9%), within entry zone, not invalidated. Neither suggestion has hit target or stop; both remain live positions to monitor into next week's hyperscaler earnings.

5. Looking Ahead

Most important event: The FOMC meeting concludes Wednesday, July 29 at 2pm ET. The decision itself (consensus: hold, with a 33.7% priced probability of a surprise hike per CME FedWatch) matters less than the language given the internal Fed dispersion of views. Core PCE data Thursday, expected above 3%, could cement the hawkish case regardless of what the Fed says Wednesday.

Key price level: S&P 500 7,400 (SPY ~$740) — the gamma flip point. Watch whether Monday's open holds above it.

Three things to watch:

Conclusion — Highest-Conviction Take

The real story this week isn't Intel's stock price — it's that the market is now punishing any capex signal regardless of the proof behind it, which means the "sell chips, buy software" rotation is a positioning unwind, not a fundamentals call. That kind of indiscriminate selling is exactly the setup that snaps back hard and fast the moment just one of next week's hyperscalers — most likely Microsoft, given Azure's capacity-constrained, already-monetizing cloud growth — reports a beat with improving free cash flow. When that happens, the entire hardware complex (INTC, MU, AMD, NVDA) should re-rate in a straight line, because it removes the market's single biggest current fear in one print. The better risk-reward right now is in already-beaten-down, proof-of-execution names like Intel — not in chasing the still-unresolved hyperscaler names into a binary event.