The most important thing that happened today wasn't the Dow's 554-point gain — it was Corning's stock crashing 12% on a quarter where its core AI growth engine (Optical Communications) accelerated to +32% year-over-year. That is the tell for the whole tape right now: investors have stopped rewarding AI-adjacent companies for accelerating fundamentals and are punishing them for any negative delta in tone, no matter how small. My stance: near-term bearish on chips/high-multiple AI infrastructure into Wednesday's Fed decision and Thursday's Apple/Amazon prints, but constructive on broad-market breadth — nine of eleven S&P sectors were green today. Conviction: moderate-high on the "discriminating market" thesis, moderate on the index-level call given how binary this week's FOMC outcome is.
| Index | Close | Change |
|---|---|---|
| S&P 500 | 7,413.18 | +0.2% |
| Dow Jones Industrial Average | 52,210.08 | +554 pts (+1.08%) |
| Nasdaq Composite | ~24,935 | -0.2% |
| Nasdaq-100 (QQQ proxy) | $675.49 | -0.97% |
| Russell 2000 (IWM proxy) | $293.37 | +0.16% |
Sector leaderboard: Best — Health Care (XLV) +2.36%, Consumer Staples (XLP) +1.99%, Communication Services (XLC) +1.87%. Worst — Technology (XLK) -1.84%, Energy (XLE) -1.35% (oil's third straight down day), Industrials (XLI) -0.39%. This was money leaving the AI-infrastructure trade and rotating into defensive quality, not a broad risk-off day — the Dow beat the Nasdaq by 130 basis points.
VIX: closed near 18.2, actually down slightly from Monday's 18.67 close. In plain English: options traders are not paying up for protection even though the chip sector just had one of its worst days and worst months in over a decade (the Philadelphia Semiconductor Index is down roughly 4% today alone, and memory names like SanDisk are down double digits). That gap between "stocks are getting crushed" and "volatility pricing is calm" is a warning sign, not a comfort signal — it means downside insurance is still cheap relative to the risk actually on the table this week.
Treasury yields: the 10-year sits around 4.63-4.65%, down modestly from last Thursday's 4.70-4.72% cycle high (the highest level since January 2025) but still elevated. What it signals: the bond market is pricing real hike risk into tomorrow's Fed decision — CME FedWatch has the odds of a surprise quarter-point hike (not a cut) at 31%, up from 26% just a week ago. A 10-year push back above 4.70% would tighten financial conditions further and compress AI-capex multiples another leg down; a drop back under 4.50% would be the clearest signal yet that the Fed is done tightening and would likely ignite the "quality rotation" names further.
The catalyst: earnings, not macro. This is the single busiest week of the season — 169 S&P 500 companies report, including four of the Magnificent Seven (Microsoft and Meta Wednesday after the close, Apple and Amazon Thursday after the close) — sandwiched around Wednesday's 2 p.m. ET Fed decision. Today's action was the market pre-positioning for that gauntlet: sell anything with AI-capex exposure and rotate into things that already reported clean, capex-light beats (Coca-Cola, Sherwin-Williams).
Narrative challenged: "Strong earnings get rewarded." According to FactSet, of the roughly 30% of S&P 500 companies that have reported, nearly 90% have beaten earnings estimates — yet Bank of America found that beat-and-raise reports are only outperforming the index by 1.2 points the next day (versus a historical 1.4-point average), and tech beats specifically are lagging the index by 2.6 points. Corning is today's extreme case study: Q2 core sales grew 17% to $4.74 billion, core EPS jumped 30% to $0.78, and its AI-linked Optical Communications segment grew sales 32% year-over-year to $2.07 billion with net income up 77%. The stock still fell 12.1% because Q3 core-sales guidance of $4.9-5.0 billion read as capacity-constrained against a smartphone-glass slump. A name can beat, raise, and accelerate its highest-growth AI segment and still get punished on a guidance nuance — that is a market pricing on tone, not on fundamentals, and it is the single biggest thing "buy the dip" investors are underestimating this week.
Narrative strengthened: quality and profitability now matter more than growth headlines. UPS actually beat earnings and raised full-year guidance today — and the stock still fell 6.6%, because investors read the raise as coming from cost discipline after Amazon further cut its delivery volumes with UPS, not from genuine demand strength. Same story, different sector, same lesson: a beat with deteriorating underlying quality is being sold, not bought.
What most investors are overlooking: the "core capital goods orders" print buried inside Monday's disappointing durable goods report. Headline durable goods rose only 0.3% in June versus a 2% estimate — a clean miss that dominated headlines. But core capital goods orders (the best proxy for actual business investment, stripping out volatile aircraft/defense orders) rose 0.9% month-over-month and 12.5% year-over-year — the fastest annual pace since November 2021, driven by computer and electronic product orders running at a 15.5% annualized rate since December, the best since 2008. Real, hard capex spending is still accelerating even as the narrative says "the AI trade is dying." That is the strongest argument that this week's selloff is a multiple re-rating, not evidence of an actual capex slowdown — which matters enormously for how violently these stocks can snap back once the Fed and mega-cap earnings are behind us.
Real-world consequence: Fitch flagged Monday that "an AI-related market correction" combined with Middle East uncertainty could become a "major credit risk," and a recent Bank of America fund manager survey found nearly half of managers now cite AI capex spending as the most likely source of a systemic credit event — up from about a third just two months ago. Hyperscaler capex is now projected near $700 billion in 2026, an estimated one full percentage point of U.S. GDP growth on its own (First Trust). If this week's earnings confirm spending discipline is real, credit markets calm down. If Microsoft, Meta, Apple, or Amazon guide capex even higher without a clearer monetization story, expect credit-spread widening to follow the equity selloff with a lag — exactly the private-credit and hyperscaler-debt risk we have been tracking all month.
Most surprising mover — and what it signals: Corning. A company that just posted its best Optical Communications growth in years crashed harder than SanDisk, a name genuinely facing new Chinese competition. That gap tells you the market isn't discriminating between "AI capex real risk" (SanDisk/memory oversupply) and "AI capex tone risk" (Corning's guidance phrasing) right now — everything capex-adjacent is getting sold first and sorted out later. That is a dangerous, indiscriminate pattern precisely because it means Wednesday and Thursday's mega-cap prints could either extend this violently or trigger an equally violent, indiscriminate relief rally the moment one of them delivers real spending discipline.
Use S&P 7,413 (the 50-day moving average) as your line in the sand into the Fed decision. A confirmed close below 7,400 tomorrow — before the 2 p.m. ET Fed announcement — would be a rare pre-Fed capitulation signal; trim high-beta semiconductor exposure (SOXX, MU, AMD, SNDK) and raise cash ahead of the decision rather than holding full size through a binary hike/hold event with 31% hike odds already priced.
Rationale: implied volatility (VIX ~18.2) is not pricing the dispersion already visible in the chip complex — protection is cheap relative to the risk of a genuine surprise hike.
LONG Corning (GLW) on further weakness in the $126-$133 zone (the stock closed today at $126.01, inside this band). Optical Communications sales accelerated to +32% YoY with net income +77% — the exact AI-infrastructure demand data the bears claim doesn't exist. Entry: $126-$133. Target: $172 (roughly 40x 2027E core EPS of $4.20), Timeframe: 1-3 months. Invalidation: two consecutive daily closes below $133, a sequential decline in Optical Communications sales, or a Q4 core sales guide below $5.0 billion. Discipline rule: do not add until the stock prints two consecutive closes back above $133 — today's flush does not yet confirm a bottom.
Rotate incremental new cash into Health Care (XLV) and Consumer Staples (XLP) — today's two best-performing sectors (+2.36% and +1.99%) — or direct names already showing genuine beat-and-raise quality (KO, SHW). Rationale: Fitch explicitly warned Monday that an AI-driven market correction could become a "major credit risk," and BofA's latest fund manager survey shows nearly half of managers now see AI capex as the top systemic risk, up from about a third two months ago. Owning quality, capex-light, dividend-paying names is the cleanest hedge against that specific, underappreciated tail risk — not a bet against AI, but insurance against the credit-market spillover if this week's earnings disappoint.
Most important event: the FOMC decision Wednesday, July 29, 2 p.m. ET. This is genuinely two-sided — CME FedWatch prices a 31% chance of a surprise hike (up from 26% a week ago), with markets also pricing a 75% chance of at least a quarter-point hike by September. Fed Chair Kevin Warsh has deliberately pulled back on pre-meeting guidance, so this is one of the least-telegraphed Fed decisions in recent memory.
Key level to monitor: S&P 500 relative to 7,413 (its 50-day moving average) heading into and out of the Fed decision — a break and hold below 7,400 opens the door to 7,185 (the 100-day moving average, roughly 3% lower); reclaiming and holding 7,470 would be the cleanest "all clear" signal since the chip selloff began.
Three things to watch and why:
The mainstream story this week will be "hike or hold" and "did Big Tech guide capex too high." The more useful, less-discussed trade is a volatility-dispersion mispricing: the VIX closed today near 18.2 — barely changed and actually a touch lower than Monday — while the underlying chip complex logged one of its worst single sessions and worst months in over a decade (SanDisk -14% today, Corning -12%, the DRAM ETF down roughly 9%). Headline index-level volatility is not capturing the dispersion already visible stock-by-stock. That means cheap, targeted hedges — SOXX puts or short-dated VIX calls positioned into Wednesday's Fed decision and Thursday's Apple/Amazon prints — are underpriced convexity, not a "sky is falling" trade. Separately, the accelerating core capital-goods-orders data (+12.5% YoY, the fastest since November 2021) tells me the underlying AI capex cycle has not actually slowed — this week's carnage in names like Corning is a multiple re-rating on tone, not a revenue re-rating on fact. When the dust settles after Thursday, I expect names punished purely on guidance language rather than actual demand deceleration — Corning chief among them — to be the sharpest snap-back candidates of the whole complex.