Thursday, July 23, 2026 ALERT · Level 3 5 min read

The AI Trade Split in Two: Capex Punished, Chip Makers Spared

Risk-off hit the companies that finance AI, not the ones that build it: Alphabet fell 7% on rising capex while chips held and Intel jumped after hours.

The risk read for this session

Sector crash

REALISED

chips −20% from peak · within 1 month

Unchanged from the previous read

Deeper crash

30%

chips −30% from peak · within 3 months

Up 2 points from 28%

Market crisis

S&P 500 −20% · within 3 months, AI at the centre

Up 1 point from 7%

SOXX

551.24

−16.0% from peak

NVDA

208.76

−11.7% from peak

VIX

18.70

fear gauge

High-yield spread

268 bp

credit stress

These are model estimates, not forecasts anyone can guarantee, and not investment advice. The dot before each probability is its band: 🟢 low · 🟡 moderate · 🟠 high · 🔴 very high, with thresholds set separately for each scenario. SOXX is the main US semiconductor ETF, the basket I track; from peak is the fall from the highest close of this cycle; VIX is the market’s fear gauge; the high-yield spread is the extra interest riskier companies pay to borrow, my early-warning line for funding stress. My internal alert is at Level 3, on a scale where 3 is the highest.

📌 Summary (3 lines)

The first real "risk-off" day in a while: the S&P 500 fell −1.21% to 7,408, the Nasdaq Composite −2.15% (briefly under 25,000), and the fear gauge (VIX) rose to 18.7.
But the selling had a precise epicenter — the companies that FINANCE AI, not the ones that build it. Alphabet fell −7% for raising its AI spending; the chip makers held (SOXX −0.8%) and Intel jumped +12% after the bell.
Nearest events: the Federal Reserve meeting on July 28–29 (odds of a rate rise have climbed) and the big-tech spending updates on July 29–30.

Where I stand today: 🟠 ALERT, and my highest internal alert (Level 3) is active for a 6th day. All figures are for the US session of Thursday, July 23, 2026.

New here? Start with this. I'm Andrea Pilotti, and this is AI Market Observatory. I invest my own money in AI and semiconductor stocks, and I work with artificial intelligence for a living — I built this because I needed to measure the risk in my own exposure instead of guessing at it. Every trading day I track these stocks with data and turn what I find into a single risk read: a traffic light for the odds of a sharp fall in this corner of the market. This is an observatory, not a recommendation service, and I am not a financial advisor: my job is to notice trouble early and say it in plain language. Every daily read stands on its own, this one included.

The traffic light, once: green = normalyellow = a down day, no change in the pictureorange = alert, tension risingred = critical

Alongside it I run an internal alert from Level 1 to Level 3, where 3 is the highest and only steps down once a full recovery checklist is met. How every gauge and threshold is defined is public in my methodology, and every previous daily read is in the archive.

📰 The day's story

The trigger was fundamental, not technical. Alphabet reported a record quarter — revenue $119.8B, Google Cloud up 82% year-on-year — yet raised its 2026 "capex" (capital spending, the money it invests in data centers and chips) to $195–205B, pushing quarterly free cash flow negative (−$5.9B). Investors read it as proof that the AI build-out is now eating cash, and sold the stock −7%. Tesla (which had reported the evening before) made it worse: profit missed, free cash flow turned negative, capex jumped 142% — the stock fell −14% to an 11-month low. The whole "who funds AI" group closed red (hyperscaler group average −5.4%).

Under the surface, though, the semiconductors held. SOXX — the main US semiconductor ETF and my benchmark — lost just 0.8% (close 551.24) on light trading volume, supported by memory chips (Micron +3.2%, SK Hynix +2.6%), equipment (KLAC +1.9%, Applied Materials +1.6%) and data-center power names (GE Vernova +4.7%). The key read: the market punished the companies that SPEND on AI and rewarded the ones that RECEIVE that spending — the chip makers. Confirmation came after the close, when Intel (revenue +25%, data-center/AI +59%, strong outlook) jumped +12%. Nvidia, the sector's flagship, fell −1.56% to 208.76 and slipped just back below its 50-day average. Oil, meanwhile, surged on Middle-East escalation, lifting bond yields and rate-rise expectations.

🚦 My nine signals (traffic light)

🟡 Moving averages: Nvidia dipped back below its 50-day average (the average price of the last 50 days — the trend's "waterline") after one day above; SOXX is a 7th day below its own, but on light volume. Both remain well above their 200-day average.
🟠 Structure: lower high and lower close than the day before; the pattern of falling peaks since the June top holds.
🟠 Momentum: SOXX's weekly trend indicator (MACD) is still in a bearish cross; Nvidia's daily reading stays positive.
🟠 Breadth (how many take part): only 44% of my universe rose (23 of 52), below the 60% I'd want to see. The average stock actually beat the mega-caps today, because the mega-caps led the fall.
🟢 Rotation / canaries: my basket of five "canary" stocks is roughly in line with the sector; one of them, Oracle, closed at a new 52-week low.
🟠 Distribution (heavy selling): two new "distribution days" (SOXX and the Nasdaq-100 ETF both fell on higher volume) — now 10 active, a heavy count.
🟢 Volatility: the VIX (Wall Street's "fear gauge") rose to 18.7 but is still below 20. Fear is building; complacency is fading.
🟢 Credit: the extra yield risky companies pay to borrow (HY OAS) is 268 basis points — still near cycle lows. No stress here.
🟠 Euphoria / fragility: SOXX is still +39% above its 200-day average (above my +35% guard line): a "stretched" market. Out of the score, but flagged.

Put together: three signal families are under tension, none broken, and SOXX is still far above its long-term line — so the picture is ALERT, not critical.

📏 Key levels (recalculated on the close)

SOXX 551.24, down −15.96% from its June peak of 655.95; 50-day average 569.00 (−3.1%), 200-day 396.36 (+39.1%).
Nvidia 208.76, −11.74% from its May peak; 50-day 209.46 (−0.3%), 200-day 192.80 (+8.3%).
S&P 500 7,408.30, only −2.65% from its record close. The −20% "bear market" line sits about 18% below here.
A 20% sector drop already happened on July 17 and still counts inside its 1-month window.

🏭 Sector groups (average move)

Green: power/data-center +1.9%, memory +1.5% (all up), equipment +1.0% (all up), networking +0.8%, servers +0.3% (Super Micro +2.1%).
Red: chip design −1.8% (all 10 names down; Nvidia −1.6%, Texas Instruments −3.1%), and hyperscalers −5.4% (Tesla −14.5%, Alphabet −7.1%, Oracle −4.6%, Amazon −4.6%). A clean split between chip suppliers (up) and AI financiers (down).

📊 My probabilities (model estimates)

Bands as dots: for the 3-month "severe drop" — 🟢 under 5%, 🟡 5–12%, 🟠 12–25%, 🔴 over 25%. For the "market crisis" — 🟢 under 5%, 🟡 5–10%, 🟠 10–20%, 🔴 over 20%.

Sector drop of 20% within a month: already triggered (July 17) and still active.
Severe drop of 30% within 3 months: 🔴 30% — up from 28% a day earlier.
Market-wide crisis within 3 months: 🟡 8% — up from 7%.

Why I nudged the 3-month severe-drop odds up by 2 points: the AI-capex worry became concrete (Alphabet punished for spending, the whole funding group down, with Microsoft/Meta/Amazon reporting next week), a broad risk-off day, two new distribution days, and Nvidia slipping back below its 50-day line. What held the increase to just +2: the chips themselves did not break — SOXX held on light volume, still +39% above its long-term line, memory and equipment rose, and Intel beat after the bell. My rule: bad news with price levels that hold is closer to noise than signal. Credit stayed calm.

💳 Credit & macro

Credit calm: the HYG/LQD ratio (risky vs safer corporate bonds) was flat; high-yield spreads 268 bp, near cycle lows. Bond yields rose (10-year 4.67%, 2-year 4.31%, as of July 22) as oil jumped on Middle-East escalation. Rate-rise odds for the July 28–29 Fed meeting have climbed toward ~36% (from ~17% mid-month; single-source, not fully cross-checked) — a live, two-way catalyst.

👀 Next 7–14 days

Intel's strong after-hours print (positive for chips) into Friday; the Fed on July 28–29; SK Hynix on July 29; and the big-tech capex updates on July 29–30 — after Alphabet, the market is hypersensitive to AI spending. Oil and the Middle East remain a rates risk. Nvidia reports August 26.

🛡️ Protecting capital (Level 3 active, day 6)

My highest internal alert only steps down with a full "all-clear" checklist (§8 of my method); today it scores 0 of 4. The fundamental engine — capex — cracked today even as the chips' price levels held. That argues for discipline on levels over prediction, and — as always — I'd suggest anyone acting on this talk to a licensed adviser.

📚 Sources

Massive/Polygon (prices, moving averages, momentum), FMP (S&P 500, VIX), Alpha Vantage (Treasury yields); FRED for high-yield spreads; company reports and market coverage (Alphabet, Tesla, Intel) via Motley Fool, Yahoo Finance, CNBC, Bloomberg, TheStreet.

Who I am. I'm Andrea Pilotti. I invest my own money in AI and semiconductor stocks, and I work with artificial intelligence for a living. I built this observatory because I needed to measure the risk in my own exposure instead of guessing at it — the agents that run it every morning are something I engineered, not a company standing behind me.

What this is for. One thing only: risk management, not returns. I publish the same measurements every trading day, with the method in the open, so that anyone can see when the picture changes and decide for themselves what, if anything, to do about it.

⚖️ What this is not. I am not a financial advisor and I am not licensed. Nothing here is a recommendation to buy, sell or hold anything, and none of it is investment advice. These are calibrated model estimates, not certainties: over short horizons noise dominates, and this sector can invalidate any estimate in a single session. Every decision is yours, ideally with professional guidance.