I’ve sat through enough “chips got smoked” weeks over fifteen years to know the drill by now: something scares the semiconductor trade, headlines get dramatic, and half of Twitter starts asking if the AI bubble finally popped. This week had all of that, plus an oil spike, a surprise Chinese AI model announcement, and Warren Buffett publicly calling the market a casino on live television. Genuinely a lot packed into five trading days. Let’s walk through what actually happened, what it means, and what’s sitting on the calendar for next week that could either calm things down or make it worse.
Table of Contents
- The Week at a Glance
- Day-by-Day Breakdown
- What Actually Drove the Sell-Off
- Winners and Losers of the Week
- Next Week’s Economic Calendar
- Next Week’s Earnings to Watch
- A Realistic Portfolio Scenario
- Veteran’s Honest Take
1. The Week at a Glance
Here’s the scoreboard first, because I know some of you scroll straight to the numbers before reading a word of my commentary, and honestly, that’s a fair way to read a market recap.
| Index | Friday Close | Weekly Change |
|---|---|---|
| S&P 500 | 7,457.69 | -1.6% |
| Nasdaq Composite | 25,520.24 | -2.9% |
| Dow Jones Industrial Average | 52,146.42 | -0.9% |
| PHLX Semiconductor Index (SOX) | Entered bear market territory | -13% over the past month |
Worth sitting with that semiconductor number for a second. The SOX index is still up something like 63% year-to-date even after this pullback, which tells you the sector rally that started the year was so strong that even a genuinely rough month barely dents the annual scoreboard. Context matters more than headlines here.
2. Day-by-Day Breakdown
I always find the day-by-day version more useful than the weekly summary alone, because it shows you the market didn’t fall in a straight line – it whipsawed on different news each day, which is its own lesson about not overreacting to any single session.
| Day | S&P 500 | What Happened |
|---|---|---|
| Monday | -0.8% | Rising oil prices and escalating US-Iran tensions pressured sentiment across the board. |
| Tuesday | +0.38% | June CPI came in cooler than expected. IBM tumbled roughly 25% after warning of soft software and infrastructure demand. |
| Wednesday | +0.38% | BlackRock beat estimates and jumped over 5%. Warren Buffett told CNBC the market is increasingly driven by speculation rather than long-term investing. |
| Thursday | -1.01% | Chinese AI startup Moonshot released a surprisingly capable open model, spooking US AI-related names. Chip stocks led the decline. |
| Friday | -1.01% | Semiconductor rout deepened, oil spiked further on renewed US-Iran strikes, and the SOX index confirmed a bear market. |
3. What Actually Drove the Sell-Off
Three separate stories got tangled together this week, and I think separating them actually matters if you’re trying to decide what to do with your own portfolio rather than just reacting to red numbers on a screen.
Geopolitical risk premium. The US and Iran traded strikes multiple times this week, and oil prices moved accordingly. That’s a genuine macro risk that touches inflation expectations, energy costs, and consumer spending power – not a sector-specific story, a broad one.
The Moonshot AI shock. A Chinese AI startup released an open model advanced enough to rattle US tech investors’ confidence in the durability of the American AI lead. Whether that fear is proportionate or overblown is genuinely debatable, but the market’s initial reaction was to sell first and debate the nuance later, which is fairly typical behavior for a surprise headline landing mid-week.
Semiconductor premium unwind. The SOX index had rallied so hard this year – 63% year-to-date at its peak – that a pullback of some size was arguably overdue on valuation grounds alone, independent of any single news catalyst. When a sector runs that hot, it doesn’t take much bad news to trigger profit-taking from investors who were already sitting on enormous year-to-date gains.
4. Winners and Losers of the Week
| Stock | Move | Reason |
|---|---|---|
| IBM | -25% | Warned Q2 profits would miss on soft software and infrastructure demand. |
| BlackRock | +5%+ | Beat Q2 estimates with EPS of $13.91 on $7.08 billion revenue. |
| Travelers Companies | +9% | Significant Q2 earnings beat, dragging insurance peers Progressive and Allstate higher too. |
| UnitedHealth Group | Beat estimates | Q2 EPS of $6.38 versus $4.94 expected, though revenue came in slightly light. |
| Netflix | -11%+ intraday | Sharp premarket decline amid the broader tech and growth-name sell-off. |
| Nvidia, Caterpillar, Goldman Sachs | Notable Dow/index drags | Semiconductor and cyclical exposure weighed on major indexes Friday. |
5. Next Week’s Economic Calendar
The single biggest item on next week’s docket isn’t a data release at all – it’s Fed Chair Kevin Warsh delivering his first congressional testimony since being sworn in back in May. Markets have been pricing in the possibility of a rate move as soon as the September meeting, and how Warsh frames the Fed’s current thinking could move rate expectations meaningfully in either direction.
| Day | Event |
|---|---|
| Monday, July 20 | Leading Economic Indicators (June) |
| Tuesday-Wednesday | Fed Chair Kevin Warsh’s first congressional testimony on monetary policy |
| Thursday, July 23 | Initial Jobless Claims (week of 7/18) |
| Friday, July 24 | Preliminary July Consumer Sentiment reading |
6. Next Week’s Earnings to Watch
This is a genuinely loaded earnings week, and it lands right in the middle of the semiconductor and AI-capex anxiety that drove this week’s sell-off – which means these reports carry more weight than they normally would.
| Day | Notable Reporters |
|---|---|
| Monday | Halliburton, 3M, General Motors, Danaher, Northrop Grumman, D.R. Horton |
| Tuesday | ServiceNow, Texas Instruments, Alphabet, Tesla, AT&T, Philip Morris, CME Group |
| Wednesday | Intel, Union Pacific, United Rentals, Deckers Outdoor, Digital Realty Trust |
Alphabet and Tesla are the two everyone’s watching most closely. Alphabet is expected to show accelerating cloud growth and, if the analyst chatter is right, could raise next year’s capex guidance into the $275-325 billion range – a number that will either reassure the market that AI infrastructure spend still has a return story behind it, or add fuel to the exact anxiety that hit chip stocks this week. Intel and Texas Instruments matter for a different reason: as pure semiconductor plays, their guidance will be read as a referendum on whether this week’s SOX bear-market move was an overreaction or the start of something longer.
7. A Realistic Portfolio Scenario
Say you’re holding a fairly typical retail growth-tilted portfolio – some Nvidia, some Microsoft, maybe an SMH semiconductor ETF position you added earlier this year when the sector was riding high. This week probably hurt. A portfolio with 20-30% tech and semiconductor exposure likely underperformed the S&P’s -1.6% weekly move by a meaningful margin, given the Nasdaq’s -2.9% week and the SOX entering a bear market outright.
Here’s the question I’d actually be asking myself in that seat, rather than just watching the account balance: has anything about the underlying demand for AI infrastructure actually changed this week, or did a scary headline and an overheated sector meet at the same time? The Moonshot AI news is a real competitive data point worth digesting, but it’s not the same thing as a hyperscaler cutting capex guidance or an earnings miss showing weaker enterprise AI adoption. Until next week’s Alphabet, Tesla, and Intel reports actually show demand deterioration in the numbers, I’d treat this week’s move as a sentiment and positioning reset rather than a fundamental break – while acknowledging plainly that I could be wrong, and that’s exactly what next week’s earnings will tell us either way.
8. Veteran’s Honest Take
Fifteen years of doing this has taught me that weeks like this one feel much scarier while you’re living through them than they usually look in hindsight six months later – though not always, and that caveat matters, because sometimes the scary week actually was the start of something. The honest answer is you don’t get to know which kind of week this is until later, and anyone telling you with total confidence right now which one it is should probably be treated with some skepticism.
What I’d actually watch next week isn’t the headline index moves, it’s the specific language coming out of Alphabet and Tesla’s earnings calls about capex plans, and how directly Fed Chair Warsh addresses the rate-hike speculation the market’s currently pricing in for September. Those two threads – AI infrastructure demand and monetary policy direction – are the actual load-bearing pillars under this market right now, far more than any single day’s semiconductor headline. Everything else this week, including the Moonshot AI scare and the Iran-driven oil spike, is real and worth tracking, but it’s noise layered on top of those two bigger questions rather than a replacement for them.
If you’re sitting on a portfolio that got knocked around this week, my honest advice is the boring one: don’t make a big allocation decision off five trading days of geopolitically-driven volatility. Wait for next week’s earnings to actually tell you something concrete about demand, and let Warsh’s testimony tell you something concrete about rates. Then reassess with real information instead of a red weekly chart.
