War, Tariffs, and Software Wreckage: The Week That Tests Everything

Look, the week of March 2 was always going to be ugly. Packed data calendar, earnings season in full swing, tariffs still hanging over everything like a fog that won't lift. Then the U.S. and Israel hit Iran on February 28 and turned "ugly" into something closer to "historically unpleasant."

Not a sternly worded letter. Actual military strikes against the world's third-largest OPEC producer.

Crude closed Friday at around $67. By Monday open, analysts expect it to gap to $90, possibly higher, as Iran has started restricting traffic through the Strait of Hormuz. For anyone keeping score at home, roughly one-fifth of all global oil and gas flows through that corridor. TD Securities says prices could move "materially above $100." Oxford Economics floated $140 in a worst-case scenario, which feels like the kind of number you throw out hoping you're wrong.

Your gas bill just became a macro event.

And this energy shock lands on a domestic data calendar that was already plenty stressful on its own. Monday brings ISM Manufacturing PMI. January came in at 52.6 first expansion in 12 months after 26 straight months of contraction, so that was nice while it lasted. Consensus for February is 52.3. A hot prices subindex reading pours gasoline on an inflation tire fire. Which, given actual oil prices, is almost too on-the-nose.

On Wednesday, we get ADP and ISM Services. Appetizers before Friday's main course: the February jobs report.

This is where things get interesting. Nonfarm payrolls consensus sits at 60,000, but estimates range from 35,000 to 125,000 because apparently nobody can agree on anything. January surprised to the upside at 130,000 versus a 55,000-70,000 consensus. Sounds great, right? Except that the 2025 benchmark revisions quietly slashed total job creation by 898,000. Turns out average monthly gains weren't 49,000. They were 15,000.

So January's big "surprise" might just be noise on top of a data graveyard. February's number carries outsized weight because, frankly, the denominator was a lie.

Friday also drops Retail Sales. Rare one-two macro punch at the end of a week already running on fumes and cortisol. Either this validates the Fed's "resilient economy, no rush to cut" narrative, or it starts cracking the foundation.

No pressure.

This week's earnings slate was designed by someone who hates sleep and has a personal vendetta against portfolio managers.

Broadcom (AVGO) Wednesday after close. Arguably the highest-stakes single name of the week. Street consensus: $19.1B revenue (up 28% YoY), EPS $1.67. The AI semiconductor business is doing the heavy lifting here: $73 billion backlog, a $7B Google TPU deal, the whole thing. Stock's pulled back to around $330 from its highs, though, and China just banned VMware, which doesn't exactly help the software segment.

Here's the thing with Broadcom: the Q1 beat won't move the stock. Broadcom beats. That's what it does. What matters is Q2 FY2026 guidance: $20.35B revenue target, 68.5% adjusted EBITDA margin. The analyst consensus price target near $433 implies about 30% upside, suggesting the market is pricing in what's next, not what's already happened. Forward confirmation or bust.

CrowdStrike (CRWD) Tuesday after close. Q4 FY2026, consensus EPS $1.10 on revenue of $1.30B. That's 22.5% YoY growth. They've beaten estimates two quarters running, including a 62.7% EPS surprise in Q3 that honestly made the sell-side look like they were throwing darts blindfolded.

D.A. Davidson's channel feedback shows the strongest sequential top-line improvement since before the July 2024 outage. Falcon Flex is reaccelerating ARR growth toward 23.5%. JPMorgan sees "healthy" results but cut the target from $582 to $472 anyway, because software multiples are compressing and nobody wants to be the last bull standing.

The real question isn't whether CrowdStrike beats. It's whether cybersecurity spending stays durable while AI eats every other software category alive. They need to prove they're on the right side of the disruption line. Not the menu.

Costco (COST) Thursday after close. Q2 FY2026, consensus EPS $4.53 on $69.2B revenue (up 8.6% YoY). They delivered 6.4% comp growth in Q1, a 20.5% surge in digital sales, and beat estimates for three straight quarters. All good. But the stock trades at 53x trailing with a $441.5B market cap, and at that valuation, disappointment doesn't get a warning shot. It gets a cliff.

Costco is the consumer bellwether. The report lands the same day as jobless claims. Thursday becomes a referendum on whether American households are handling inflation or quietly going under.

Marvell (MRVL) Thursday after close. Q4 FY2026, consensus EPS $0.79, revenue $2.2B (up 21% YoY). The AI data center connectivity business is the swing factor in Q3, with $0.76 EPS versus $0.67 consensus and 36.8% revenue growth, but shares have been all over the place because everything even tangentially related to AI is being treated like a suspect in a lineup right now.

This one matters as data point number two on AI semiconductor demand, right after Broadcom the night before. Two weak guides back-to-back, and the AI infrastructure trade is pronounced dead. Two strong ones, and it gets a second life. One of each? Chaos, which is apparently the market's preferred state of existence.

The rest of the card: Target and Best Buy on Tuesday for discretionary spending reads. Ross Stores and Burlington for the off-price channel, which tends to do well when consumers trade down. Which they are. Obviously.

Iran resets the energy equation. The strikes are the most significant military action against a major oil producer since the 2003 invasion of Iraq. Iran pumps 3.3 million barrels a day, but it's the Hormuz chokepoint that turns a regional mess into everyone's problem. Capital Economics warns of $100 oil if the disruption sustains. Whether this is a one-week spike or a regime change in energy pricing depends entirely on duration. The cascading effects on inflation, central bank policy, and consumer sentiment are none of them good. You already know that.

Software-mageddon rolls on. February was brutal. Anthropic dropped AI-powered legal automation tools, and the market responded by erasing roughly $1 trillion in software market cap over three days. Palantir down 22%. Salesforce and ServiceNow are down 25-30%. Intuit is off 34% year to date. Then Dorsey announces that Block is cutting its workforce in half because AI has made humans redundant.

It's every automation horror story from every conference keynote for the last five years, except now it's actually happening, and it's happening to the companies that thought they were the ones doing the disrupting. Morningstar says tech now holds the highest percentage of undervalued stocks. Maybe the repricing overshot. This week's earnings will test whether the market can tell the builders from the victims.

Tariffs remain stuck in limbo. The Supreme Court struck down the IEEPA tariffs on February 20. The administration pivoted to Section 122 within hours, imposing a 10% global tariff that was bumped to 15% and expires in 150 days without congressional approval. So we've got lower tariffs with an expiration date nobody trusts, sitting atop ongoing Section 232 and 301 investigations that could add more duties whenever. Tariff-sensitive sectors are one headline away from a bad day. Every single day. For five months.

CNN Fear & Greed closed February at 43. Solidly in "Fear" territory. The 2026 average is 48.67, with 38% of trading days in the fear zone and exactly 0% in extreme greed. VIX peaked around 21.5 for the year, up 42% YTD.

Now, VIX at 21 doesn't scream panic on its own. But beneath the surface, the options market is telling a very different story.

Goldman says institutions are selling U.S. equities at a four-year high clip while simultaneously loading up on downside protection at unprecedented levels. One-month S&P 500 options skew hit a four-year extreme. Puts are expensive, calls are cheap, and Goldman's trading desk says there's "no demand for S&P 500 call options on the floor."

Think about that for a second. VIX reads 21. Institutional positioning reads 35. The pros are braced for a move that headline volatility hasn't yet caught up to. They're not panicking. They're already in position.

Polymarket has an 87% chance of a gap-down open on Monday. $813K in volume behind that bet. On rate cuts, the March FOMC carries 4% odds. For 2026 overall, "2 cuts" leads at 27%, and "0 cuts" still holds 10%, which is a non-trivial number of people betting the Fed does absolutely nothing all year.

Defensive rotation tells the same story. Utilities up roughly 10% in February, best month since 2003. Staples up about 8%. Energy leads YTD at +24%, and that gap probably widens from here. Nasdaq dropped 3% in February. Money is leaving tech for what people are calling "Real Economy" assets.

The smart money is already in the bunker. The rest of us are reading newsletters about it. Which, I'll admit, is a choice.

Oil stays above $90, Hormuz escalates. Energy keeps outperforming. CPI expectations jump. Rate cuts get pushed back further. The second-order hit on gas prices, crushing consumer confidence, starts showing up in retail data within a few weeks. If it all de-escalates quickly? Risk premium vanishes overnight, energy names reverse hard, growth stocks get a relief bounce. Either outcome is plausible. That's the whole problem.

Jobs come in hot again. Another 130K+ print and the market reprices rate cuts downward. CME FedWatch June probability drops below 50%. Ricchiuto over at U.S. Bank thinks we might get zero cuts in 2026. He calls that "unfavorable for equities" and expects "a sideways correction for an extended period," which is analyst-speak for "this is going to suck for a while." A weak print under 35K flips the script entirely, recession fears, Treasury flight, and gold catches another bid.

Broadcom guides down. AI semiconductor sentiment takes a real hit. It cascades straight into Marvell's report on Thursday and breathes new life into the "AI spending has peaked" camp. An upside surprise does the reverse selective tech bounce, particularly in names that got dragged down with the garbage.

Private credit cracks widen. This one's been building quietly. Blue Owl already restricted withdrawals and liquidated $1.4B in fund assets. Market Financial Solutions in the UK collapsed entirely. Software companies back roughly 40% of PE loans, so Software-mageddon has a direct pipeline into credit markets. The default rate at 5.8% as of January. BDCs are down 11.5% on the year. BlackRock TCP Capital is sitting at record lows. One bad headline from a major BDC this week, and credit spreads blow out in ways the equity market hasn't even started to price.

Defensives vs. Growth, the rotation keeps deepening. Defensive sector market cap weight bottomed at an all-time low late last year, which mirrored dot-com era extremes. It's been ticking back up since. That's not bullish when defensives start outperforming; the broader market is usually heading into rougher waters. Passive investors are sitting on historically low exposure to utilities, healthcare, and staples. Every further leg down in tech hits those portfolios harder than people realize.

Congrats on the "diversified" index fund, by the way.

AI Infrastructure vs. AI Disruption is increasingly a two-speed market. February drew a sharp line between companies building the infrastructure and companies getting disrupted by it. This week tests whether that line holds or starts to blur. The VanEck BUZZ Sentiment Index dropped 9.1% in early February. The hype trade got humbled.

Consumer resilience gets a full workup this week. Target, Best Buy, Costco, Kroger, and Burlington full cross-section from mass-market discretionary through warehouse clubs to off-price. Costco's Q1 says the upper-middle consumer is still spending, but energy costs and tariffs are headwinds that don't just go away. Ross and Burlington's commentary on traffic patterns will tell you where things are headed into spring.

Gold pushed past $5,300 after the strikes, closing in on January's record near $5,600. JPMorgan is calling it a "structural repricing phase" with targets up to $6,000. Others think $5,500 is reachable, but gains beyond that could run into a stronger dollar. Sustained buying above $5,300 signals a real flight-to-safety bid, not just a knee-jerk reaction.

When gold is the most interesting thing in your portfolio, the market is trying to tell you something. Might want to listen.

More catalysts are packed into five sessions than any week I can remember. And the real risk isn't any single one of them; it's how they interact.

Iran rewrites energy pricing. Oil at $90+ is something markets haven't had to deal with this cycle. That slams into a data calendar trying to answer whether the labor market is strong enough to keep the Fed parked or whether January was just a statistical anomaly. Broadcom and Marvell either validate AI semiconductor demand or they don't. CrowdStrike and Okta test software durability as Software-mageddon grinds into month two. And the retail names tell us if the consumer is bending or breaking.

Under the surface, institutions are behaving like the VIX should be 35. Skew at four-year highs. Zero interest in upside. Polymarket has 87% odds on a Monday gap-down. The 6,900 level on the S&P is the line below that; we're testing mid-December lows around 6,730. Getting back above 7,000 needs a catalyst that, as far as I can tell, doesn't exist yet.

Everything feeds into everything else. Tariffs in limbo. Iran's next move is unknown. The Fed hasn't been tested by a simultaneous oil shock and labor data in this cycle. Private credit stress keeps building while the software valuations backing 40% of PE loans are actively repricing.

None of it resolves this week. But every thread produces a data point that either loosens or tightens the pressure. All of them, hitting in the same five-day window.

Cross-asset volatility isn't the tail risk here. It's the base case.

Enjoy your week.