Palo Alto Networks did everything right and still got punished for it. Beat on earnings — $1.03 per share against the $0.94 consensus. Beat on revenue — $2.59 billion, edging past expectations. And then opened its mouth about next quarter and watched $25 billion in market cap evaporate overnight.
PANW dropped roughly 9% on Wednesday after management guided fiscal Q3 EPS to $0.78–$0.80, against a Street expectation of $0.92. Full-year EPS outlook revised down to $3.65–$3.70 from $3.80–$3.90. The culprit, more or less openly admitted: the $25 billion CyberArk acquisition is consuming the margin structure. Integration is expensive, platformization is capital-intensive, and the market, having been burned a dozen times this cycle by M&A promises, didn't wait around for the explanation.
This is the story of 2026 in a single earnings card. Execution is no longer enough. The question being asked of every major technology company right now isn't "did you grow?" It's "what did growth cost you, and when do we stop paying for it?" The era of giving management teams credit for vision without demanding near-term proof has quietly ended.
Meanwhile, AMD lost almost 4% on the same day Nvidia climbed 2%. Same sector. Same macro backdrop. Completely different trajectories. The Meta–Nvidia infrastructure deal — millions of chips for a new data center — is doing what all landmark supply agreements do: it validates the vendor and quietly sentences the runner-up. Amazon and Micron caught the draft and both rose sharply as major asset managers announced increased positions. The rotation inside tech isn't random noise. It's a genuine repricing of who owns the actual infrastructure of the AI buildout versus who benefits theoretically from it.
CrowdStrike is probably watching all of this with one hand on the phone. While Palo Alto digests CyberArk and absorbs Koi, its XDR acquisition, the window for a competitor to grab enterprise security conversations is widening. Not permanently — Palo Alto's platformization strategy is architecturally sound — but there's an 18-month window of margin distraction that sharp competitors don't ignore.
Brent crude sat at $67.50 on Tuesday and barely moved. That restraint is geopolitically remarkable given what's actually happening: the United States and Iran are sitting in Switzerland discussing nuclear guiding principles, with a second carrier group deployed to the Middle East and explicit threats of military action if talks collapse. Under almost any historical analog, that combination of facts would have oil at $80 and climbing. Instead, the market is treating the Strait of Hormuz risk as manageable and the diplomatic track as credible enough to short the fear premium.
Iran's foreign minister went out of his way to remind journalists that reaching an understanding on guiding principles is not a comprehensive agreement. Technically accurate. The market already knew that. What it heard instead was: nobody is bombing anything this week. Brent dropped 2% on the day the "guiding principles" headline crossed. Energy markets now pricing a deal not as a question of if but of when, which brings Iranian barrels — 3 million per day or thereabouts — back into the theoretical supply picture. OPEC+ is watching. They always are.
Japan reported Q4 GDP growth at an annualized 0.2%, against a forecast of 1.6%. The quarter before was a 2.6% contraction. So: contraction, then near-rounding-error growth. The Nikkei extended losses. The Bank of Japan, which hiked into this weakness last year, is now sitting on a decision that has no clean answer. Inflation is sticky. The yen is fragile. Domestic demand is depressed by cost-of-living pressure. The new prime minister has promised stimulus. The central bank has promised discipline. Those two commitments will eventually conflict, and when they do, the yen trade will move faster than anyone's model suggests.
Asia was also running at partial volume this week — Lunar New Year holidays had China, Hong Kong, Taiwan, South Korea, and Singapore either closed or muted. That thinned liquidity across the board and gave price moves more amplitude than the underlying news probably warranted. Keep that context when reading the day's charts.
The S&P 500 closed February 18 at 6,881. Forward PE sits at 22.2 against a 10-year average of 18.8. The Federal Reserve flagged this valuation in September. The Financial Stability Report flagged it again in November. And here we are, still 22x, still grinding. The CBO has formally stated that current tariffs will reduce real GDP from what it otherwise would have been. That's not a think-tank opinion — that's the official nonpartisan scorekeeper saying the tax on imported goods is a tax on growth.
The bull case requires AI productivity gains to offset that drag. Maybe they will. The bear case requires nothing — just the arithmetic of 22x forward earnings in a 4% rate environment with slowing GDP and a central bank that just admitted it doesn't know which direction it moves next. Historically, valuations this stretched have only survived two other periods in the last four decades: the dot-com peak and the COVID-era liquidity flood. Both eventually cleared.
None of that means sell everything tomorrow. It means the margin for error is thin, and thin margins for error are precisely when guidance misses like PANW's $0.14 EPS shortfall cost 9% of market cap in a single session.
PCE data lands Friday alongside the Q4 GDP advance estimate — expected at 3.0%, down from Q3's 4.4%. Walmart reports Thursday morning and will tell us something real about whether the American consumer is still spending through tariff-induced price pressure or beginning to pull back. Deere reports too, which tends to be a quiet but revealing read on industrial and agricultural capital spending. Nvidia earnings are next week. Between now and then, the Fed minutes from Wednesday will continue being parsed for that one sentence about rate hikes — because once language like that is in the public record, it doesn't go away.
Gold drifted back below $5,000. The dollar held firm. Bitcoin sat near $67,000 and generated no particular headlines, which is its own kind of news.
The guidance graveyard keeps filling up. Bring flowers.
