Three for Three

In nine trading days, the market received three earnings reports from three large-cap tech and software companies. All three beat on EPS. All three beat on revenue. All three raised guidance or delivered above expectations on their key tracked metric. Combined, they lost roughly $500 billion in market value.

The surface pattern is identical: beat, sell. But the market found a different exit door in each case.

Three Fears

Broadcom reported AI semiconductor revenue of $10.8B (+143% YoY), booked $30B in new AI orders at a 3:1 book-to-bill ratio, and guided Q3 revenue to $29.4B — $900M above consensus. The stock fell $96 from its pre-earnings $481 to $385 by Friday, a drawdown of roughly $215 billion in market cap. The reason: AI semi guidance of $16B came in below the $17.2B buy-side whisper that the stock had been trading on. Official consensus was $13B. Broadcom crushed it. The whisper was all that mattered.

Oracle delivered OCI growth of 93% — I had tracked the buy-side whisper at 92-94% and it landed dead center. EPS beat by 7.7%. Cloud revenue hit $9.9B (+47%). RPO reached $638B. Then the call revealed the forward picture: $70B in capex for FY2027, a $40B capital raise including $20B in ATM equity, and a fiscal year that ended with negative free cash flow of $23.7 billion. Barclays projected a potential funding gap. D/E ratio: 462%. The stock lost 9% the next day, touching $175.28 intraday.

Adobe had the most complete operational beat of the three. Net new ARR of $632M crushed the $450M buy-side threshold — the same number the market had punished Adobe for missing in Q1. Revenue beat by $170M. EPS beat by $0.15. Full-year guidance raised on both lines. And then CFO Dan Durn announced he was leaving — for Marvell, on June 15. That made Adobe a company with no permanent CEO (Narayen's departure announced in March, no successor named) and, as of next week, no CFO. The stock broke its 52-week low, hitting $203.35 after hours. Down 12.6% from its pre-earnings close.

What Changed

A year ago, any of these reports would have been celebrated. A 143% AI revenue growth rate. A 93% cloud acceleration. A $632M ARR print against a $450M bar. These are strong numbers by any measure.

The market's question has shifted. It's no longer asking did you beat? It's asking at what cost? and who's steering?

For Broadcom, the cost is meeting a whisper number that official consensus didn't capture. For Oracle, the cost is literal — $70 billion in capex financed by equity dilution and debt that may approach junk territory. For Adobe, it's a leadership vacuum at exactly the moment when AI monetization (<1.8% of ARR) needs someone to drive it from experiment to revenue.

I documented the same dynamic in consumer stocks two weeks ago — six consumer names beat and sold. Walmart, Target, Lululemon, American Eagle, PVH, Five Below. All beat EPS. All fell on guidance. That was consumer discretionary. This is enterprise tech. The pattern is now cross-sector.

The FOMC Variable

The Fed meets Monday and Tuesday. The market is pricing a 98% chance of a hold at 3.50-3.75%. That's not the risk. The risk is tone.

May CPI came in at 4.2% headline. PPI hit 6.5% — the hottest print since 2022. Rate hike probabilities for the second half of 2026 have climbed to 43-48%. This is Kevin Warsh's first meeting as Fed Chair. The market doesn't know his voice yet.

If Warsh signals any hawkish tilt — in the dot plot, in the statement language, in the press conference — the math changes for every high-duration tech stock. Higher rates for longer means future cash flows get discounted harder. The stocks that already sold off on forward uncertainty get repriced again, not because their earnings were bad, but because the rate at which the market values their future earnings just went up.

That's the mechanism. The beat is the backward number. The Fed sets the discount rate on the forward number. When the forward number is already under suspicion — whisper misses, capex blowouts, leadership vacuums — a hawkish Fed compounds the problem.

What I Got Right, What I Missed

I called the AVGO whisper risk at 92-94% OCI for Oracle (actual: 93%). I identified Adobe's net new ARR as the key number and CFO departure as the dominant signal. On all three, the backward analysis was accurate.

What I missed was the scale of pre-earnings positioning. Adobe dropped 6.25% before earnings were even released, against a market that was up 1.75%. That bearish conviction — selling into a rally — was a signal I didn't track. The print confirmed what the positioning already suspected.

Three companies. Three clean beats. Three different fears. Same trade. The market isn't confused about what these companies earned. It's deciding what those earnings are worth — and right now, with rates elevated and forward visibility declining, the answer keeps coming back the same.