Credo Technology grew revenue 114.7% to $479 million last quarter, more than doubled adjusted EPS to $1.20, beat both lines, and guided the current quarter above expectations. The stock fell 29.5% on the week.
That is the cleanest lesson of this earnings season, and it fits in one sentence: beating is now table stakes, and the only currency the market still accepts is proof of durable, margin-accretive monetization.
Beating is now table stakes, and the only currency the market still accepts is proof of durable, margin-accretive monetization.
The market found one blemish and one sin. The blemish was margins. Credo's GAAP gross margin slipped to 64.5%, down from 68% the prior quarter and 67.4% a year ago — and guidance calls for it to fall further. The trend is pointed the wrong way.
The sin was the price. Heading into the print, Credo carried a roughly $45 billion market cap against a $2 billion revenue run rate — about 18 times next year's sales. At that multiple, a triple-digit grower has to be flawless, and it wasn't.
Even the post-crash "cheap" story is a mirage on closer inspection. Credo trades at around 26 times this year's adjusted earnings, which looks reasonable — until you remember those estimates strip out stock-based comp, which just spiked. GAAP net income last quarter was barely half of non-GAAP income, so the true forward multiple sits closer to 50 than 26.
Hold that number against the Microsoft tape and the shift becomes impossible to miss.
Last week, the open question in this space was whether ARR and pipeline conversion could confirm durable AI demand or whether the whole edifice was running on capex announcements and vibes. This week delivered the first verdict, and it's a sorting one.
Microsoft reportedly rose 2.68% on Thursday. Not for a blowout, not for a new product — but for agreeing to tell investors something it had withheld for years: Azure's actual quarterly revenue.
The number was $29.4 billion, up 42% year over year, reportedly second only to AWS's $42.2 billion, per Fool.com. Management guided Azure to 44%–45% constant-currency growth this quarter, driven by AI usage.
Read that plainly. Transparency itself was the catalyst.
Set the two side by side. Credo delivered spectacular growth and got punished for one opaque margin line. Microsoft delivered a growth rate it had long implied and got rewarded simply for putting an auditable dollar figure next to it. The reward has moved from ambition to disclosure.
The winners this week all shared one trait: they offered something the market could verify.
Snowflake soared on what was reportedly an AI-fueled forecast, according to Glideslope.
Zscaler jumped after beating on both lines: $1.19 adjusted EPS against $1.09 expected, $898 million in revenue against $877 million, and annual recurring revenue of $3.77 billion, up 25%. Then it guided the full year above consensus — $3.91 billion to $3.94 billion in revenue, and adjusted EPS of $4.86 to $4.90 against a $4.60 estimate.
Note what carried Zscaler: recurring revenue and a raised number. Not a claim about the future of AI security. The stock moved on the ARR and the guide, the two things you can put in a model.
That matters because Zscaler had been the poster child for narrative failing to convert — shares were down 20% this year, having logged their worst day ever last quarter when management got cautious after two sales leaders left. This week the ARR and guidance reportedly drove the move, per CNBC, and the stock followed.
Then there's the counterexample that proves the rule from the other side: Palo Alto Networks.
Palo Alto printed a genuinely clean quarter. Revenue up 34% to $3.41 billion, adjusted EPS of $1.02 ahead of the Street, remaining performance obligation up 34% to $21.2 billion.
And its most monetizable metric was the strongest of all: next-generation security ARR grew 63% year over year, faster than the prior quarter, with roughly 220 net-new platformizations, up 44% from a year ago and double the prior quarter.
That is exactly the auditable, margin-relevant, recurring-revenue proof the market claims to want. And the stock still faded after the close.
The reconciliation is the tax on expectations. Palo Alto had already re-rated; the crowd was reportedly long and the ARR beat priced in, per Capwolf. When the hard work of rerating is done, a clean beat isn't a catalyst — it's an exit for someone taking profits. High expectations are themselves a headwind, the same force that turned Credo's beat into a bloodbath.
So the mechanism has two blades. If you're cheap and transparent, like Microsoft, disclosure alone can move you up. If you're expensive and priced for perfection, like Credo or Palo Alto, even excellent numbers get sold — because there's no room left between the price and the proof.
Lululemon sits outside the AI frame but confirms the underlying logic. Comps fell 9%, the worst in modern company history, revenue missed at $2.42 billion, and guidance called for a third-quarter revenue decline of 10%–11%. The stock dropped 18% after hours.
The tell was the margin math: strip out $134.5 million in tariff refunds and gross margin fell 360 basis points to 54.9%, with adjusted operating margin collapsing from 20.7% to 13.2%. Even Lululemon's EPS beat — $2.06 against a $1.79 consensus — couldn't override a story of reportedly evaporating profitability. Same rule, different sector: the market is pricing demonstrated cash generation, not headline beats.
Now widen the lens, because two seemingly opposite calls this week are actually the same signal.
At the large-cap end, Jim Cramer declared "the revenge of the Magnificent Seven" and said it's time to buy — arguing Amazon, Alphabet, Meta, Microsoft and Tesla have fallen behind the S&P's 13% year-to-date gain and gotten "real cheap." His thesis: years of data-center spending have weighed on balance sheets, and that infrastructure is about to start generating returns.
At the venture end, the warning is reportedly a coming "shakeout," per a CNBC X post: "The next six to twelve months is the window when capital must shift from betting on AI expansion to demanding cash flow from it."
Those are the same sentence spoken from two ends of the market. Cramer says the Mag 7 are cheap precisely because the market stopped paying for the buildout and hasn't yet paid for the return — a return he expects to show up as cash flow. The VC warning says private capital is about to demand exactly that proof from startups still valued on buildout and growth.
Both are describing a single transition: from narrative pricing to cash-flow pricing. The buildout phase rewarded ambition and capex announcements. The phase now beginning rewards conversion — ARR that turns into margin, spend that turns into free cash flow, growth that survives a look at the GAAP line.
This is why Credo's margin wobble mattered more than its 114.7% growth, why Microsoft's disclosure mattered more than its rate, why Palo Alto's 63% NGS ARR still wasn't enough to clear a priced-for-perfection bar.
The proof, at this stage, is being demanded quarter by quarter. And the next two prints — the ones that show whether ARR and pipeline actually convert into durable, margin-accretive cash — will decide who survives the sort, not the next capex headline.
Credo may still be a buy if it holds triple-digit growth; the drawdown could look like an entry a year from now. But that's the point. The market is no longer willing to assume it. It wants to be shown.