Date: 2026-03-03 | Analyst: StockNovice (Joe) | Quarter: Q3 FY26 (Jan 2026)
Credo started March with what I consider an absolute monster of a quarter. And I don't throw that word around lightly.
$407M revenue. +201% YoY. +52% QoQ. Beat their own guide by $67M — that's not a rounding error, that's a whole extra quarter for the company they were two years ago. Fifth consecutive beat-and-raise. Every margin metric at record levels. FCF inflecting to $140M from $39M last quarter.
Here's the thing — I had four red flags pinned to my board going into this report. Every single one resolved in Credo's favor. That... doesn't happen very often.
When the prelim dropped on February 9th, I flagged these open items:
1. The gross margin mystery. There was this weird 45% "gross margin" guidance floating around that didn't reconcile with the 67.5% GAAP GM they'd been printing. Was the company actually guiding for a 22-point margin collapse?
Verdict: Nope. It was a mislabeled operating margin figure, just like I suspected. Actual Q3 GAAP GM: 68.5%. Non-GAAP GM: 68.6%. Beat the guided range of 63.8-65.8% by 270 basis points on the high end. This wasn't close. The margin concern is dead.
2. Customer concentration direction. Top-3 customers at 82% in Q2, with the 4th hyperscaler just crossing 10%. Were they diversifying?
Verdict: Mixed. Top-1 dropped from 42% to 39% — that's the right direction. But Top-2 surged from 24% to 32%, and Top-3 held at 17%. Combined top-3: 88%. That's worse than Q2's 82%. The 4th hyperscaler actually fell back below 10%. I'll talk more about this below because it's the one wart on an otherwise flawless report.
3. Q4 formal guidance. Would they just say "mid-single-digit sequential" or give real numbers?
Verdict: Real numbers: $425-435M. GM 64-66%. Based on the beat pattern, I'd pencil in something more like $460-490M actual and 67%+ GM. But let's not get ahead of ourselves.
4. Competitive signals. Any signs of Broadcom or Marvell gaining traction in AECs?
Verdict: Not a whisper. Brennan talked about the fifth hyperscaler starting to ramp. The patent enforcement activity (Amphenol settled, Molex and TE Connectivity cases ongoing) actually suggests the competition is trying to copy, not innovate around. That's a moat signal.
| Quarter | Revenue | YoY% | QoQ% | GM% [NG] | OpM% [NG] | Net Margin [NG] | FCF | FCF% |
|---|---|---|---|---|---|---|---|---|
| Q3 FY25 (Jan-25) | $135M | +154% | +88% | 63.9% | 31.4% | 33.6% | — | — |
| Q4 FY25 (Apr-25) | $170M | +180% | +26% | 67.4% | 36.8% | 38.4% | $54M | 31.9% |
| Q1 FY26 (Jul-25) | $223M | +274% | +31% | 67.5% | 43.1% | 44.1% | $51M | 23.0% |
| Q2 FY26 (Oct-25) | $268M | +272% | +20% | 67.7% | 46.3% | 47.7% | $39M | 14.4% |
| Q3 FY26 (Jan-26) | $407M | +201% | +52% | 68.6% | 49.6% | 51.3% | $140M | 34.3% |
I mean... look at that operating leverage. NG operating margin went from 31% to 50% in five quarters. Net margin over 51%. For a semiconductor company. This is the kind of profile I usually only see in software companies.
| Quarter | Guide Mid | Actual | Beat | Beat % |
|---|---|---|---|---|
| Q3 FY25 | $120M | $135M | +$15M | +12.5% |
| Q4 FY25 | $150M | $170M | +$20M | +13.3% |
| Q1 FY26 | $190M | $223M | +$33M | +17.4% |
| Q2 FY26 | $235M | $268M | +$33M | +14.0% |
| Q3 FY26 | $340M | $407M | +$67M | +19.7% |
Five in a row. And the dollar beats are getting bigger, not smaller. That's the single best leading indicator I know for management quality — they're not just sandbagging by a fixed percentage, the business keeps surprising them to the upside.
Original FY26 guide (set in June 2025): >800M.Actual : trackingto 1.33B. They underguided by $530M. Sixty-six percent. Shows what I know about "conservative guidance" — I didn't realize there was a level beyond conservative.
When I first looked at CRDO on the prelim, I put it firmly in the "Is" column. The full results only strengthen that call.
What "Is":
What "Could Be" (option value on top):
Here's how I'm thinking about it — the "Is" alone justifies a starter position. The "Could Be" makes the ceiling meaningfully higher. If ZF Optics ramps to $200-400M in FY27 (per the revenue build from Fleming's commentary), that's a genuine second pillar that transforms the bull case from "great AEC company" to "connectivity platform."
I always read the call. The numbers are the headline; the call is the context. A few things stood out:
Brennan on the NVIDIA optical narrative: "There has been a bit of a signal-to-noise ratio issue in the market, and the noise right now is dominating the signal. It is not an either-or type of situation."
I thought this was well said. For those who missed it — NVIDIA announced $2B partnerships with Lumentum and Coherent (optical component companies) on the same day as CRDO's earnings. Market read: NVIDIA is picking optical winners and CRDO isn't one. My read: Lumentum and Coherent make lasers and silicon photonics. CRDO makes copper Active Electrical Cables and system-level optical solutions with reliability software. Different products, different reach distances, different value propositions. Brennan was right to address it directly without being defensive.
Brennan on copper vs optical: "They [NVIDIA] have been really outspoken that where you can use copper, you will use copper. The reasons are very basic — reliability, number one; power efficiency, number two; and ultimately, total cost of ownership, number three."
Fleming on FY27: "Just north of $1,300,000,000 [FY26]. Fifty percent growth gets you to nearly $2,000,000,000 for next year."
That's... pretty direct for a CFO. FY27 at ~$2B. Not "aspirational." Not "in the right conditions." Just — here's the math, you can check it.
Fleming on what drove the Q3 upside: "Our top three customers all grew sequentially from Q2 to Q3."
No one-offs. No pull-forwards. Broad-based demand across their biggest customers. That's the best possible answer to "what drove the beat?"
I need to be honest about this because it's the one thing that keeps me from banging the table:
| Quarter | Top 1 | Top 2 | Top 3 | Combined | 10%+ Customers |
|---|---|---|---|---|---|
| Q4 FY25 | 61% | 12% | 11% | 84% | 3 |
| Q1 FY26 | 35% | 33% | 20% | 88% | 3 |
| Q2 FY26 | 42% | 24% | 16% | 82% | 4 |
| Q3 FY26 | 39% | 32% | 17% | 88% | 3 |
88% from three customers. The 4th hyperscaler fell back below 10% — not because they ordered less, but because the top customer grew so fast it changed the denominator. The individual trajectories are healthy. But the aggregate concentration is stubborn and it went backward from Q2.
Here's why this matters: if Customer #1 at 39% pauses a build cycle for even one quarter, that's ~$160M of revenue at risk. One hyperscaler decision can swing the trajectory. That's not a thesis killer — hyperscalers don't tend to pause AI infrastructure buildouts right now — but it's a risk I have to size around.
Look, I've seen the market do this before — latch onto a narrative and ignore what's actually happening in the business. The NVIDIA-Lumentum (2B)andNVIDIA − Coherent(2B) deals are about optical components: lasers, silicon photonics, transceiver technology. CRDO's core business is copper AECs for short-reach (<7 meter) data center connections. These are different products for different applications.
Could the industry eventually shift more toward optical at shorter reaches? Sure, someday. Physics makes copper better at short distances today (reliability, power, cost). Brennan's right that it's not an either/or. And CRDO has their own optical product (ZF Optics) that competes on reliability and telemetry, not on laser technology.
I think the selloff on this narrative is overblown. But I also understand why the market is nervous — when NVIDIA writes $4B in checks to two optical companies on the same day you report, the optics (no pun intended) aren't great. It creates a perception drag even if it's not a fundamental problem.
If NVIDIA ever announces a partnership with an AEC competitor, or if optical technology improves enough to displace copper at short reach — that changes things. Neither has happened.
| Metric | Value |
|---|---|
| Market Cap | ~$20.3B |
| Run-rate P/S (Q3 x4) | 11.7x |
| Run-rate P/E (GAAP) | 32x |
| Run-rate P/E (NG) | 24x |
| EV/FCF (run-rate) | 34x |
| FY27 P/S (~$2B) | ~10x |
At 200%+ growth, 11.7x P/S looks cheap. But that growth rate isn't sustainable — even management says FY27 is >50% YoY. So the fairer comparison is 50-60% growth at 11.7x P/S, which is... still pretty reasonable actually. Not screaming cheap, but I've paid a lot more for a lot less growth.
32x GAAP P/E on a company growing 200% YoY with 68% gross margins and $1.3B cash. I've seen worse risk/reward setups.
Establishing → Strengthening
Here's my reasoning: every open question from the prelim resolved favorably. The GM mystery was conclusively solved. The beat-and-raise pattern continued and widened. FCF inflected massively. The product pipeline expanded with customer-driven pull-ins (ZF Optics). Management delivered specific FY27 commentary (~$2B) with the kind of directness that earns credibility.
The one thing holding me back from full conviction is the customer concentration. 88% from three customers is real risk that I can't wish away with good vibes. But everything else — the growth, the margins, the execution, the management, the competitive position — is elite.
Sizing: Starter. This is a full-position company for me. The "Is" evidence is overwhelming. The "Could Be" upside from ZF Optics and the broader platform makes the ceiling even higher.
If I were building from scratch today, I'd go 8-10% and look to add on any pullback toward $100 (or after Q4 confirms). The NVIDIA optical narrative might create that opportunity.
The one thing I wouldn't do is wait. The "one more quarter" instinct can work both ways — sometimes you wait for confirmation and the stock runs away from you. With five consecutive beats, 200% growth, 34% FCF margins, and the stock selling off on an optical narrative that doesn't touch the core business... I think the evidence is here.
CRDO just put up 407 points in a quarter where the oddsmakers had them at 340. Five games in a row they've beaten the spread, and the margin of victory keeps growing. The defense (margins) is getting better, not worse. They've got a star player (AECs) in their prime and a first-round draft pick (ZF Optics) warming up on the bench.
The one concern? They're relying on three teammates for 88% of their scoring. If one goes cold, the whole team feels it. But right now? Right now they're playing their best basketball of the season and the market is selling because the other team signed some free agents who play a different position.
I'll take that mismatch all day.
As usual, thanks for reading.