Credo Technology Group Holding (CRDO) Q3 FY26 earnings call: guidance and Q&A

How does the ZF Optics ramp compare to AECs, and will customer engagement be similar in scale and pattern?

Bill Brennan stated that the ZF Optics ramp is ahead of schedule, with production units already shipped and a significant ramp expected in Q1 FY27. Customer activity is very similar to AECs, with strong interest from hyperscalers and Neoclouds, driven by the product's superior reliability.

“moving from providing samples to going right into qualification with the customers, what we're seeing. And so that's why we highlighted the fact that although last quarter, we signaled that the ramp would occur in second half fiscal '27, we feel confident now saying that, that ramp is going to start in first quarter, noting that we've already shipped production units.”

asked by Thomas O'Malley · answered by William Brennan

Can you elaborate on AEC use cases for FY27/FY28 and how they compare to CPO, given market noise around CPO?

Bill Brennan reiterated that AECs are compelling for intra-rack and rack-to-rack connectivity (up to 7 meters) due to reliability and power efficiency. He clarified that CPO is not an either/or situation but part of a heterogeneous mix, and Credo's focus is on delivering bulletproof reliability for all its optical solutions, including ALCs, which offer copper-like reliability for longer reaches.

“It's about deploying the right technology at the right reach and the right power head below. And we see the industry is evolving even more so to a heterogeneous mix of short-reach copper, pluggable optics, near package optics and eventually CPO.”

asked by Tore Svanberg · answered by William Brennan

Will non-AEC offerings like PCIe solutions and optical products contribute materially to the 50%+ growth in FY27, or is that more of an FY28 story?

Bill Brennan confirmed that FY27 will see a different composition of growth, with Zero Flat Optics ramping alongside continued growth in AECs and ICs. He specified that FY28 will layer in Active LED Cables (ALCs) and the first OmniConnect gearbox, indicating a phased contribution from new product lines.

“I think that it's fair to say that we'll see a different composition between copper and optical in fiscal '27 specifically as ZF Optics round. That's -- with that, I'll say that we do expect growth in AEC. We expect growth in ICs and then the new wave of growth will come with Zero Flat optics.”

asked by Joseph Cardoso · answered by William Brennan

What drove the Q3 revenue upside? And how do AECs compare to optical solutions, especially given NVIDIA's investments in optical peers?

Dan Fleming stated that Q3 upside was driven by sequential strength across all hyperscale customers, with the top three customers growing sequentially and the largest customer contributing 39% of revenue. Bill Brennan reiterated NVIDIA's preference for copper where possible, emphasizing AEC's advantages in reliability, power, and total cost of ownership over laser-based optical modules. He stressed Credo's focus on bulletproof reliability for its optical products.

“Our top 3 customers all grew sequentially from Q2 to Q3. So that really drove that growth. And our largest 3 customers in Q3 were also our largest in Q2, as you would expect but in a different order. Let me just talk briefly about our largest customer. They were 39% of revenue, and they were also the same customer that was our largest customer in Q1.”

asked by Vivek Arya · answered by Daniel Fleming

Is the fourth hyperscaler still expected to be >10% of FY26 revenue? And why is OpEx guided flat after a significant Q3 increase, despite new products and an acquisition?

Dan Fleming indicated that the fourth hyperscaler might not reach >10% of revenue for the full FY26 due to the significant upside from other customers. He explained that Q3 OpEx saw a large step-up, and the Q4 guidance reflects a balance between project-related spend and hiring, which happened to offset each other.

“So that may make the math, while they're still in line with our expectations from 90 days ago, they may not be a 10% customer for the full quarter, if that makes sense or for the full year.”

asked by Sean O'Loughlin · answered by Daniel Fleming

How many customers do you expect to ramp Zero Flat Optics across in FY27? And can you elaborate on the OmniConnect gearbox's ability to handle training and inference in the same architecture?

Bill Brennan expects to ramp Zero Flat Optics across more than the current four customers in FY27. He detailed OmniConnect's composable architecture, where gearboxes convert VSR SerDes to different memory types (DDR, LPDDR) or protocols (PCIe Gen6/7, Ethernet, UALink, E.SUN) and speeds (100G, 200G, 400G), allowing XPU optimization for various workloads and future-proofing by simply upgrading the gearbox.

“You can imagine when 200 gig per lane is really ready for that given customer you could just simply drop in a new gearbox that would support 200 gig per lane with any of the protocols we've talked about being Ethernet or UAL or E.SUN. And you can extend that case to scale out as well.”

asked by Sujeeva De Silva · answered by William Brennan

How should we think about the $5 billion AEC TAM split between front-end vs. back-end links? And why are gross margins guided down for Q4?

Bill Brennan stated that scale-out networks represent the largest AEC opportunity, with front-end links likely 20-25% of the scale-out market, and disaggregated/switch markets having significant potential. Dan Fleming explained that the Q4 gross margin guide is due to product mix and conservative forecasting, reiterating the long-term expectation of 63-65%.

“front end is going to be smaller than scale out probably on the order of -- it will probably be 20% to 25% of the total scale-out market as we see it.”

asked by Karl Ackerman · answered by William Brennan