Finn Stockinger(@FinnStockinger):Quick thoughts on $CRDO after their earnings call 👇 ➡️Optics is taking the crown from copper AECs built the Credo we know today, but optics is officially taking over as the primary growth engine. Credo reiterated >$600M in optical revenue for FY27, with ZeroFlap optics, PICs, and optical DSPs each contributing >$100M. They are successfully moving past the "just a cable company" narrative. ➡️Dust Photonics looking like a brilliant M&A play The Dust Photonics acquisition paid off immediately with its first recognized PIC revenue in Q1 and confirmed FY28 design wins at 2 major players. Pairing in-house silicon photonics (PICs) with their custom DSPs gives Credo a clear cost (COGS) and system-level performance edge over standard module builders. ➡️Fixing one of AI’s biggest bottlenecks (ZeroFlap + Pilot) Link flaps in massive AI clusters can waste 10% to nearly 20% of expensive GPU utilization. Credo’s Pilot software acts like a "check engine" light—spotting fiber dust and latent ESD damage before a link actually fails. Slashing cluster bring-up time from 6–8 weeks down to just 5–6 days is a massive value prop for hyperscalers. ➡️Concentrated hyperscaler traction (and product expansion) Customer concentration remains high - their top 4 customers drove 84% of Q1 revenue (33%, 28%, 13%, and 10%). However, Credo now holds deep relationships with 5 top hyperscalers overall and is expanding across Neo Clouds. Crucially, these Tier-1s aren't just buying AECs anymore; they're adopting Credo's broader optical and retimer portfolio. To support this wave, Credo has been prepping its supply chain for 18–24 months - leaning heavily into working capital with inventory up $62M+ sequentially to $313M to ensure they can scale production effortlessly for the 2H FY27 ramp. ➡️Sneaky monetization play on Inference (OmniConnect) Credo is positioning early for the memory fanout and bandwidth wall in inference. Their Weaver gearboxes interface with LPDDR5/LPDDR6 to boost memory capacity up to 2TB (starting with customer Positron). Management expects this to represent thousands of dollars in Credo content per GPU starting in FY28. ➡️Agnostic strategy over religious tech battles Instead of picking sides in copper vs. optics vs. NPO/CPO, Credo is playing the whole board: > Short-reach copper (AECs up to 6.5m for 1.6T) > Micro-LED active cables (ALCs up to 30m, ramping FY28) > Full optics & NPO (joined Open CPM MSA consortium for scale-up) They win regardless of which topology hyperscalers or Neo Clouds choose. What’s your take on $CRDO here?

2026.09.01 22:29

Quick thoughts on $CRDO after their earnings call 👇 ➡️Optics is taking the crown from copper AECs built the Credo we know today, but optics is officially taking over as the primary growth engine. Credo reiterated >$600M in optical revenue for FY27, with ZeroFlap optics, PICs, and optical DSPs each contributing >$100M. They are successfully moving past the "just a cable company" narrative. ➡️Dust Photonics looking like a brilliant M&A play The Dust Photonics acquisition paid off immediately with its first recognized PIC revenue in Q1 and confirmed FY28 design wins at 2 major players. Pairing in-house silicon photonics (PICs) with their custom DSPs gives Credo a clear cost (COGS) and system-level performance edge over standard module builders. ➡️Fixing one of AI’s biggest bottlenecks (ZeroFlap + Pilot) Link flaps in massive AI clusters can waste 10% to nearly 20% of expensive GPU utilization. Credo’s Pilot software acts like a "check engine" light—spotting fiber dust and latent ESD damage before a link actually fails. Slashing cluster bring-up time from 6–8 weeks down to just 5–6 days is a massive value prop for hyperscalers. ➡️Concentrated hyperscaler traction (and product expansion) Customer concentration remains high - their top 4 customers drove 84% of Q1 revenue (33%, 28%, 13%, and 10%). However, Credo now holds deep relationships with 5 top hyperscalers overall and is expanding across Neo Clouds. Crucially, these Tier-1s aren't just buying AECs anymore; they're adopting Credo's broader optical and retimer portfolio. To support this wave, Credo has been prepping its supply chain for 18–24 months - leaning heavily into working capital with inventory up $62M+ sequentially to $313M to ensure they can scale production effortlessly for the 2H FY27 ramp. ➡️Sneaky monetization play on Inference (OmniConnect) Credo is positioning early for the memory fanout and bandwidth wall in inference. Their Weaver gearboxes interface with LPDDR5/LPDDR6 to boost memory capacity up to 2TB (starting with customer Positron). Management expects this to represent thousands of dollars in Credo content per GPU starting in FY28. ➡️Agnostic strategy over religious tech battles Instead of picking sides in copper vs. optics vs. NPO/CPO, Credo is playing the whole board: > Short-reach copper (AECs up to 6.5m for 1.6T) > Micro-LED active cables (ALCs up to 30m, ramping FY28) > Full optics & NPO (joined Open CPM MSA consortium for scale-up) They win regardless of which topology hyperscalers or Neo Clouds choose. What’s your take on $CRDO here?

Finn Stockinger@FinnStockinger

2026.09.01 20:32

Just as I predicted - a classic beat & raise for $CRDO Credo blew past Wall Street estimates across the board: revenue hit $479M vs the $435M expected, and Non-GAAP EPS crushed it at $1.20 vs $0.98 consensus. The Q2 guidance of $525M–$535M completely wiped out analyst expectations ($465M). AI infrastructure demand is real, Non-GAAP margins hold strong at 68%, and operational leverage is insane. Wall Street will be forced to aggressively hike price targets after this clear execution showcase. How market react? Of course - sell out, what the times.