Credo Technology Group
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
Specialty connectivity silicon — Active Electrical Cables, optical DSPs and, since the May DustPhotonics close, silicon-photonics PICs — designed into hyperscaler AI cluster reference architectures. The moat is multi-year design-win stickiness inside the platform cycle, not software lock-in.
Credo's moat is multi-year design wins inside hyperscaler AI cluster reference architectures — once a SerDes IP block or AEC SKU is qualified, the customer sticks with it for the platform life:
- AEC Performance Lead at 800G/1.6T: Credo's ZeroFlap AECs remain the practical inside-rack option at 800G/1.6T versus passive DACs and active optical cables. Q1 was the seventh consecutive quarter of triple-digit YoY revenue growth, with AECs still the core engine even as optics starts to print. First-mover lead vs Marvell and Broadcom AEC offerings is intact; it is not a perpetual franchise — it lasts the platform cycle.
- Hyperscaler Design-In Stickiness: Once a hyperscaler qualifies a SerDes or AEC into a reference architecture, the design persists for the platform's 3–5 year deployment cycle. Q1 again had four domestic end-customers at or above 10% of revenue; management still expects three-to-four 10% customers through FY27. That is visibility, and it is concentration. Neo-cloud diversification is progress, not a completed de-risking.
- Vertical Integration into Optics: The May DustPhotonics acquisition (cash plus stock; goodwill stepped to $986M in Q1) puts silicon-photonics PICs next to Credo's DSP and SerDes IP. FY27 optical revenue is guided above $600M, with ZeroFlap Optics, PICs and optical DSPs each expected to contribute more than $100M — about half of FY27's absolute-dollar growth, per the June call, now showing up in the Q1 print and Q2 guide. Integration risk is the other side of the same coin.
Moat Verdict
Q1 does not change the moat labels: talent and design-win embedding stay strong, regulatory lock-in stays weakened, everything else is intact or N/A. Optics at a guided $600M is a second growth engine, not a new pillar — DustPhotonics is still an integration, not a completed SiPho franchise. The thesis remains hypergrowth plus share gain sized as a high-conviction speculative position, not a core compounder. Next trip is Q2 $525–$535M; missing it with four 10% customers is how concentration converts to volatility.
67.7 resilient · 65.0 vulnerable · 80/20 = 67.2 · = 67
Open a moat to read its note.
semiconductor IP/component vendor.
SerDes IP, PILOT diagnostics and the SDK ecosystem are real embedment in customer reference designs. DustPhotonics adds PIC design-in, not a second copy of customer-owned logic. Still a hardware franchise.
N/A.
High-speed SerDes and silicon-photonics designers are scarce, and Credo's team still explains its design-win pace. Broadcom, Marvell and Astera hire from the same pool, and no figure shows the bench blocks entry, so it rates intact.
AEC + optical DSP + SiPho PIC + IP licensing is a wider bundle after DustPhotonics, but Credo still does not set the rack BOM. Limited bundling power at the customer level.
N/A.
Standard SerDes export-control compliance only; there is no regulatory moat to rate. Previously weakened, which scored a pillar that never applied.
N/A.
Once a SerDes IP block or AEC SKU is qualified into a reference architecture, the design persists for 3–5 years — Q1's four 10% customers are that embedment showing up in the P&L. Strong inside the platform cycle; the cycle itself is the constraint.
N/A.
Fabless or equipment franchise whose edge is IP and process know-how, rated elsewhere; manufacturing volume does not give it a unit-cost lead rivals cannot buy.
Buyers are enterprises choosing on switching cost, integration and performance, which the other pillars rate. The name carries reputation, not a price premium it could hold on brand alone.
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
Specialty connectivity silicon — Active Electrical Cables, optical DSPs and, since the May DustPhotonics close, silicon-photonics PICs — designed into hyperscaler AI cluster reference architectures. The moat is multi-year design-win stickiness inside the platform cycle, not software lock-in.
Growth Score
Q1 FY2027 (ended August 1, reported Sep 1) printed revenue $479.0M (+114.7% YoY, +9.6% QoQ) and non-GAAP EPS $1.20, beating ~$470M / $1.17. Non-GAAP net income $236.3M (+140% YoY); non-GAAP operating margin 48.2%. Q2 guided $525–$535M revenue with non-GAAP GM 67–69%. FY26 was $1.335B (+206% YoY) with non-GAAP NI $662M; FY27 growth was lifted to >85% YoY with >$600M of optical revenue and a ~50% non-GAAP net-margin target. Close $206.63 on Sep 1; after-hours ~$185 on a take-profit print after a large run-up. Top-four customers each ≥10% of Q1 revenue.
Valuation Score
At the Sep 1 close of $206.63 (~$39B) CRDO is ~17× FY26 sales and ~33× a ~$6 FY27 non-GAAP EPS sketch on the >85% / ~50% net-margin guide — a hypergrowth multiple that still prices in share gains. After-hours ~$185 is ~25% below the $240 base. Street mean target ~$283. The old $45/$110/$160 ladder was overrun; this one is reset to the FY27 optical + AEC path.
The AEC Design-Win Moat
Credo's moat is multi-year design wins inside hyperscaler AI cluster reference architectures — once a SerDes IP block or AEC SKU is qualified, the customer sticks with it for the platform life:
- AEC Performance Lead at 800G/1.6T: Credo's ZeroFlap AECs remain the practical inside-rack option at 800G/1.6T versus passive DACs and active optical cables. Q1 was the seventh consecutive quarter of triple-digit YoY revenue growth, with AECs still the core engine even as optics starts to print. First-mover lead vs Marvell and Broadcom AEC offerings is intact; it is not a perpetual franchise — it lasts the platform cycle.
- Hyperscaler Design-In Stickiness: Once a hyperscaler qualifies a SerDes or AEC into a reference architecture, the design persists for the platform's 3–5 year deployment cycle. Q1 again had four domestic end-customers at or above 10% of revenue; management still expects three-to-four 10% customers through FY27. That is visibility, and it is concentration. Neo-cloud diversification is progress, not a completed de-risking.
- Vertical Integration into Optics: The May DustPhotonics acquisition (cash plus stock; goodwill stepped to $986M in Q1) puts silicon-photonics PICs next to Credo's DSP and SerDes IP. FY27 optical revenue is guided above $600M, with ZeroFlap Optics, PICs and optical DSPs each expected to contribute more than $100M — about half of FY27's absolute-dollar growth, per the June call, now showing up in the Q1 print and Q2 guide. Integration risk is the other side of the same coin.
Moat Verdict
Q1 does not change the moat labels: talent and design-win embedding stay strong, regulatory lock-in stays weakened, everything else is intact or N/A. Optics at a guided $600M is a second growth engine, not a new pillar — DustPhotonics is still an integration, not a completed SiPho franchise. The thesis remains hypergrowth plus share gain sized as a high-conviction speculative position, not a core compounder. Next trip is Q2 $525–$535M; missing it with four 10% customers is how concentration converts to volatility.
67.7 resilient · 65.0 vulnerable · 80/20 = 67.2 · = 67
Open a moat to read its note.
semiconductor IP/component vendor.
SerDes IP, PILOT diagnostics and the SDK ecosystem are real embedment in customer reference designs. DustPhotonics adds PIC design-in, not a second copy of customer-owned logic. Still a hardware franchise.
N/A.
High-speed SerDes and silicon-photonics designers are scarce, and Credo's team still explains its design-win pace. Broadcom, Marvell and Astera hire from the same pool, and no figure shows the bench blocks entry, so it rates intact.
AEC + optical DSP + SiPho PIC + IP licensing is a wider bundle after DustPhotonics, but Credo still does not set the rack BOM. Limited bundling power at the customer level.
N/A.
Standard SerDes export-control compliance only; there is no regulatory moat to rate. Previously weakened, which scored a pillar that never applied.
N/A.
Once a SerDes IP block or AEC SKU is qualified into a reference architecture, the design persists for 3–5 years — Q1's four 10% customers are that embedment showing up in the P&L. Strong inside the platform cycle; the cycle itself is the constraint.
N/A.
Fabless or equipment franchise whose edge is IP and process know-how, rated elsewhere; manufacturing volume does not give it a unit-cost lead rivals cannot buy.
Buyers are enterprises choosing on switching cost, integration and performance, which the other pillars rate. The name carries reputation, not a price premium it could hold on brand alone.
Growth Analysis
Growth Drivers
Key Risk
Four end-customers each ≥10% of Q1 revenue. If Marvell or Broadcom take the next-platform AEC socket, or DustPhotonics slips the optical $600M, a single-customer share loss still compresses revenue 30%+ in a quarter as the platform cycle turns.
Score Derivation
91.3 base + 2.7 trajectory − 10 risk = 84
Base 91 (30–40% CAGR mid 35%) + 3 trajectory (AEC and optical accelerating; IP stable) + 0 margin (Q1 non-GAAP GM 68%, FY27 guided in line with FY26) − 10 high concentration (top-4 each ≥10%; DustPhotonics integration) = 84
Price Scenarios (12–24 Months)
Valuation Multiples
| Forward P/E (FY27) | ~33× |
| Price / Sales (FY26) | ~17× |
| Price / Sales (FY27) | ~16× |
| PEG Ratio | ~0.9× |
| Net cash | $764M |
The multiple is a claim that 800G/1.6T design wins persist and that optics becomes a second engine at $600M+. If either fails, 33× becomes 20× quickly. After-hours weakness is take-profit, not a thesis break — unless Q2 misses $525–$535M.
Approximate figures as of September 2, 2026.
Where We Are vs Targets
Loading live price…
Marvell/Broadcom take next-platform AEC sockets, optical $600M slips, customer concentration shows up as a down quarter, and the multiple compresses toward 18–20× on a slower FY28.
- Lose a top-2 hyperscaler socket in the 1.6T/3.2T transition
- FY27 optical lands well below $600M as DustPhotonics integration slips
- Hyperscaler capex digestion in 2027 turns concentration into a 30%+ revenue air-pocket
FY27 converts near the >85% guide (~$2.47B), optics prints $600M+, Q2 $525–$535M holds, and the multiple settles ~35× on ~$7 FY28 EPS as 1.6T ramps.
- FY27 revenue ~$2.47B; Q2 in the $525–$535M range; H2 optical inflection arrives
- Three-to-four 10% customers persist; neo-cloud mix rises without a hyperscaler loss
- Non-GAAP GM stays 67–69%; ~50% net-margin target is approached, not missed
AEC franchise holds across top-5 hyperscalers, optics overshoots $600M, CPO/NPO attach starts to print, and the multiple holds ~40× on accelerating FY28.
- Design wins at all top-5 hyperscalers and a material neo-cloud slice by FY28
- Optical revenue exceeds $600M and SiPho/CPO becomes a disclosed third leg
- Net margin holds ~50% while revenue compounds through the 1.6T cycle