Credo (CRDO) first look: strong growth, rich price, concentrated customers

Ozkent starts a new format in the September 24, 2026 recording: a five-minute first look . The goal is not a deep report but a quick filter for whether a business merits longer study. First up is Credo, a name the author holds himself, selling connectivity for artificial intelligence data centers.

What it sells and who buys

The firm moves data between graphics processors, central processors and memory. When links stall, costly processors sit idle, which no operator wants. Revenue comes in three forms: copper cables , connectivity chips and usage rights for the technology. Buyers are large data centers, hyperscale operators, Neo Cloud builders and AI hardware makers. Sales run direct and through licensing, so third parties can fabricate chips from its rights. An AEC (active electrical cable) is a copper cable carrying a small circuit that cleans the signal. A DSP (digital signal processor) is the circuit that restores degraded signals. The analogy is a concert engineer removing hiss at the mixing desk.

Rivals are Marvell, Broadcom and Astera Labs. The author also owns Marvell and Astera, convinced by the connectivity trade. Credo stands out with low-power AEC design plus a narrow, chiplet-centered lineup. Its edge is stickiness. Tailoring chips and cables to one buyer from first contact to volume shipment usually takes two to three years. Once a system is embedded, buyers rarely switch, which steadies revenue.

Balance sheet and guidance

Latest quarter revenue reached 479 million dollars, up 114.7 percent from a year earlier and 9.6 percent from the prior quarter. GAAP gross margin printed 64.5 percent , non-GAAP 68 percent, a firm level for a hardware maker. Margin slipped 2.9 points year on year; still elevated, yet the cause deserves a closer read. Operating spend was 188.4 million on GAAP and 95.2 million ex-items. Net income came at 129.4 million GAAP and 236.3 million adjusted. Per-share earnings were 0.67 and 1.20 dollars. Cash plus short-term holdings stood at 764.3 million against 20 million of debt. The author cites 438.6 million of free cash generation. Guidance points to 525-535 million in revenue with GAAP gross margin of 62.9-64.9 percent.

Valuation, chart levels and risks

The price is demanding. Trailing twelve-month earnings multiple is 67, forward 30.8. The PEG ratio the author watches reads 0.21, below one, which favors the growth case. His scorecard gives 49 of 100: quality, expansion and debt resilience score well, valuation drags. The chart shows a pullback from highs, a base forming, then a turn upward, though not yet powerful. Below 177 dollars the structure breaks down; a daily finish above 198.97 signals repair. The recording notes a September 23 close of 193.87. A system named DEER flags 182.11 as an exit, built to stop losses from compounding. The author suggests choosing between small staged trims and the technical stop.

Three open questions remain. First, concentration : in 2026 three buyers produced 84 percent of revenue, so one design or spending shift could hurt. Second, margin: a supplier this fast-growing with firm demand should normally defend pricing, making the 2.9-point slip worth probing. Third, dilution: staff option grants plus capital raises keep lifting share count and thinning existing holders. Add multiple risk: when markets sour or headlines wobble, richly priced names can fall hard.

My closing take is balanced: a critical product in a fast-expanding field, worth deeper study. Still, nobody should buy on a five-minute screen or on another person's tip. My watch list is short: does the buyer base widen, does gross margin recover, does dilution stay contained. Until those answer, the picture stays incomplete.