This analysis is based on completed internal research memos dated May 29, 2026, including a deep dive into the four critical questions (CPO timeline, earnings setup, customer concentration, and margin defensibility), not a fresh market-data refresh. All figures are from the review timestamps.
Stock and Call
Credo Technology (CRDO) is rated as a Watchlist at $222.35. The call: wait for or trade around the June 1 FY2026 earnings print. This is a hypergrowth company (+226% YoY) with a PEG of 0.40 that implies the market expects growth to collapse — a pricing disconnect that June 1 earnings could resolve. If the FY2027 revenue guide exceeds $2.15B consensus and margins stabilize toward 65%, the re-rating case is strong. If the guide disappoints (below $1.8B), the deceleration is sharper than priced. Conviction: Moderate — the growth numbers are extraordinary, but $41B market cap means this is no longer under-covered. Entry: $222 pre-earnings is speculative; post-earnings on a guide that validates +50% FY2027 growth is more attractive. Target: $260 (Barclays, April 28, the high-analyst target). Stop-loss: below $150, which would imply a PEG re-rating toward 1.0 at current earnings.
Why It Is Promising
Credo's Active Electrical Cables (AECs) have become the de facto standard for high-speed intra-rack connectivity in AI clusters. Every GPU deployed in a hyperscaler data center generates 4-8 high-speed interconnects — a 10,000-GPU cluster creates 40,000-80,000 links, each requiring Credo's retimer silicon or AEC cable assemblies. The scaling relationship is near-linear: more GPUs = more links = more Credo revenue.
The operating metrics are extraordinary — revenue/employee of $1.72M (approaching NVIDIA's $6M at one-tenth the scale), 68% gross margins, 60.6% ROIC, and 226% YoY revenue growth to $1.07B TTM. The PEG of 0.40 is so low that it borders on a pricing error: at a 46.7x forward P/E, the market is implying that 54% consensus growth will decelerate to roughly 12%. There is no evidence in Credo's four consecutive beat-and-raise quarters that this deceleration is imminent.
Co-Packaged Optics (CPO) — the most-cited bear argument — is a 2029-2030 threat, not a 2026-2027 one. Three independent sources (SemiAnalysis, The Next Platform, Broadcom) align on late-decade volume deployment. Critically, ~70% of Credo's revenue (AECs and related retimers) is structurally insulated from CPO: CPO switches eliminate optical DSPs but still need copper AECs for intra-rack switch-to-server connections. The recent DustPhotonics acquisition adds silicon photonics capability, giving Credo a path to participate in the optical transition rather than be displaced by it.
Customer concentration (top 3 at 88%, top 1 likely Amazon at 64%) is diversifying steadily — four hyperscalers are now each >10% of revenue, up from one at 67% in FY2025. And the AEC architecture is already the hyperscalers' standard, creating requalification switching costs that would take 24-36 months to overcome even if someone tried to build a captive solution.
Evidence Quality
Catalyst Path
The immediate catalyst is FY2026 Q4 earnings on June 1, 2026. Consensus expects $425-435M revenue (management's own guide). The real number to watch is the FY2027 revenue guide — consensus is $2.15B (+58.6%). If management guides above $2.15B, the PEG of 0.40 re-rates upward. If they guide below $1.8B, the deceleration is sharper than priced. Beyond earnings: the DustPhotonics acquisition adds optical capability that addresses the CPO risk head-on. The 1.6T product roadmap (AECs and retimers at 224Gbps SerDes) positions Credo for the next GPU generation (NVIDIA Rubin, 2027). A fifth hyperscaler is reportedly ramping, further diversifying the customer base.
Main Risks
Customer concentration is the single biggest risk. The top customer (likely Amazon AWS) represents ~64% of revenue. If a hyperscaler designs captive AEC/retimer silicon — as Amazon has done with Graviton CPUs and Trainium AI chips — Credo loses its anchor customer. The 88% top-3 concentration means any single customer loss is existential. CPO is a genuine long-term threat to the optical DSP business (~15-20% of revenue), even though the AEC core (~70%) is insulated. At $41B market cap, Credo is no longer under-covered — 18 analysts follow it, and the stock is up 249% in a year. A growth deceleration to below 100% YoY for two consecutive quarters (possible by Q2-Q3 FY2027 if consensus 58.6% holds) would trigger sharp multiple compression from 38x sales toward 15-20x.
What Would Change the Call
Watchlist Action
Next checkpoint: Q4 FY2026 earnings, June 1, 2026 after market close. Watch for: (1) FY2027 revenue guide vs $2.15B consensus, (2) Q4 gross margin actual vs 64-66% guide, (3) customer count update — is the 5th hyperscaler scaling?, (4) DustPhotonics integration timeline, and (5) ALC sampling timeline confirmation. The stock is already moving in pre-market ($229.43, +3.18%), suggesting buying ahead of earnings. A post-earnings entry on a strong guide is the prudent approach. If the stock sells off on a guide miss toward $150-170, that could be a much more attractive entry for a company growing 50%+ — but a miss this dramatic would also raise questions about whether the hypergrowth chapter is closing.