Credo Technology (CRDO): The AI Interconnect Small-Cap Breaking Into Hyperscale

Recommendation: Buy (with staggered entry). Credo Technology has transitioned from a money-losing AI-interconnect hopeful to the fastest-growing profitable semiconductor small-mid cap on the tape. Fiscal Q3 2026 (quarter ended January) delivered $407M in revenue, up 201.5% year-over-year and 51.9% sequentially, with non-GAAP gross margin of 68.6% and non-GAAP diluted EPS of $1.07 — a quarter that alone nearly matched all of fiscal 2025's $437M in revenue. Management guided Q4 to $425–$435M, implying another step-up into fiscal year-end.

The balance sheet is pristine: $1.30B in cash, $16M in debt, net cash of $6.97 per share, a current ratio of 10.8, and ROIC of 60.6%. Credo used that cash on April 16, 2026 to announce a $750M acquisition of DustPhotonics, adding silicon photonics to a product stack that already includes Active Electrical Cables (AECs), PAM4 optical DSPs, SerDes IP, line-card PHYs, and PCIe retimers. Management has framed three new multi-billion-dollar TAM expansions — ZeroFlap optics, Active Linear Cables (ALCs), and OmniConnect — layered on top of the core AEC and DSP businesses.

The valuation is high on trailing numbers (PE 88, EV/Sales 27) but reasonable on forward estimates: forward PE of 36, forward P/S of 15, PEG of 0.31, and a consensus 5-year EPS growth forecast of 74% per year across 14 analysts. Consensus rating is Strong Buy with an average target of $179.62, roughly 12% above the current print.

The core risk is pace, not thesis. The stock has rallied +326% in twelve months, re-rated from ~$96 to $160 in eight trading sessions, carries a beta of 2.72, a 14-day RSI above 71, and Bollinger %B near 94. That argues for scaling in on pullbacks toward the 20-day EMA (~$126) and 50-day SMA (~$116) rather than chasing the breakout. Customer concentration at a single hyperscaler and the ongoing copper-to-optical transition remain the two items we are watching most closely.

Daily close vs. 14-period Kaufman Adaptive Moving Average (KAMA) and 50-day Simple Moving Average. The KAMA flattens through chop and accelerates through real trend changes; the recent turn higher is the most decisive signal on this chart.

Credo Technology Group Holding Ltd designs high-speed connectivity silicon and cable solutions for AI, cloud, and enterprise infrastructure. The product portfolio spans five engines: Active Electrical Cables (HiWire AECs) for short-reach in-rack links, PAM4 optical DSPs for pluggable transceivers, line-card PHYs, SerDes IP and chiplets, and PCIe retimers for GPU and accelerator fabrics. Customers are predominantly hyperscalers and the optical module, ODM, and OEM supply chain that serves them.

Credo went public in January 2022 and is incorporated in the Cayman Islands with operating headquarters in the United States. The fiscal year ends in late April or early May. Reported employee count is 622, which puts revenue per employee above $1.7M — a signature of a highly leveraged, IP-rich semiconductor business.

The AI training and inference buildout is the demand backdrop for every Credo product line. Hyperscaler capex remains the single biggest driver: each new accelerator cluster needs an order of magnitude more high-speed interconnect — inside the rack (copper AECs and ALCs), across racks (DSP-driven optics), and across scale-out fabrics (SerDes and retimers). Credo's pitch is that it participates in all three pockets while avoiding the thermal and power costs of purely optical solutions in the shortest-reach segment.

The three TAM expansions management highlighted with Q3 results stack into that picture. ZeroFlap optics addresses reliability gaps in pluggable 800G/1.6T transceivers. Active Linear Cables (ALCs) target a low-power copper/optical hybrid segment that sits between AECs and full optical. OmniConnect is positioned as a platform for memory and GPU scale-out connectivity. The April 2026 acquisition of DustPhotonics extends that roadmap into silicon photonics — the missing piece needed to compete more directly with integrated optical vendors.

The consensus view on the sell side is constructive. StockAnalysis aggregates 14 analysts on CRDO with an average price target of $179.62 (about 11.8% above the April 17 close) and a Strong Buy consensus; Credo was upgraded to Strong Buy on April 17 on the back of rising forward estimates. The five-year revenue growth forecast is 44.8% per year; five-year EPS growth is forecast at 74.2%.

CRDO's chart has three distinct regimes over the last 90 days: a January high above $160, a sharp drawdown to an $87.81 low on March 30, and an explosive eight-session rally that has round-tripped the stock back to new 52-week highs. The 50-day SMA has flattened around $116 and is now turning up; the 14-period KAMA crossed back above price and has accelerated to $138 — its steepest slope in the series and the single cleanest confirmation that the trend is back on.

Momentum has stretched. The 14-day RSI is 71.1 (mildly overbought but not extreme for a stock coming out of a 30% drawdown). Stochastic %K is 94 and MFI is 76, so short-term mean reversion is possible. Bollinger %B printed 93–133 over the last five sessions, consistent with a legitimate breakout but a high-risk chase zone. ADX finally broke above 25 (25.9) on April 17, confirming that the new trend has real directional strength — the first time the tape has carried that signature in months.

Practical read: trend direction is long, but entry quality matters more than usual. A retest of the April 10–13 breakout shelf ($107–$120) would be a more favorable risk/reward zone than buying the gap.

The fundamental profile is unusual for a semiconductor small-mid cap: Credo is simultaneously growing triple digits, already profitable, net-cash rich, and generating meaningful free cash flow. Trailing twelve-month revenue is $1.07B (+226% YoY), gross margin is 67.8%, operating margin is 30.2%, net margin is 31.8%, and free cash flow margin is 26.6%. The Piotroski F-score is 7 and the Altman Z-score is 60.5, which is as clean a balance-sheet signature as this category produces.

The Q3 FY2026 print (reported March 2, 2026) was the inflection quarter. Revenue of $407M crossed a threshold management had been signaling for several quarters, and importantly the non-GAAP gross margin held at 68.6% while operating expenses were tightly contained at $77.4M non-GAAP. That combination drove non-GAAP EPS to $1.07 and net income to $208.8M in a single quarter. The Q4 guide of $425–$435M implies another ~5% sequential step-up, with non-GAAP gross margin guided modestly lower (64–66%) on mix — a typical pattern as lower-margin AEC volume scales.

Valuation needs to be read on forward numbers because the P&L is still catching up. Trailing P/E is 88 and EV/Sales is 27 — numbers that only make sense alongside the 200%-plus growth. Forward P/E drops to 36.3, forward P/S to 15.2, and the PEG ratio (based on 5-year EPS growth) is 0.31. On a growth-adjusted basis CRDO is cheaper than the median AI-hardware name.

The bull case is compelling, but the risk surface is also unusually broad for a $30B market cap. We think about it in five buckets:

Credo Technology has graduated from "speculative AI small cap" to "profitable, cash-rich, fast-growing AI infrastructure name" in the span of four quarters. The chart will look expensive on any near-term screen; the fundamentals, backed by a $1.3B cash cushion, 30%+ operating margins, and a visibly expanding product surface, are what justifies owning the stock through normal volatility. We rate CRDO a Buy with a preference for staggered accumulation rather than chasing the current breakout.