CRDO · Credo Technology Group Holding Ltd · Stock research

Current Thesis

The catalyst void closed early. Credo scheduled its Q1 FY2027 results for 2026-09-01 after the market close, call at 2:00 p.m. Pacific, covering the fiscal quarter ended 2026-08-01. Third-party earnings calendars had been carrying 2026-09-09; the company's own scheduling release pulls the binary forward, and it now sits about seven sessions out rather than a month.

The product story is unchanged in substance. Credo sells the interconnect layer — active electrical cables, optical DSPs, SerDes IP — that hyperscalers buy by the rack, and fiscal 2026 turned that into roughly $1.34B of revenue and $662M of non-GAAP net income. What has moved is where price sits against the enthusiasm that produced it. The 2026-08-21 close was $230.57, 23.8% under the $302.52 52-week high set during the late-June initiation wave (Evercore $325, Stifel $350, BofA $340, all 2026-06-22/23). RSI(14) sits at 53.4 and the three-month price change is +5.6%. The August recovery stalled: the 2026-08-07 close was $249.89, and 2026-08-18 saw chip stocks sold on rising bond yields — the same session TD Cowen maintained Buy and lifted its target to $300 from $260.

One correction to the prior frame published here. The DustPhotonics acquisition was carried as a forward catalyst closing around Oct–Nov 2026. It did not: Credo completed the acquisition on 2026-05-28, three sessions before the Q4 FY2026 print. The share component of consideration is therefore already in the float and is not a pending overhang. Coverage of the closing put the transaction at up to $1.3B including an earnout; the completion release confirms an earnout structure exists but does not itemize the terms.

The narrative is maturing. The fresh-attention phase is dated and behind — the initiation cluster and the 52-week high both landed 2026-06-22 through the July peak. The narrative still works (TD Cowen's 2026-08-18 raise, an average sell-side target above the tape), but the flow is maintenance revisions rather than new coverage, and price has not reclaimed the June high in eight weeks. It is not saturated: the stock trades 23.8% below that high with a mid-range RSI, which is not the shape of a crowded late-cycle bid.

Bullish and bearish views on Credo Technology Group Holding Ltd

The model's bull view on Credo Technology Group Holding Ltd (CRDO), in brief: Q4 FY2026, reported 2026-06-01: revenue $437.0M, +157% YoY, against $432.0M consensus; adjusted EPS $1.16 vs $1.03 consensus. The bear view: The price-target ceiling has not moved since 2026-06-22. Both cases follow in full.

Bull Case

• Q4 FY2026, reported 2026-06-01: revenue $437.0M, +157% YoY, against $432.0M consensus; adjusted EPS $1.16 vs $1.03 consensus. A single quarter outsold the whole of fiscal 2025.
• Full-year FY2026 roughly $1.34B revenue with $662M non-GAAP net income — realized P&L behind the interconnect content story, not a serviceable-market slide.
• The guide is not a stretch bar. Q1 FY2027 was guided to $465–475M on 2026-06-01 against $461.6M consensus at the time; the calendar consensus into the print sat at $470.38M revenue and $1.16 EPS as of 2026-08-08 — inside the company's own band.
• Sell-side marks still sit above the tape. MarketBeat's panel of 19 analysts showed a Strong Buy consensus with an average 12-month target of $283.23 as of 2026-08-22, versus $279.29 on 2026-07-26. TD Cowen went to $300 on 2026-08-18, into weakness rather than into strength.
• The optical layer is owned, not pending. DustPhotonics closed 2026-05-28, adding silicon-photonics PIC capability across 800G, 1.6T and 3.2T near-packaged and co-packaged optics — the layer where AVGO and MRVL concentrate.
• Demand framing restated in August: BofA's 2026-08-03 work put hyperscaler capex above $1.2T and named nine beneficiaries.

Bear Case

• The price-target ceiling has not moved since 2026-06-22. Stifel's $350 is still the top mark. The newest raise, TD Cowen's on 2026-08-18, landed at $300 — below the ceiling and roughly 30% above the 2026-08-21 close, which is a normal Street cushion rather than a re-rating.
• Three officers sold into the July high. Form 4 activity summarized 2026-07-20: COO and director Yat Tung Lam ~56,000 shares (~$12.6M); CFO Daniel W.
• Valuation leaves no slack for a merely in-line quarter. Trailing P/E near 94.8 against a semiconductor industry median of 57.2, with a GF Value estimate of $219.90 versus a $237.92 print (both 2026-08-04).
• The July–August drawdown had no company cause. From $302.52 to a reported $195.39 low with no guidance change, then a partial rebuild and another fade on 2026-08-18 yields. Rate-path and sector-beta sensitivity dominates the near-term tape.
• Customer concentration is unresolved. AWS has historically been over 40% of revenue; the 2026-04-28 custom-silicon report produced a roughly -15% session on a headline alone.
• The print is a genuine binary and nothing resolves before it. The number that matters on 2026-09-01 is the Q2 FY2027 guide, not the quarter delivered.

Setup & Price Structure

• Reference levels. 52-week high $302.52 (late June). Reported early-August low $195.39. Close of $249.89 on 2026-08-07; close of $230.57 on 2026-08-21. The last close sits inside a roughly $195–$250 August range, nearer its upper half.
• Momentum is neutral. RSI(14) at 53.4 as of 2026-08-21, with a three-month price change of +5.6%. There is no extension above a rising short-term average to unwind, and equally no reclaimed base — the June breakout structure that produced the high has not been rebuilt.
• Crowding observables, stated plainly: the July insider cluster above; a static top-end target since 2026-06-22; listicle-grade retail coverage on 2026-08-14 ("5 AI Stocks with Strong Growth to Buy on the Rebound"); and an earnings date inside the next seven sessions. The share-issuance overhang that was previously listed as forward supply has been in the float since 2026-05-28.
• Sequencing matters this cycle. Nvidia reports 2026-08-26 after close and Marvell around 2026-08-27 (est.) — both land before Credo, so the setup carries two sector read-throughs it does not control, then its own print, then Broadcom around 2026-09-04 (est.).

Catalyst Calendar (next 30 days)

• 2026-08-26 — Nvidia Q2 FY2027 results, after close. Company guided $91.0B ±2%; a 40-analyst consensus at $91.85B revenue / $2.08 EPS. Sets AI-networking risk appetite three sessions before Credo reports.
• ~2026-08-27 (est.) — Marvell quarterly print (prior-year cadence, unconfirmed). Closest listed comparable on custom silicon and interconnect attach commentary.
• 2026-09-01 — Credo Q1 FY2027 results, after close; call 2:00 p.m. PT (company-confirmed). Quarter ended 2026-08-01. Tests the $465–475M guide against a $470.38M consensus, and the Q2 FY2027 guide against the sequential-acceleration claim.
• ~2026-09-04 (est.) — Broadcom quarterly print (prior-year cadence, unconfirmed). Largest single read on AI networking dollars and on how much of the optical layer hyperscalers keep in-house.
• ~2026-09-15 to 2026-09-16 (est.) — FOMC meeting. At a trailing multiple near 94.8x, rate-path commentary is the direct transmission channel into the valuation.

What Would Change Our Mind

The structure that has to hold is the early-August low. The rebuild from $195.39 to $249.89 on 2026-08-07 was the market's answer to the June-high drawdown; losing it on a weekly close below $195 says the rebuild was a bounce inside a topping process and that the June breakout base is gone, not merely untested. That is the gradeable line.

Three non-price conditions would flip the read independently. First, a Q2 FY2027 guide issued on 2026-09-01 implying a lower sequential growth rate than the quarter just reported — revenue can land inside $465–475M while the 800G-to-1.6T ramp visibly decelerates. Second, post-print revisions clustering below the $283.23 average target rather than above it, with the $350 ceiling still unmoved from 2026-06-22, which would date the sell-side fuel as exhausted. Third, further officer Form 4 sales filed in the weeks after the print, extending the July cluster into a pattern.

Conversely, a reclaim of the $250 area on a weekly close, with a Q2 guide above consensus and at least one target above $350, would re-date the narrative as accelerating rather than maturing.

Correlation Notes

• Sector beta dominates the near-term tape. The 2026-08-18 session — "chip stocks crater as bond yields bite" — moved CRDO with the complex on no company news, the same pattern as the drawdown from $302.52 with no guidance change.
• NVDA and AVGO are the read-through anchors. Nvidia's 2026-08-26 print lands before Credo's; Broadcom's, around 2026-09-04 (est.), lands after and can re-price the post-print move regardless of what Credo delivers.
• MRVL is the closest business comparable, and its attach-rate commentary around 2026-08-27 (est.) is a leading indicator for the interconnect content assumption. Optical peers AAOI and LITE trade the same headline flow with more volatility and less revenue.
• AWS is the single-name correlation that is not a market factor. With historic revenue concentration above 40%, one hyperscaler roadmap headline can decouple CRDO from SOXX/SMH in either direction.
• Rates are the multiple channel. A trailing P/E near 94.8 against a 57.2 industry median means the discount rate, not the order book, sets the day-to-day range between prints.