Stock Analysis · Bulls vs Bears

Credo Technology Group CRDO

🐂 Bulls vs Bears 🐻Report #212 Oct 2026
★ Willem owns this stock · 2% of the portfolio
37
moderate risk▼ −14 vs last
🐂 9vs11 🐻
Bears lead
36%
▲ Chance of the biggest tailwind the forecasts miss
17%
▼ Chance of the biggest headwind the forecasts miss
The verdict

The evidence still leans bull on the business. Credo beat the top of its revenue guide in 8 of 8 quarters, still has no AEC rival shipping in volume (Astera is pre-production at 800G, and Marvell showed no AEC customer at its 2026-10-06 Investor Day), and FY28 EPS estimates rose from $8.85 to $9.70 while the stock halved and recovered. The risk is concentrated and binary: a few buyers with $4.2M of contracted…

🎙️ Bulls vs Bears podcast — Credo Technology Group
Alex and Sam debate every bull and bear argument · 13:00
0:0013:00
⤓ Download

Written by Claude (Anthropic’s AI) from public sources · Willem owns CRDO · AI can make mistakes — numbers in brackets point to the sources · Not financial advice

🐂 Bull arguments 9

  1. 1No AEC rival is shipping in volume at 800G yetStrong
  2. 2Eight straight beats above the top of the guided rangeMedium
  3. 3A cash machine: ~$439M FCF in twelve months, no debtStrong
  4. 4FY28 estimates rose while the multiple fell by halfMedium
  5. 5Optics is guided past $600M with three $100M linesMedium
  6. 6The customer base is slowly broadeningMedium
  7. 7Scale-up retimer revenue has started, ahead of standards siliconMedium
  8. 8New mid-reach products target the gap copper is losingWeak
  9. 9Hyperscaler capex for 2027 is guided up againMedium

🐻 Bear arguments 11

  1. 1A few buyers, no contracts, $4.2M of backlogHigh
  2. 2At 200G per lane the copper reach shrinksMedium
  3. 3Astera and Marvell are building ready-made second sourcesMedium
  4. 4Profit converts to cash poorly as working capital swellsMedium
  5. 5Stock pay is now ~18–20% of revenueMedium
  6. 6FY27 needs a ~45% bigger second halfMedium
  7. 7Optics is new, laser-constrained and not in the risk factorsMedium
  8. 8A capex pause would show up within one quarterMedium
  9. 9Expectations and a beta of 3.25 make every print binaryMedium
  10. 10Insiders only sell, including one CEO sale outside a planLow
  11. 11One foundry, Taiwan, China R&D and an unresolved export issueLow

Bull arguments show how well they survive the devil’s advocate (strong / medium / weak); bear arguments show the damage they could do after the counter-argument (high / medium / low).

Bull case

🐂 Bull #1No AEC rival is shipping in volume at 800G yet

  • The point: Credo's core product still has no proven second source at hyperscale, ten months after Marvell launched a kit designed to create one.
  • Evidence:
    • AECs drove more than 90% of the year-on-year revenue increase in Q1 FY27, and management says AEC revenue "more than tripled" in FY26 [1,4].
    • Astera Labs' Q2 2026 call (2026-08-04): it "shipped pre-production volume" of 100G-per-lane Taurus modules for 800G AECs. It is not yet in production volume [23,24].
    • Marvell's Investor Day (2026-10-06) set FY31 targets of $70–90B revenue and a ~$65B interconnect market. Neither the transcript nor Futurum's write-up shows an AEC customer, AEC revenue or a Golden Cable production win [21,22].
    • Marvell's Golden Cable programme (2025-12-09) names cable makers (Foxconn Interconnect, Luxshare, "20+" partners) but no hyperscaler [20].
    • Credo settled its AEC patent fight: Volex in August 2025, and TE and Molex cross-licences in March 2026 (secondary). Amphenol's status was not found [51].
  • Why it matters: each quarter without a qualified rival at 800G is another quarter of ~68% gross margin on the largest product. The 800G installed base also sets the default for 1.6T requalification.
  • 😈 Devil's advocate:
    • "Pre-production" is one step from production. Astera says it is designed into "multiple platforms" at its "lead hyperscaler" [24].
    • The battle that matters is 1.6T (200G per lane). Credo says its 1.6T ramp follows "a similar kind of time frame as the rest of the market" [4], which means no head start.
    • The only third-party AEC market size (650 Group: $644M in 2025, $1.4B in 2029 [20]) is smaller than Credo's own AEC revenue. Nobody outside the company can measure its share.
  • Verdict: Strong Re-tested with new evidence (Astera Q2, Marvell Investor Day), and it holds for FY27. FY28 is open.

🐂 Bull #2Eight straight beats above the top of the guided range

  • The point: management keeps under-promising, which lowers the chance of a guidance shock.
  • Evidence:
    • Revenue beat the top of the range in all 8 quarters from Q2 FY25 to Q1 FY27. Examples: Q3 FY26 $407.0M vs $335–345M; Q1 FY27 $479.0M vs $465–475M [2,7,8,9,10,11].
    • The FY27 growth guide moved from ">50%" (March) to ">80%" (June) to ">85%" (September) [4,5,6].
    • FY26 grew 206%; the September 2025 call had pointed to about 120% [3,6].
    • H1 FY27 was guided in June to "mid single digit" sequential growth. Q1 came in at +9.6%, and the Q2 guide midpoint is +10.6% [4,5].
    • Ernst & Young has been auditor since 2018. Internal control is effective, with no restatement and no material weakness [3,14].
  • Why it matters: the ~$2.47B FY27 plan is back-end loaded (Bear #6), so a record of conservative guides is the main reason to believe it.
  • 😈 Devil's advocate:
    • The beats shrank from ~$70M (Q3 FY26) to $2M and $4M in the last two quarters [2,7].
    • Non-GAAP opex beat the top of its guide in each of the last three quarters ($77.4M vs $68–72M; $81.7M vs $76–80M; $95.2M vs $86–90M) [2,7,8].
    • The Q1 GAAP guide (gross margin 66.9–68.9%, opex $167.6–171.6M) ignored DustPhotonics purchase accounting, even though the deal closed on 2026-05-28. The actuals were 64.5% and $188.4M [2,7].
  • Verdict: Medium Strong. The revenue record is perfect, but the cushion is now thin and cost guides are not reliable.

🐂 Bull #3A cash machine: ~$439M FCF in twelve months, no debt

  • The point: Credo funds 3nm tape-outs, acquisitions and supply deposits from its own cash flow.
  • Evidence:
    • FCF over the last four quarters was $38.4M + $139.7M + $177.5M + $82.9M = $438.5M, on $1,591M revenue. That is a 27.6% FCF margin [1,3][derived].
    • Cash plus short-term investments were $764.3M at 2026-08-01, after paying $735.6M net cash for DustPhotonics [1].
    • No borrowings, revolver or going-concern language [1,3].
    • Non-GAAP operating margin was 48.2% in Q1 FY27 [2].
  • Why it matters: the business cannot be forced into a dilutive raise, and it can carry $313M of inventory and $190M of capacity deposits through a demand air pocket.
  • 😈 Devil's advocate:
    • Q1 FCF fell to 17.3% of revenue because working capital absorbed $151.3M [1].
    • A new capacity-reservation agreement signed after quarter-end needs another $102.4M of refundable deposits within FY27 [1].
    • Cash went from $1.44B to $0.76B in one quarter [1].
  • Verdict: Strong Liquidity is not the risk. Cash conversion is a quality question, covered in Bear #4.

🐂 Bull #4FY28 estimates rose while the multiple fell by half

  • The point: the June–September sell-off was a de-rating, not a downgrade of the business. The stock now trades at a discount to its connectivity peers.
  • Evidence:
    • FY28 consensus EPS rose from $8.85 to $9.70 in 90 days, with 14 upward and 1 downward revisions in 30 days (2026-10-05) [36].
    • Consensus (20 analysts, 2026-10-12): FY27 revenue $2.50B and EPS $6.31; FY28 $3.87B and $9.69 [34].
    • Forward P/E is about 27.9x for CRDO, against 58.7x for Astera, 46.7x for Marvell and 46.0x for Lumentum [35].
    • CRDO's own forward P/E was about 44x on 2026-06-30 [38].
    • The stock fell from a $302.52 close (2026-06-22) to $150.09 (2026-09-14), then recovered to about $203–205 on 2026-10-12 [44,31].
  • Why it matters: with estimates rising, the multiple, not earnings, did the damage. That limits how much further a de-rating alone can hurt.
  • 😈 Devil's advocate:
    • A discount to Astera and Marvell may simply reflect more concentration and a single-product business (Bear #1).
    • At about $205, the stock is 21x FY28 EPS that assumes another +55% growth year [34].
    • Price targets are falling even as estimates rise: Stifel cut to $310 from $350 on 2026-10-05, and Mizuho to $245 from $290 on 2026-09-21 [32,33].
  • Verdict: Medium This is new this report. The valuation is no longer extreme, but it is far from cheap for a company with this concentration.

🐂 Bull #5Optics is guided past $600M with three $100M lines

  • The point: Credo is becoming a two-engine company, which reduces the copper-substitution risk.
  • Evidence:
    • FY27 optical revenue is guided above $600M, with ZF optics, silicon-photonics PICs and optical DSPs each above $100M. The guide was raised from >$500M (April) and reiterated in September [4,5,13].
    • ZF optics production shipments are under way. First PIC revenue was booked in Q1, and first 1.6T DSP revenue is "on track for later this fiscal year" [4].
    • Brennan calls FY27 "just a steppingstone" and expects "outsized growth again" in FY28 [4].
    • At OCP (from 2026-10-12), Credo is showing 1.6T ZeroFlap optics that pair its DSP with silicon photonics [30].
  • Why it matters: if optics reaches the ~50/50 balance with copper that management describes [5], a copper setback no longer threatens most of the revenue.
  • 😈 Devil's advocate:
    • Q1 optical revenue was not disclosed, so the run-rate cannot be checked [4].
    • The guide is "accelerating in the second half", at the same time as an industry-wide indium-phosphide laser shortage: Lumentum says it undershipped demand by ~30% [27].
    • The two major PIC design wins "do not include the DSP" [4], so the full-stack pitch is not yet in the wins.
  • Verdict: Medium This is unchanged from report #1, because no new disclosure tests it until Q2 (early December).

🐂 Bull #6The customer base is slowly broadening

  • The point: concentration is falling while revenue grows.
  • Evidence:
    • The largest contracting party fell from 67% of revenue (FY25) to 49% (FY26) and 43% (Q1 FY27). The second is 28% [1,3].
    • End customers were 33/28/13/10% in Q1 FY27, against 39/32/17% in Q3 FY26 [4,6].
    • The fourth 10% customer in Q1 is a new name, different from Q4's fourth [4].
    • Management says neoclouds could reach "on the order of 20%" of revenue [5]. TensorWave, an AMD-based neocloud, is a named partner [30].
  • Why it matters: every new 10% customer makes the loss of one top buyer survivable rather than a cliff.
  • 😈 Devil's advocate:
    • The top end customer has sat at 33–35% for a year (35% in Q1 FY26, 33% in Q1 FY27) [4,6].
    • Two contracting parties still hold 85% of receivables [1].
    • The fourth 10% slot rotates between customers, so it is not yet a stable fourth pillar.
  • Verdict: Medium

🐂 Bull #7Scale-up retimer revenue has started, ahead of standards silicon

  • The point: Credo already earns money in the scale-up network, the links that join AI chips into one pod.
  • Evidence:
    • In March 2026 management said: "We do not have any revenue for scale-up" [6].
    • By September, retimer growth was "driven by scale up deployments" [4]. In June management said FY28 would be "more substantial" [5].
    • UALink 2.0 was ratified on 2026-04-07, with first products late 2026 to early 2027. ESUN is in progress at OCP [42].
    • Astera's CEO says "the scale-up within a rack will continue to be copper" [24].
  • Why it matters: scale-up needs many more links per accelerator than scale-out, so it raises Credo content per AI chip.
  • 😈 Devil's advocate:
    • Marvell says copper reach at 200G per lane is about 2.5 m, so scale-up beyond one rack needs optics. It expects "multi-hundred-million" dollars of scale-up optics revenue from next year [21,22].
    • Credo has given no scale-up dollar figure.
  • Verdict: Medium The revenue is real and new since March, but its size is unknown.

🐂 Bull #8New mid-reach products target the gap copper is losing

  • The point: if copper's reach shrinks, Credo has products for the 7–50 m gap.
  • Evidence:
    • At OCP, Credo is showing a 1.6T ZeroFlap Active Light Cable (ALC). Coverage cites about 30% less power than comparable AECs and up to 50 m reach, with commercial shipments in 2027 [30,31].
    • A micro-emitter optical backplane developed with SENKO is claimed at up to 10x the bandwidth of copper backplanes [30].
    • Weaver/OmniConnect is sized by management at $2,000–3,000 per GPU, with Positron as the first customer [5].
  • Why it matters: these products hedge Bear #2 directly. A hyperscaler win on any one would be a new multi-hundred-million-dollar line.
  • 😈 Devil's advocate:
    • The stock fell about 6% on 2026-10-12, which Investing.com partly attributes to a "sell the news" reaction to a product with a distant ship date [31].
    • Avicena already ships microLED evaluation kits [45].
    • No hyperscaler customer is named for any of these products.
  • Verdict: Weak Medium. These are options, not earnings.

🐂 Bull #9Hyperscaler capex for 2027 is guided up again

  • The point: demand for Credo's products is a function of AI capex, and capex is still rising.
  • Evidence:
    • The top four's 2026 capex guides total $735–750B: Amazon $220B, Alphabet $195–205B, Microsoft ~$175B, Meta $130–145B. In January the total was ~$640B [28].
    • Alphabet expects capex to "increase significantly in 2027" [28].
    • Mizuho's channel checks (2026-09-21) point to Amazon Trainium3 ramps in 2027 and widening work at Microsoft, Meta and Oracle (secondary) [33].
  • Why it matters: it underwrites the FY27 second-half step and the FY28 consensus.
  • 😈 Devil's advocate: this is the best-known fact in the market and is fully in consensus. It is support, not edge.
  • Verdict: Medium

Dropped from report #1: "Management incentives point at a much larger company". The CEO's discretionary sale outside a plan (Bear #10) undercuts it, and on its own it was Weak.

Bear case

🐻 Bear #1A few buyers, no contracts, $4.2M of backlog

  • The point: three or four buyers decide Credo's revenue, and none of them is contractually committed.
  • Evidence:
    • Contracting parties A and B were 43% and 28% of Q1 revenue. They hold 57% and 28% of receivables [1].
    • The top four end customers were 84% of revenue [4]. The top 10 were about 90% of FY26 [3].
    • Remaining performance obligations were only $4.2M. Orders can be cancelled or rescheduled on short notice, with no minimum purchases [1,3].
    • Precedent: in February 2023, the largest customer (then ~30% of sales) cut its forecast. Quarterly revenue fell from $54.3M to $32.1M (−41%), and the stock roughly halved in a day [40,41].
  • Why it matters: a lost next-generation design at one top customer could remove 25–30% of revenue within two quarters.
  • 🛡️ Counter-argument:
    • Revenue is now 7x the 2023 base, with four 10% customers rather than one.
    • Credo's links are qualified across "five of six" hyperscalers [5].
    • FY24 revenue still grew 4.8% after the 2023 cut [41].
  • Severity: High Diversification is real but slow, and the backlog gives no visibility.

🐻 Bear #2At 200G per lane the copper reach shrinks

  • The point: at 200G per lane, copper reach shortens, and the industry is openly planning optical alternatives for in-rack links.
  • Evidence:
    • Credo's own 1.6T AEC reaches 6.5 m, against 7 m at 800G [4].
    • Marvell claims 2.5 m copper reach at 200G per lane, presumably for passive copper; that is not stated [21].
    • Nvidia will offer its 2028 Feynman platform with "copper or CPO NVLink" and has not said which [25].
    • Astera expects NPO in 2027 and CPO from 2028 [24].
    • Nvidia invested $2B each in Coherent and Lumentum and $2B in Marvell for optical I/O in March 2026 [27,46].
  • Why it matters: AECs are the largest and highest-margin business. A platform shift would hit FY29 revenue and margin together.
  • 🛡️ Counter-argument:
    • Broadcom said its custom-chip customers stay on copper inside racks through 2028 [47].
    • The Kyber copper midplane rack has reportedly slipped to 2028. That is reported by analysts, not confirmed by Nvidia, and pushes any change further out [26].
    • Credo sells optics, ALCs and NPO/CPO parts, so it can win either way. NPO/CPO design wins ramp in FY28 [4].
  • Severity: Medium High. It is a 2028–29 risk, not a 2027 one. New since report #1: Marvell now states the copper limit publicly as its sales pitch.

🐻 Bear #3Astera and Marvell are building ready-made second sources

  • The point: Credo's AEC edge is know-how, and rivals package that know-how for cable makers and hyperscalers.
  • Evidence:
    • Astera's 800G Taurus is in pre-production, with "multiple platforms" designed in at its lead hyperscaler. Astera sizes the Taurus opportunity at over $4B by 2030 [24].
    • Astera's non-GAAP gross margin is 73.7%, and it also has an Amazon warrant agreement [23].
    • Marvell's kit lets Foxconn Interconnect design a 1.6T AEC in about two months [20].
    • Hyperscalers dual-source by habit, and Credo's 10-K allows that it may grant "exclusivity or most favored nation pricing" [3].
  • Why it matters: a qualified second source at 1.6T caps Credo's price and share at the very generation that drives FY28.
  • 🛡️ Counter-argument:
    • No production win for either rival is found as of 2026-10-12 [21,22,23].
    • Credo's non-GAAP gross margin held at 68.0% in Q1 [2].
    • Credo sells a tested link with PILOT telemetry, not just a chip.
  • Severity: Medium High. The threat is closer than a year ago: Astera has moved from sampling to pre-production.

🐻 Bear #4Profit converts to cash poorly as working capital swells

  • The point: reported profits are running well ahead of cash.
  • Evidence:
    • Q1 FY27 non-GAAP net income was $236.3M against FCF of $82.9M [2,1].
    • Inventory rose from $90.0M to $313.1M over five quarters (3.5x), while quarterly revenue rose 2.1x [1].
    • Receivables were $288.8M, against $162.1M a year earlier [1].
    • Refundable capacity deposits are $88.4M today, with another $102.4M due under a new agreement. Purchase obligations are $222.2M [1].
  • Why it matters: Credo is pre-buying supply for a second half that has not been ordered (Bear #1). If demand slips, inventory write-downs and margin pressure follow.
  • 🛡️ Counter-argument:
    • The build is deliberate for a guided H2 ramp. Brennan: "We are leaning in from a supply chain standpoint" [4].
    • Excess-and-obsolete write-downs were only $3.9M [1].
    • FCF was 34–41% of revenue in Q3–Q4 FY26 [1,3].
  • Severity: Medium This is new this report. It becomes High if Q2 FCF is again below 20% while revenue growth slows.

🐻 Bear #5Stock pay is now ~18–20% of revenue

  • The point: the non-GAAP numbers exclude a large and rising real cost.
  • Evidence:
    • SBC was $88.0M in Q1 FY27 (18.4% of revenue), against $35.5M (15.9%) a year earlier [1][derived].
    • The Q2 guide implies about $104M (~19–20%) [2][derived].
    • Shares outstanding rose 8.7% in a year to 187.95M, from the $750M ATM (4.82M shares at ~$155.6), the Amazon warrant (3.76M) and DustPhotonics (0.76M) [1,3,12].
    • Overhang: 13.1M unvested RSUs, PSUs and options (~7%). There is also the DustPhotonics earn-out of up to ~2.8M shares plus $31.6M, and the CEO's 1.44M-share PSU [1,3,15].
    • Tax withheld on vesting rose to $25.9M a quarter, from $3.7M [1].
  • Why it matters: GAAP diluted EPS was $0.67 against $1.20 non-GAAP in Q1 [2]. At ~$400M of SBC a year, that is ~1% of market cap annually.
  • 🛡️ Counter-argument:
    • The ATM is fully used, and management has "no current plans" to raise [3,5].
    • The ATM and the Amazon warrant were one-offs, so the run-rate dilution is ~2–3% a year from awards.
  • Severity: Medium Low.

🐻 Bear #6FY27 needs a ~45% bigger second half

  • The point: the FY27 guide relies on a steep, unproven H2 step.
  • Evidence:
    • FY26 revenue was $1,335.1M, so >85% growth means at least ~$2,470M for FY27 [3,4].
    • Q1 ($479M) plus the Q2 midpoint ($530M) is $1,009M. H2 therefore needs about $1,461M, ~45% more than H1 [derived].
    • The CFO's path is reported as about +20% sequential in Q3 and +30% in Q4 (secondary) [50].
    • An analyst on the Q1 call put the H2 increment at about $500M [4].
    • Consensus FY27 ($2.50B) sits just above the floor [34].
  • Why it matters: H2 depends on the newest products (1.6T, optics, PICs), all on tight 3nm and laser supply.
  • 🛡️ Counter-argument:
    • Management has beaten 8 of 8 guides [Bull #2].
    • It has pre-paid capacity for exactly this ramp [1].
    • FY26 delivered a similar back-half step: Q3 FY26 was +52% sequential [2,8].
  • Severity: Medium

🐻 Bear #7Optics is new, laser-constrained and not in the risk factors

  • The point: the second engine runs on a supply chain Credo has never scaled, and its own filings don't describe the risk.
  • Evidence:
    • Indium-phosphide lasers are short: Lumentum says EML capacity is committed through 2027, and Coherent's growth is constrained [27].
    • The FY26 10-K and the Q1 10-Q have no risk factor on DustPhotonics, silicon photonics, lasers, optical module manufacturing or Israel. The 10-Q says there are "no material changes" [1,3].
    • Goodwill plus intangibles rose to $1,365M, 45% of total assets (from $122M) [1].
    • The DustPhotonics earn-out metrics are undisclosed ("certain financial milestones") [3].
  • Why it matters: a miss on the >$600M optics guide would hit the H2 plan and the "two-engine" story at the same time. It could also lead to an impairment test.
  • 🛡️ Counter-argument:
    • Credo says the Dust design uses "substantially fewer lasers" [5].
    • ZF optics are shipping, and first PIC revenue arrived on schedule [4].
  • Severity: Medium

🐻 Bear #8A capex pause would show up within one quarter

  • The point: Credo's growth depends on a few companies continuing to raise AI spending.
  • Evidence:
    • Customers can cancel or reschedule on short notice, and backlog is $4.2M [1,3].
    • The 2023 forecast cut took 41% off quarterly revenue within one quarter [40].
    • Alphabet's FCF turned negative in Q2 2026, and Meta's fell sharply (tracker, secondary) [28].
  • Why it matters: with purchase-order-only sales, a pause hits revenue immediately, while ~$500M of inventory and deposits is already committed [1].
  • 🛡️ Counter-argument: every large spender guided capex up for 2026 and 2027 [28]. Supply, not demand, is the stated constraint in lasers and 3nm [27].
  • Severity: Medium

🐻 Bear #9Expectations and a beta of 3.25 make every print binary

  • The point: small disappointments cost a third of the value.
  • Evidence:
    • The stock fell 20% on 2026-09-02 after a beat-and-raise, citing GAAP margin, opex and the optics guide not being raised [37].
    • It fell 50% from close to close between 2026-06-22 and 2026-09-14 [44].
    • It fell ~6% on 2026-10-12 on an OCP product reveal [31].
    • Beta is about 3.25. Short interest is 6.6% of float (11.28M shares, up from 8.61M a month earlier) [35].
  • Why it matters: this is a risk to holders more than to the business. Timing of any position matters a lot.
  • 🛡️ Counter-argument: the multiple has already compressed to ~28x forward (Bull #4), and short interest is low with 1.2 days to cover [35].
  • Severity: Medium

🐻 Bear #10Insiders only sell, including one CEO sale outside a plan

  • The point: no insider buys, and the CEO sold outside a pre-set plan.
  • Evidence:
    • Zero open-market purchases in Form 4s from April to October 2026 [18].
    • CEO Brennan sold 54,984 shares at $218.6 ($12.0M) on 2026-06-05, with the 10b5-1 box unchecked. He adopted a 440,000-share plan five days later. Its first sales were at ~$151.9 on 2026-09-14 [16,17,1].
    • CTO Cheng has sold 581k shares ($114M) since March. COO Lam sold 361k ($77M), including 150k at ~$205–206 in late September and early October. Both sold under 10b5-1 plans [18].
    • CLO Laufman made non-plan sales at $185 and $205 in September–October [18].
  • Why it matters: this is a mild signal. A discretionary CEO sale right after earnings is the one data point that is not mechanical.
  • 🛡️ Counter-argument:
    • The founders still hold millions of shares each: Cheng ~5.6M, Lam ~2.25M plus trusts [18].
    • The selling ran at $150 as well as at $300, which is consistent with diversification rather than timing.
  • Severity: Low Report #1's open question is answered: the June sale was outside a plan.

🐻 Bear #11One foundry, Taiwan, China R&D and an unresolved export issue

  • The point: the supply chain and the R&D base sit in politically exposed places.
  • Evidence:
    • TSMC is the exclusive foundry, and all 200G-per-lane parts are on 3nm, where supply is tight [3,5].
    • BizLink assembles the AECs. Switching suppliers takes 9–12 months [3].
    • 28% of FY26 revenue was billed in Hong Kong. Credo has R&D in mainland China, Hong Kong and Taiwan, with 477 of 807 staff in Asia [3].
    • The 10-K says some past exports "may have been made in violation of Trade Controls laws". No outcome or penalty has been disclosed in any filing [3].
    • DustPhotonics brings Israeli operations, and the 10-K cites "armed conflict in Israel" [3].
    • Pillar Two minimum tax may apply from FY28, against a Q1 effective rate of −0.5% [1,3].
  • Why it matters: a Taiwan event, a tightened US outbound-investment rule (revision expected in 2027) or a 232 tariff without the data-centre exemption would hit supply or costs.
  • 🛡️ Counter-argument:
    • Mainland China revenue is down to $1.1M a quarter [1].
    • The Section 232 chip tariffs exempt US data centres [43].
    • Every fabless peer shares the TSMC profile.
  • Severity: Low Medium. Pillar Two is the only quantifiable item, and it is not in the FY28 consensus (see headwinds).

Tailwinds NOT in the forecasts

Base: consensus FY28 (ends ~May 2028) revenue $3.87B and EPS $9.69; FY29 revenue $5.07B and EPS $12.45 (secondary) [34,50]. EPS maths uses ~200M diluted shares [4]. The incremental net margin is an estimate: ~55% on copper/retimers (68% gross margin, opex largely fixed, ~13% tax) and ~45% on optics.

  • Scale-up Ethernet becomes a volume socket for retimers and short AECs — 36% chance by FY28

    • Impact on revenue: +$300M, +7.8% vs consensus FY28
    • Impact on EPS: +$0.83, +8.5% vs consensus FY28
    • How we got the number (estimate): $300M × 55% incremental net margin ÷ 200M shares ≈ +$0.83
    • Expected value: 36% × $0.83 ≈ +$0.30
    • What it is: custom-ASIC pods (UALink, ESUN) and merchant GPUs use Credo's Blue Heron and Toucan retimers and in-rack AECs, per accelerator.
    • Why it's not priced in: management gives no figure beyond "more substantial" in FY28 [5]. The FY28 consensus of +55% reads as a continuation of AECs plus optics [34].
    • Trigger to watch: a scale-up revenue figure on the Q2 or Q3 FY27 call; first UALink/ESUN silicon in 2027 [42].
    • Why this probability: revenue already exists and Astera agrees in-rack scale-up stays copper [24]. But Marvell's 2.5 m copper claim at 200G [21] puts the cross-rack part in doubt. I cut it from 42% to 36%.

  • Optics clearly beats >$600M and steps up again in FY28 — 33% chance by FY28

    • Impact on revenue: +$350M, +9.0% vs consensus FY28
    • Impact on EPS: +$0.79, +8.1% vs consensus FY28
    • How we got the number (estimate): $350M × 45% incremental net margin ÷ 200M ≈ +$0.79
    • Expected value: 33% × $0.79 ≈ +$0.26
    • What it is: ZF optics ramps at a second hyperscaler, the two PIC wins ramp early, and 1.6T DSPs ship in volume.
    • Why it's not priced in: the >$600M guide was held, not raised, in September, and the stock fell 20% partly on that [37]. Brennan's "more than double or triple" comment is not in numbers.
    • Trigger to watch: the optics guide raised at Q2 (early December 2026); a named ZF optics customer.
    • Why this probability: management beats revenue guides, but the laser shortage persists past 2027 [27]. Optics is the part of the plan it has never executed.

  • A hyperscaler win for ALC, the optical backplane or Weaver — 24% chance by FY29

    • Impact on revenue: +$250M, +4.9% vs consensus FY29
    • Impact on EPS: +$0.63, +5.0% vs consensus FY29
    • How we got the number (estimate): ~100k GPUs × $2,500 Weaver content, or an equivalent ALC volume, = $250M × 50% ÷ 200M ≈ +$0.63
    • Expected value: 24% × $0.63 ≈ +$0.15
    • What it is: one of the FY28 product lines lands a top-four customer.
    • Why it's not priced in: none has revenue. The market sold the 2026-10-12 ALC reveal [31].
    • Trigger to watch: customer commentary after OCP; ALC sampling news; a second Weaver customer beyond Positron.
    • Why this probability: there are three shots on goal, but each is a new product category, and Avicena is competing for microLED links [45].

  • A US rule on Chinese optical transceivers shifts share to Credo's modules — 14% chance within 36 months

    • Impact on revenue: +$150M, +3.9% vs consensus FY28
    • Impact on EPS: +$0.34, +3.5% vs consensus FY28
    • How we got the number (estimate): $150M × 45% ÷ 200M ≈ +$0.34
    • Expected value: 14% × $0.34 ≈ +$0.05
    • What it is: the FCC drafts a Covered-List measure on new Chinese transceiver models. Chinese suppliers hold ~56–60% of the market [29].
    • Why it's not priced in: the rule is unpublished, existing models are reportedly grandfathered, and nobody quantifies a benefit [29].
    • Trigger to watch: the FCC publishes rule text; the target is end-2026 [29].
    • Why this probability: reported but unpublished, and hyperscalers would lobby against a supply squeeze during a laser shortage.

Headwinds NOT in the forecasts

  • Hyperscaler capex growth stalls in 2027–28 — 17% chance by FY28

    • Impact on revenue: −$870M, −22% vs consensus FY28 (growth ~+20% instead of +55%)
    • Impact on EPS: −$2.61, −27% vs consensus FY28
    • How we got the number (estimate): $870M × 60% decremental net margin (opex and deposits are fixed) ÷ 200M ≈ −$2.61
    • Expected value: 17% × −$2.61 ≈ −$0.44
    • What it is: two or more of the top four spenders digest in 2027–28, with Credo's purchase-order book cut within a quarter.
    • Why it's not priced in: consensus assumes +55% in FY28 [34], and all spenders guide up [28].
    • Trigger to watch: Q3 calendar-2026 hyperscaler calls (late October 2026); 2027 capex guides (January–February 2027).
    • Why this probability: 2027 guides are up, but hyperscaler FCF strain is visible [28], and the 2023 precedent shows how fast it hits [40].

  • A top-three customer dual-sources 1.6T AECs with Astera or a Marvell-based cable — 26% chance by FY28

    • Impact on revenue: −$400M, −10.3% vs consensus FY28
    • Impact on EPS: −$1.20, −12.4% vs consensus FY28
    • How we got the number (estimate): 40% of one ~28% customer's FY28 AEC volume ($1.0B), plus ~5% AEC price pressure, ≈ −$400M × 60% ÷ 200M ≈ −$1.20
    • Expected value: 26% × −$1.20 ≈ −$0.31
    • What it is: Astera's Taurus or a Golden Cable AEC qualifies at a top-three customer for 1.6T.
    • Why it's not priced in: consensus FY28 gross margin and growth assume Credo keeps share [34]. No rival win has been announced.
    • Trigger to watch: Astera's Q3 call (2026-11-03) moving Taurus from pre-production to production; Credo's top customer falling below ~28%.
    • Why this probability: Astera is designed into multiple platforms at a lead hyperscaler [24], and hyperscalers dual-source by habit. Credo's 800G incumbency and the 9–12 month requalification [3] slow it. I lowered it from 30% to 26% because there is still no win ten months into Golden Cable.

  • A major platform chooses CPO/NPO over AECs for in-rack links — 22% chance by FY29

    • Impact on revenue: −$600M, −11.8% vs consensus FY29
    • Impact on EPS: −$1.80, −14.5% vs consensus FY29 (partly offset if Credo's NPO parts win; not netted)
    • How we got the number (estimate): ~25% of a ~$2.4B FY29 AEC base lost = −$600M × 60% ÷ 200M ≈ −$1.80
    • Expected value: 22% × −$1.80 ≈ −$0.40
    • What it is: Nvidia's Feynman (2028) or a hyperscaler custom rack goes optical for scale-up and short scale-out.
    • Why it's not priced in: the FY29 consensus of +31% assumes the copper and optics mix holds [50].
    • Trigger to watch: GTC March 2027 architecture details; Marvell's scale-up optics revenue in 2027; Astera's NPO deployments in 2027 [24].
    • Why this probability: Nvidia keeps copper as one option and Broadcom stays copper through 2028 [47]. But three major suppliers are now selling the optical alternative [21,24,25].

  • AEC price competition pushes non-GAAP gross margin to ~64% — 28% chance by FY28

    • Impact on revenue: −$100M, −2.6% vs consensus FY28 (price only)
    • Impact on EPS: −$0.67, −6.9% vs consensus FY28
    • How we got the number (estimate): 4 pts × $3.87B = −$155M gross profit × 87% after tax ÷ 200M ≈ −$0.67
    • Expected value: 28% × −$0.67 ≈ −$0.19
    • What it is: second sources plus a lower-margin optics mix push margin down toward the 63–65% long-term model management itself published [6].
    • Why it's not priced in: guidance says FY27 gross margin is "broadly consistent" with FY26's 68.1% [4].
    • Trigger to watch: the Q3 gross margin guide at or below 67%; optics share of revenue disclosed.
    • Why this probability: management's own long-term model is 63–65% [6], and Credo is running above it.

  • Pillar Two minimum tax raises the FY28 tax rate — 55% chance in FY28

    • Impact on revenue: none
    • Impact on EPS: −$0.30, −3.1% vs consensus FY28
    • How we got the number (estimate): ~$2.1B FY28 pre-tax non-GAAP profit (estimate) × ~3 pts higher effective rate ≈ −$63M ÷ 200M ≈ −$0.32, rounded to −$0.30
    • Expected value: 55% × −$0.30 ≈ −$0.17
    • What it is: the 10-K says Credo "may be subject to" the OECD 15% regime from FY28. Q1 FY27's effective rate was −0.5% [1,3].
    • Why it's not priced in: no quantification was found in guidance or in the consensus write-ups. Whether analysts model it is not found.
    • Trigger to watch: a FY28 tax-rate comment at Q4 FY27 (June 2027).
    • Why this probability: the 10-K language is conditional ("may"), but Credo's Cayman structure is the kind of setup the regime targets. The non-GAAP rate already used by analysts is not found, so the delta is uncertain.

  • The laser shortage caps the FY27 optics ramp — 25% chance by FY27

    • Impact on revenue: −$120M, −4.8% vs consensus FY27
    • Impact on EPS: −$0.27, −4.3% vs consensus FY27
    • How we got the number (estimate): $120M × 45% ÷ 200M ≈ −$0.27
    • Expected value: 25% × −$0.27 ≈ −$0.07
    • What it is: ZF optics ships below plan because of indium-phosphide laser supply.
    • Why it's not priced in: the >$600M guide is reiterated, and the consensus sits at the >85% floor [4,34].
    • Trigger to watch: optics commentary at Q2; Lumentum and Coherent laser-capacity updates [27].
    • Why this probability: the shortage is real, but Credo designs for fewer lasers and has pre-paid capacity [1,5].

  • An export-control finding, outbound-rule tightening or a Taiwan event — 12% chance within 36 months

    • Impact on revenue: not quantifiable
    • Impact on EPS: not quantifiable
    • How we got the number: no basis. Credo has not disclosed the scope of the past "apparent violations", and the 2027 outbound-rule revision is unwritten [3].
    • Expected value: n/a
    • What it is: a BIS penalty, a forced restructuring of China R&D, or a Taiwan supply disruption.
    • Why it's not priced in: guidance assumes the "current tariff regime" [4].
    • Trigger to watch: the outbound-investment rule text in 2027; any 8-K on trade controls.
    • Why this probability: policy is volatile, but chips for US data centres have so far been carved out [43].

Investment Risk Score: 37/100 (moderate)

# Factor Weight Measured value Score Why
1 Liquidity & funding runway 15% FCF-positive; LTM FCF $438.5M; cash + short-term investments $764.3M, no debt; ~$102M new deposits due in FY27 10 Top band. Not 5, because Q1 FCF fell to 17% of revenue and deposit calls continue.
2 Leverage 10% Net cash $764M; leases ~$26M 5 No debt of any kind.
3 Dilution 10% Shares +8.7% y/y (187.95M); overhang ~13.1M awards (7%) + up to 2.8M earn-out shares (1.5%) + 1.44M CEO PSUs; SBC 18.4% of revenue 52 Mid-band (+3–10%). The one-offs (ATM, Amazon warrant) are done, but SBC at ~$400M a year keeps ~2–3% a year coming.
4 Profitability 10% LTM FCF margin 27.6%; Q1 FCF margin 17.3%; GAAP operating margin 25.2% 15 Top band on FCF. Pushed up from 10 because SBC (18% of revenue) is excluded from FCF.
5 Revenue quality & concentration 10% Contracting party A 43%, B 28%; top-four end customers 84%; RPO $4.2M; purchase orders only 68 ">35% top customer" band. Kept below 75 because there are four 10% customers and revenue is repeat in practice.
6 Execution & track record 10% Revenue beat the top of the guide 8/8; opex above guide 3 of the last 3; no restatement or material weakness; EY since 2018 22 Hits guidance ≥80% on revenue. Small cost-guide misses and the GAAP guide ignoring Dust keep it out of 0–20.
7 Competitive position & disruption 10% AEC leader, no rival in volume; Astera pre-production at 800G; Marvell pitching the "copper wall" at 200G; Nvidia Feynman CPO option 48 Contested. A strong position today against a well-funded second-source effort and a substitution debate for 2028+.
8 Cyclicality & macro sensitivity 10% ~100% AI data-centre capex; 2023 customer cut −41% q/q; ~$500M of inventory and deposits committed; beta 3.25 62 Highly cyclical in practice. One buyer's pause can take ~30% off a quarter.
9 Regulatory, legal & geopolitical 5% TSMC sole source, Taiwan packaging; 28% billed via Hong Kong; China R&D; undisclosed past export "apparent violations"; Israel; Pillar Two FY28; no material litigation 50 Material but contained. No pending ruling, but China/Taiwan exposure is structural.
10 Valuation & expectations 10% ~$205; ~28x forward and 21x FY28 consensus EPS; EV/sales 25x trailing; short interest 6.6%; −20% on a beat-and-raise 55 Rich. Cheaper than ALAB, MRVL and LITE, but consensus assumes +87% and +55% growth.

Overall: weighted average = 1.5 + 0.5 + 5.2 + 1.5 + 6.8 + 2.2 + 4.8 + 6.2 + 2.5 + 5.5 = 36.7 → 37.

Change vs last: −14 vs 51 (old 6-area method). It is not like-for-like. The old method scored "dependency & concentration" at 75 with a 15% weight and folded cyclicality into business risk. The v3 method gives liquidity 15% weight and separates leverage and profitability, where Credo scores 5–15. New evidence moved two inputs:

  • Valuation is lower: the multiple compressed from ~44x to ~28x.
  • Cash-conversion and SBC quality is worse: Q1 FCF 17%, SBC 18.4%.
  • Interpretation: on a like-for-like basis the risk is about unchanged.

Ranking check: At 37, CRDO sits just above AVGO and GRAB (34). It is riskier than AVGO, with the same AI-capex exposure but one product line and 43% from one buyer instead of a diversified franchise. It is safer than VIST (41), which depends on a single country and a single commodity with dilution, while Credo has net cash and ~28% FCF margins.

Conclusion

  • What you are really betting on:
    • Credo keeps the default short-reach link through the 1.6T generation.
    • Optics becomes a second engine before Astera, Marvell or CPO erode the AEC franchise.
    • All of this has to happen while three or four buyers keep raising capex without any contract obliging them to buy from Credo.
  • The 2–3 things that will decide the outcome:
    1. 1.6T share at the top three customers, where Astera's Taurus is the first credible rival.
    2. Whether the FY27 second half (~$1.46B, ~45% above H1) arrives, with optics above $600M despite the laser shortage.
    3. Hyperscaler 2027–28 capex. One pause would hit within a quarter.
  • Signposts:
    • Confirm the bull case:
      • Q2 FY27 (early December) beats $525–535M, and the optics guide is raised.
      • Q2 FCF recovers above 30% of revenue as inventory growth slows.
      • A quantified scale-up revenue figure for FY28.
      • A hyperscaler is named for ALC or ZF optics.
    • Break the bull case:
      • Astera announces Taurus production volume at a top-three Credo customer, or Credo's top customer falls below ~28% with flat revenue.
      • The non-GAAP gross margin guide falls to 67% or lower.
      • Q2 FCF is again below 20% of revenue with inventory still rising.
      • Nvidia or a hyperscaler commits to CPO/NPO for in-rack scale-up in a 2027–28 platform.
  • New since the last report:
    • The competitive check came back clean for now. Astera is pre-production at 800G, and Marvell's Investor Day showed no AEC win. Marvell is now openly selling the "copper wall" at 200G per lane, which sharpens Bear #2.
    • Cash quality is the new concern. Q1 FCF was $82.9M against $236M non-GAAP net income, inventory is up 3.5x in five quarters, a new $102.4M deposit is due, and SBC is ~18% of revenue.
    • The June CEO sale was outside a 10b5-1 plan. The 10-K and 10-Q added no optics, laser or Israel risk factor. The DustPhotonics earn-out is up to ~2.8M shares plus $31.6M on undisclosed milestones.

Open questions & data gaps

  • Identity and share of the top three end customers: not disclosed. The customer letters change between filings, and no source assigns percentages to named companies. This matters because the identities decide which capex plan and which Astera fight count most.
  • Optical revenue in Q1 FY27: not disclosed. This matters for testing the >$600M guide before December.
  • DustPhotonics earn-out milestones and period: not found. The share purchase agreement is not among the exhibits seen. This affects dilution and the scale of a possible impairment.
  • Past export-control "apparent violations": there is no BIS self-disclosure outcome in any filing or search. Low to medium importance.
  • Why the 650 Group's $644M 2025 AEC market is below Credo's AEC revenue: the methodology was not found.
  • Pillar Two: the expected FY28 rate and whether consensus includes it were not found. The estimate in the headwind is labelled as an estimate.
  • The 2026-06-23 drop: report #1 linked it to CPO worries via one secondary source. This run found no source confirming that cause, and the link should be treated as unconfirmed.
  • Consensus FY29 comes from a Substack citing Koyfin (secondary).
  • Valuation multiples differ by aggregator. Peer multiples are stockanalysis.com as of 2026-10-12.
  • Call transcripts are third-party (Motley Fool, stockanalysis.com, Benzinga), not company-issued. The Q2 FY26 (December 2025) transcript could not be obtained.
  • Whether Marvell's 2.5 m figure at 200G per lane refers to passive copper was not confirmed. Credo's own 1.6T AEC claim is 6.5 m.
Sources
  1. Credo Form 10-Q, quarter ended 2026-08-01 — SEC EDGAR — 2026-09-02 — https://www.sec.gov/Archives/edgar/data/1807794/000162828026060111/crdo-20260801.htm
  2. Credo Q1 FY27 results press release (8-K Ex. 99.1) — SEC EDGAR — 2026-09-01 — https://www.sec.gov/Archives/edgar/data/1807794/000162828026059795/credoq12027ex-991.htm
  3. Credo Form 10-K, fiscal year ended 2026-05-02 — SEC EDGAR — 2026-06-15 — https://www.sec.gov/Archives/edgar/data/1807794/000162828026043303/crdo-20260502.htm
  4. Credo (CRDO) Q1 2027 earnings call transcript (third-party) — Motley Fool — 2026-09-08 (call 2026-09-01) — https://www.fool.com/earnings/call-transcripts/2026/09/08/credo-crdo-q1-2027-earnings-call-transcript/
  5. Credo (CRDO) Q4 2026 earnings call transcript (third-party) — Motley Fool — 2026-06-01 — https://www.fool.com/earnings/call-transcripts/2026/06/01/credo-crdo-q4-2026-earnings-transcript/
  6. Credo (CRDO) Q3 2026 earnings call transcript (third-party) — Motley Fool — 2026-03-02 — https://www.fool.com/earnings/call-transcripts/2026/03/02/credo-tech-crdo-q3-2026-earnings-call-transcript/
  7. Credo Q4 and FY26 results press release (8-K Ex. 99.1) — SEC EDGAR — 2026-06-01 — https://www.sec.gov/Archives/edgar/data/1807794/000162828026039474/credoq42026ex-991.htm
  8. Credo Q3 FY26 results press release — SEC EDGAR — 2026-03-02 — https://www.sec.gov/Archives/edgar/data/1807794/000162828026013205/credoq32026ex-9911.htm
  9. Credo Q2 FY26 results press release — SEC EDGAR — 2025-12-01 — https://www.sec.gov/Archives/edgar/data/1807794/000162828025054428/credoq22026ex-991.htm
  10. Credo Q1 FY26 results press release — SEC EDGAR — 2025-09-03 — https://www.sec.gov/Archives/edgar/data/1807794/000180779425000018/credoq12026ex-991.htm
  11. Credo Q1–Q4 FY25 results press releases — SEC EDGAR — 2024-09-04 to 2025-06-02 — https://www.sec.gov/Archives/edgar/data/1807794/000162828024039309/credoq12025ex-991.htm
  12. Prospectus supplement, $750M at-the-market offering (424B5) — SEC EDGAR — 2025-10-06 — https://www.sec.gov/Archives/edgar/data/1807794/000162828025044356/credo-prospectussupplement.htm
  13. "Credo Agrees to Acquire DustPhotonics" (8-K Ex. 99.1) — SEC EDGAR — 2026-04-13 — https://www.sec.gov/Archives/edgar/data/1807794/000162828026024892/april20268-kex991.htm
  14. Credo 2026 proxy statement (DEF 14A) — SEC EDGAR — 2026-08-25 — https://www.sec.gov/Archives/edgar/data/1807794/000162828026058936/crdo-20260825.htm
  15. Credo 8-K Item 5.02, CEO special performance award — SEC EDGAR — 2026-06-01 — https://www.sec.gov/Archives/edgar/data/1807794/000162828026039474/crdo-20260528.htm
  16. Form 4, William J. Brennan (sale 2026-06-05, no 10b5-1) — SEC EDGAR — 2026-06-09 — https://www.sec.gov/Archives/edgar/data/1807794/000162828026041881/wk-form4_1781035389.xml
  17. Form 4, William J. Brennan (10b5-1 plan sale 2026-09-14) — SEC EDGAR — 2026-09-16 — https://www.sec.gov/Archives/edgar/data/1807794/000162828026062295/wk-form4_1789591996.xml
  18. Credo Form 4 filings, April–October 2026 (submissions index) — SEC EDGAR — retrieved 2026-10-12 — https://data.sec.gov/submissions/CIK0001807794.json
  19. BlackRock Schedule 13G/A on Credo (7.7%) — SEC EDGAR — 2026-07-27 — https://www.sec.gov/Archives/edgar/data/1807794/000201238326002151/primary_doc.xml
  20. "Marvell Launches Strategic Initiative to Accelerate AEC Ecosystem and Hyperscaler Adoption" — Marvell — 2025-12-09 — https://investor.marvell.com/news-events/press-releases/detail/1002/marvell-launches-strategic-initiative-to-accelerate-aec-ecosystem-and-hyperscaler-adoption
  21. "Marvell Investor Day 2026 Targets $80B FY2031 Revenue With Custom Silicon and Optics" — Futurum — 2026-10-06/07 — https://futurumgroup.com/?p=96506
  22. Marvell Technology Investor Day 2026 transcript (third-party) — stockanalysis.com — 2026-10-06 — https://stockanalysis.com/stocks/mrvl/transcripts/724088-investor-day-2026/
  23. "Astera Labs Reports Second Quarter 2026 Financial Results" (8-K Ex. 99.1) — Astera Labs / SEC EDGAR — 2026-08-04 — https://www.sec.gov/Archives/edgar/data/0001736297/000173629726000033/q226exhibit991.htm
  24. Earnings call transcript: Astera Labs Q2 2026 — Investing.com — 2026-08-04 — https://www.investing.com/news/transcripts/earnings-call-transcript-astera-labs-tops-q2-2026-eps-forecast-shares-swing-93CH-4836051
  25. Nvidia optical scale-up at GTC 2026 — The Register — 2026-04-05 — https://www.theregister.com/2026/04/05/nvidia_optical_scale_up/
  26. "Nvidia's Kyber AI rack slips to 2028" (SemiAnalysis report, relayed) — TheNextWeb — 2026-07-06 — https://thenextweb.com/news/nvidia-kyber-rack-delay-2028
  27. "Lumentum CEO says the indium phosphide shortage will become worse than memory" — Tom's Hardware — 2026 — https://www.tomshardware.com/tech-industry/semiconductors/lumentum-ceo-says-the-indium-phosphide-shortage-will-become-worse-than-memory
  28. "Follow the CAPEX: Q2 2026 Scoreboard" (tracker, secondary) — Platformonomics — 2026-07/08 — https://platformonomics.com/?p=41900
  29. "FCC reportedly drafting ban on Chinese optical transceivers" (Reuters-based) — eWeek / TrendForce — 2026-08-05 — https://www.trendforce.com/presscenter/news/20260805-13169.html
  30. "Credo to Showcase Broad Portfolio of Connectivity Solutions for AI Scale Out and Scale Up Fabrics at 2026 OCP Global Summit" — Credo / Business Wire via Finviz — 2026-10-09 — https://finviz.com/news/400883/credo-to-showcase-broad-portfolio-of-connectivity-solutions-for-ai-scale-out-and-scale-up-fabrics-at-2026-ocp-global-summit
  31. "Why is Credo Technology stock sliding today?" (AI-assisted, editor-reviewed) — Investing.com — 2026-10-12 — https://www.investing.com/news/stock-market-news/why-is-credo-technology-stock-sliding-today-93CH-4944844
  32. "Stifel cuts Credo Tech stock price target on valuation" — Investing.com — 2026-10-05 — https://www.investing.com/news/analyst-ratings/stifel-cuts-credo-tech-stock-price-target-on-valuation-93CH-4932163
  33. "Why is Credo Technology stock rallying today?" (Mizuho, Rosenblatt notes) — Investing.com — 2026-09-21 — https://www.investing.com/news/stock-market-news/why-is-credo-technology-stock-rallying-today-93CH-4909330
  34. CRDO stock forecast and consensus estimates (S&P Global data) — stockanalysis.com — 2026-10-12 — https://stockanalysis.com/stocks/crdo/forecast/
  35. CRDO and peer statistics (short interest, multiples) — stockanalysis.com — 2026-10-12 — https://stockanalysis.com/stocks/crdo/statistics/
  36. "Credo Has One Number That Could Change the Stock's Long-Term Story" (estimate revisions) — 24/7 Wall St — 2026-10-05 — https://247wallst.com/investing/2026/10/05/credo-has-one-number-that-could-change-the-stocks-long-term-story/
  37. "Why Credo Technology Stock Plunged 20% Today" — The Motley Fool — 2026-09-02 — https://www.fool.com/investing/2026/09/02/why-credo-technology-stock-plunged-20-today/
  38. "Credo Technology Has Fallen 30% Since June 22. Its Forward Estimates Have Not Followed" — TIKR — 2026-07-26 — https://www.tikr.com/blog/credo-technology-has-fallen-30-since-june-22-its-forward-estimates-have-not-followed
  39. "Credo's $236 Million Adjusted Profit Came With Just $83 Million in Free Cash Flow" — Yahoo Finance — 2026-09 — https://finance.yahoo.com/markets/stocks/articles/credo-236-million-adjusted-profit-154239820.html
  40. "Shares in data center electric cable supplier Credo halve as biggest customer cuts orders" — DatacenterDynamics — 2023-02 — https://www.datacenterdynamics.com/en/news/shares-in-data-center-electric-cable-supplier-credo-halve-as-biggest-customer-cuts-orders/
  41. "Credo Reports Fourth Quarter and Fiscal Year 2024 Financial Results" — Credo IR — 2024-05-29 — https://investors.credosemi.com/news-events/news/news-details/2024/Credo-Technology-Group-Holding-Ltd-Reports-Fourth-Quarter-and-Fiscal-Year-2024-Financial-Results-05-29-2024/default.aspx
  42. "Cornelis to build scale-up interconnect using UALink and ESUN" — HPCwire — 2026-09-14 — https://www.hpcwire.com/aiwire/2026/09/14/cornelis-to-build-scale-up-interconnect-using-ualink-and-esun/
  43. "US President Trump issues Section 232 proclamations on semiconductors and critical minerals" — Baker McKenzie — 2026-01-27 — https://globalimportblog.bakermckenzie.com/2026/01/27/us-president-trump-issues-section-232-proclamations-on-semiconductors-and-critical-minerals/
  44. CRDO daily price history — Yahoo Finance chart data — retrieved 2026-10-12 — https://query1.finance.yahoo.com/v8/finance/chart/CRDO?range=6mo&interval=1d
  45. "Avicena launches the world's first microLED optical interconnect evaluation kit" — Business Wire syndication — 2026-03-12 — https://www.eagletribune.com/region/avicena-launches-the-world-s-first-microled-optical-interconnect-evaluation-kit-for-ai-infrastructure-innovators/article_8c5e53b4-b79c-5905-85e8-7d0a1b7f80f7.html
  46. "Nvidia invests $2bn in Marvell as part of wider NVLink Fusion deal" — DatacenterDynamics — 2026-03-31 — https://datacenterdynamics.com/en/news/nvidia-invests-2bn-in-marvell-as-part-of-wider-nvlink-fusion-deal
  47. "Credo Shares Jump 10% as Broadcom Backs Copper Over Optical Tech" — GuruFocus — 2026-03-05 — https://www.gurufocus.com/news/8681840/credo-shares-jump-10-as-broadcom-backs-copper-over-optical-tech
  48. CRDO price targets and analyst actions — MarketBeat — 2026-10-12 — https://www.marketbeat.com/stocks/NASDAQ/CRDO/forecast/
  49. Credo Q4 2026 earnings call complete transcript (automated) — Benzinga — 2026-06 — https://cdn1.benzinga.com/insights/news/26/06/52927097/credo-technology-group-q4-2026-earnings-call-complete-transcript
  50. "Credo: 50% Consensus Growth for 24x Earnings" (secondary; FY29 consensus via Koyfin, CFO sequential path) — longtermpick (Substack) — 2026-09-18 — https://longtermpick.com/p/credo-analysis
  51. "Credo, Volex reach settlement in active electrical cable patent dispute" (and later TE/Molex settlement reports) — TipRanks/TheFly — 2025-08 (TE/Molex ~2026-03, secondary) — https://www.tipranks.com/news/the-fly/credo-volex-reach-settlement-in-active-electrical-cable-patent-dispute-thefly
All CRDO Bulls vs Bears reports
Disclaimer. AI-generated research and podcast (Claude by Anthropic, voices by Google Gemini) for educational purposes — not financial advice and not a recommendation to buy or sell. Willem (MR.INVEST) owns this stock. Check the company’s own filings before making any decision.
← All Bulls vs Bears reports