Altair
Skip to content

Trading Analysis Report: CRDO

Analysis date: 2026-10-14 · Model: claude-opus-5-5 · AI-generated, not financial advice.

  • Analysis date: 2026-10-14
  • Rating: Underweight
  • Generated: 2026-10-14 15:08:37
  • TradingAgents 0.6.0: anthropic, deep claude-opus-5-5, quick claude-sonnet-5-5
  • Analysts: market, sentiment, news, fundamentals; research debate rounds 5, risk debate rounds 5
  • Data vendors: core_stock_apis yfinance, technical_indicators yfinance, fundamental_data sec_edgar,yfinance, news_data yfinance, macro_data fred, prediction_markets polymarket

V. Portfolio Manager Decision

Rating: Underweight

Executive Summary: Underweight CRDO. Cut exposure to the lower of 0.5x a standard allocation or 2% of the portfolio, cap any ALAB/MRVL/CRWV/NBIS overlap separately as a basket, and if you are flat do not initiate at $179.23. Sell about three-quarters of the intended reduction tomorrow between 10 a.m. and noon, or two-thirds only if the peer test, a clean 10-Q read and a verified event window of five or more sessions all pass. Put the last quarter on a $189 limit live from the open (add $195 only in the long window), and sell anything unfilled if CRDO is below $169 at the 1 p.m. or 3:30 p.m. check, or at expiry (default Friday's close if event dates are unverified); hold the residual at half size unless the filing and guide clear it, take it to zero on a weekly close below the weekly SuperTrend stop (now $159.83), make no add-back before the early-December print, and use a 2-3 month horizon with a $155 target.

Current Price: 179.23

Price Target: 155.0

Confidence: Medium

Time Horizon: 2-3 months

IV. Risk Management Team Decision

Aggressive Analyst

Aggressive Analyst: Nobody else has spoken yet, so I'll set the frame and then take on the objections I expect.

Aggressive doesn't mean long. The boldest move on CRDO right now is refusing to be paid 0.74% in free cash flow yield to carry a stock with a beta of 3.25. We're at about 67x run-rate EPS and 21x trailing sales. The bull case in the research plan only gets to roughly 4.4% a year. Polymarket has a 97% chance of no cuts in 2026 and 86% on another hike, with three-hike odds up almost 10 points this week. In that world you hold a stock that gaps 20% on a bad day for under 5% a year of upside. That isn't conviction. It's donating the risk premium.

The fundamentals are also turning, and at this multiple that matters. Sequential growth went from 52% to about 7% to about 10%. Gross margin dropped 410 basis points with no explanation. Operating income fell from $150M to $121M even though revenue rose from $407M to $479M. Operating cash flow was only about 70% of net income, and TTM free cash flow is under half of earnings. The Zacks headline cites a $525M to $535M outlook. If that's next-quarter guidance, which the headline doesn't confirm, it's roughly 10 to 12% sequential growth. That's the new plateau, not a re-acceleration, and a 67x multiple prices in re-acceleration.

Now the objections.

The conservative will probably say the plan is too timid, and that if the stock is this risky we should dump everything at market. I disagree. ADX is 13.5, so there is no confirmed trend. The DI cross is real but unreliable at that reading, and price is sitting on the daily stop, the 200 SMA and the session low all at once. In that kind of chop, panic-selling the low is how you get whipsawed. The plan does the opposite. It cuts half the intended reduction now, sells the rest into a bounce toward 199.68 to 201.41 where the Bollinger midline and 10 EMA stack up, and forces the remainder out on two closes below 178.98. You get de-risked either way without needing to guess the direction.

The neutral will probably say to wait for resolution at 179. I'd point out that resolution costs money. ATR is 15.29, so one day of waiting is a roughly 8.5% swing. I'll also be honest about the numbers. From 179, the move to 199.68 is about plus 11%, and the move to 159.83 is about minus 11%. Technically that's a coin flip. What tilts it is valuation, decelerating fundamentals and the rate backdrop. The plan also arranges to sell the up-leg 11% higher than here.

Someone will say the weekly and monthly SuperTrends are still up, so why sell. Those signals lag. The weekly stop is 159.83. The highest close was 302.52, so waiting for the primary trend to flip means selling about 47% off the top. The stock is already 41% below its peak with the weekly trend "up." That's why I like the trim.

The contrarian case will be that StockTwits looks like capitulation, so buy. The sentiment report itself says nothing shows selling pressure is exhausted, and its confidence is low on a one-day sample. RSI is 40.5, not oversold. The z-score is minus 1.34, so not stretched. TD is at 3 of 9. MFI is 38 against 28 at the September base. OBV has given back about 46% of the September accumulation, while price has given back about 59% of the rally, so the money flow has more to unwind. Real capitulation is people quitting. What I'm reading is people saying "loading up before earnings," selling 180 puts and buying high-beta names on margin. Those are early dip-buyers, and margin buyers are the fuel for the next leg down. Even the September low near 150 was only about 56x run-rate EPS, so valuation never got cheap there either.

On the quality argument, I agree. It's a great business, with no debt, a 7.4 current ratio and 115% year-over-year growth. That's why this is a trim and not an exit. But the stock went from an $87.81 close in March to $302.52 in June, nearly tripling in under three months, and the fortress balance sheet didn't stop a 41% drawdown. Quality is already in the price. It doesn't give you a margin of safety.

On insiders, I'm not leaning hard on it. The cadence looks like plan-based selling, which is a weak signal. Still, it's about $215M since April with zero open-market buys. The CEO sold roughly $2.5M at about $151 on 9/14, right at the base, and about $43M went out in the last four weeks. Nobody with a view that this was cheap acted on it.

Last, two data points to keep honest. The fundamentals report worried that $179 conflicted with insider sales near $213. The technicals resolve that, because the stock really did fall from 220.83 on 10/6 to 179.23. So the 67x multiple is the real, post-drop number. And the 18.5x forward P/E rests on a $9.69 forward EPS that is 3.4x trailing, so I wouldn't anchor on it.

What would change my mind? A reclaim of 201 on rising OBV and volume would. So would a washout near 150 to 160 with MFI under 20, a z-score under minus 2, a TD 9, and a real explanation for the margin drop. Then I'd be a buyer. Until then, being wrong on the trim costs me some upside on a residual half-size position capped at 2% through the December print. Being wrong on holding full size costs me a gap through 159.83. I don't know the caller's book, so all of this is relative to their own position, but I'd take the first risk over the second. Aggressive Analyst: Let me start by handing back the points you two have earned.

The conservative is right that "nearly tripled" was sloppy. 87.81 to 302.52 is 3.4x, and CRDO is still about double its March low, so this drawdown hasn't put a floor under anything. He's also right that my operating income comparison spans two quarters with the fiscal Q4 missing, that the 7% sequential figure is backed out of trailing numbers, and that $43M of insider selling is rounding error against a $2 billion tape. The neutral is right on two bigger things. If the margin drop and slower growth came with the quarter that ended August 1, the 9/2 print already put them in the price, and the stock then rallied 47%. So I can't say fundamentals are what's hitting it today. And my 178.98 two-close trigger is bad. At 0.016 ATR from the close in a 13.5 ADX tape, it fires on noise within days. I'll take the close below 169 and the expiry date for the last tranche.

Now what survives.

Neutral, you say today is a multiple and factor move and treat that as benign. I don't. The 150 to 220 rally was pure multiple expansion: the same $2.68 run-rate EPS went from 56x to 82x. At 179 we're at 67x, the middle of that range, so I'll admit valuation isn't a timing tool. It tells me the reward is thin, not that the direction is down. But if the factor is rates, which is the only evidence we have (hot inflation, the 10-year up, a global bond selloff, 97% no cuts on $8.9M of volume), a factor-driven drop doesn't mean-revert on its own. It reverts when the factor does, and the next inflation prints and the Fed meeting fall inside this window. I agree the hike-count markets are thin and I won't lean on the 86%. The no-cuts leg is deeper, and the headlines back it up independently.

Your 52x on next-twelve-month math is fair, and I accept it's not 67x. But it needs 10% sequential growth to hold for a year, which is about 46% growth against 115% now. Even the bull card decelerates: $530M against $268M a year ago is about 98%, down from 115%. 52x for a decelerating grower with a 3.25 beta isn't a margin of safety. On OBV, I'll call it a wash.

Conservative, your hit on my wording is fair. ATR is variance, not expected cost, so what I'm paying for is the tilt, not the noise. But look at what you're asking for. The gap between "most or all now" and the 60/25/15 structure is about a quarter of standard size. At an 11% move either way, that's under 3% of the position. You want me to sell that quarter at the point where the 200 SMA, the daily stop and the session low all converge, after 17% in three sessions and the heaviest volume since 9/18, to avoid a swing equally likely to go the other way. You said low ADX means whipsaw in both directions. At a fork, going all-in on one branch tells you the least.

I also reject "capped upside, uncapped downside." Resting limits at 189, 195 and 200 mean I'm getting paid for the volatility instead of paying it. By the neutral's math, 195 fills about half the time before the Fed meeting, and that's a sale 9% above where you'd dump it. If it doesn't fill, the expiry gets me out before the meeting, which answers your gauntlet point. On the loss budget, the 2% cap already is one. At that weight, a fall to 159.83 costs about 0.2% of the portfolio, 150 about 0.3%, and even the monthly stop near 110 under 0.8%. The residual isn't where the risk lives. You are right about verification, though. Read the 10-Q on the margin drop and the equity jump, and confirm the earnings and Fed dates, before sizing the residual.

So I'm moving to the neutral's structure. Sell about 60% of the intended reduction, worked through the morning and not at the opening print. Put about 25% on limits near 189, 195 and 200, expiring the day before the Fed meeting. Force the last 15% out on a close below 169 or at that expiry, whichever comes first. Keep the residual at the 2% cap or 0.5x of standard size, whichever is lower, with a weekly close below 159.83 as the backstop. If flat, don't initiate at 179.

You two haven't given me credit for the aggressive part of this. The trim is funding. If the washout comes near 150 to 160, with MFI under 20, z-score under minus 2, a TD 9 and the margin drop explained, I'm a buyer with the capital I freed. I'd be buying on flow and exhaustion, since valuation was never cheap at 150. If CRDO closes back above 201 on rising OBV and volume, I stop selling. Until one of those happens, I'd rather sell volatility than carry it. Aggressive Analyst: Let me say where we've landed, because it's closer than the tone suggests. Two-thirds now versus three-quarters now is a gap of about 8% of a half-size reduction, roughly 0.04 of a standard position. An 11% move on that slice, either way, is about half a percent of a standard position. I won't fight over that, so let me spend the time on the pieces that could actually hurt someone.

Conservative, you're right about the rungs, and I'll stop calling them paid volatility. With no drift, a 195 limit that fills half the time means the unfilled half has to average about 163 for the ladder to break even with selling now. You're also right that 195 is where today's bounce died: open at 190.79, high at 195.18, close 23 cents off the low. But I don't accept that my thesis implies negative drift over the next nine sessions. My case is about thin reward over the next year: a 0.74% free cash flow yield and a bull scenario around 4.4% a year. I've already said valuation isn't a timing tool. The multiple ran from 56x to 82x on the same run-rate earnings and now sits at 67x, the middle of that range. Over this window I treat drift as roughly zero. The rungs are fair bets, and what they buy is time for facts to arrive. We don't know yet whether CRDO fell alone or with ALAB, MRVL, CRWV and NBIS, and nobody has read the 10-Q on the 410bp margin drop. Neutral's peer test is the best idea in this round. If the peers fell in lockstep, it's a factor move and two-thirds stands. If CRDO fell alone on something stock-specific, I go to the conservative's 75%.

On the stop near 169, I'm with the neutral. Today's range was about 16 points, more than a full ATR, and 169 is only two-thirds of an ATR below the close. A normal wick touches it and recovers, so you'd sell the low of a wick. A stop also doesn't protect against an overnight gap, because it fills at the gap open. Your worry was that a close-based rule prints the sale after the gap. Judging the break in the last half hour and selling into that close fixes exactly that, because you're out before the overnight gap, and it doesn't sell every wick. That gets you what you wanted, and it avoids the noise problem you criticized in my 178.98 trigger.

Neutral, you've got my re-entry rule. The 0.281 MFI at the 9/14 base is on the page, so "MFI under 20" would have kept me out of the best entry of the year. I also don't have z-score or TD history at that low, so I can't claim the rest of the checklist would have fired. And you're right that buying at 150 to 160 collides with a weekly-close exit at 159.83. The exit wins. It's a 2% position, and I'd rather re-buy after the stock proves a low than try to catch it. A daily close back above the 169 lower band with OBV ceasing to fall is a fair test, and the 2% cap still binds any buy-back.

Conservative, you're also right that "buying with the freed capital" was loose. If the freed capital goes back into NBIS, CRWV, ALAB or MRVL, we've reset the clock on the same factor. The freed capital is dry powder, and any re-entry belongs inside one factor budget for the whole basket. If the book holds those names, the 0.2% loss math understates the combined hit, as you said.

On rates, I overreached. The deep no-cuts market moved 1.1 points on the week, the 86% hike market trades on tens of thousands of dollars, and a soft CPI print could help this stock as easily as hurt it. I'm not forecasting the next print. The point is that a 3.25 beta name sitting above its risk budget carries both outcomes, and that's an argument about size, not direction. You're also right that the Fed isn't the first event. Confirm TSMC and the hyperscaler dates, which came from the analyst's own calendar knowledge and not the tools, and set the expiry before the first print we wouldn't hold above target through. If that's this week, the plan collapses toward selling now, and I'm fine with that.

Here's where I disagree with the conservative's framing. When expected value is a wash, taking the lower-variance side argues for selling the residual too. What stops us is the residual's job. The company is still doubling revenue, and if the $525M to $535M outlook is next-quarter guidance, that's roughly +98% year over year against $268M. This stock has fallen 41% from its peak and still sits about double its March low. The residual is the part that pays if we're wrong, so I want it at the lower of 2% or 0.5x, rising to the top of that range if the 10-Q and the earnings date check out and shrinking if the margin drop looks structural.

The bold move here isn't dumping everything at the session low on the heaviest volume since 9/18, and it isn't holding full size and hoping. It's cutting hard enough to survive a gap, keeping the part with convexity, and having a rule for when to stop selling. So I land here: sell about two-thirds of the intended reduction tomorrow, worked through the morning and sized after the peer check. Put the last third on rungs at 189 and 195. Sell whatever is left at the earlier of a late-session break below 169 or the confirmed expiry. Take the residual to zero on a weekly close below 159.83. Stop selling on a close above 201 on rising OBV and volume. If someone is flat, don't initiate at 179. It's still a Sell, sized as a trim. Aggressive Analyst: I'll give credit quickly, because most of the last round was right. Conservative, the residual isn't convex. A long stock is linear, and I used the word loosely. What the residual buys is participation in a business that may still be doubling revenue, and the cap makes that affordable. You're also right that the peer test only works if the threshold is written before tomorrow's tape. Neutral, the rungs don't buy facts, tonight's lookups do. Your ordering point on the 10-Q is correct too, since you can't halve the residual first and un-halve it later without buying back.

So I'm taking Neutral's package as the base, with a few edits.

First, the stakes. Conservative, you're right that the cap can bind, so I'll run it at a 10% standard weight. The reduction is 8 points, the disputed one-twelfth is about two-thirds of a point, and an 11% swing on that is roughly 7 basis points. Even at that weight we've spent four rounds on single-digit basis points. I agree the answer is a written default, not more machinery. Sell two-thirds only if the peer test passes cleanly and the calendar is confirmed long. In every other case, including "couldn't check," sell three-quarters.

Second, the line about stopping the selling on a close above 201 is now a dead letter. To close above 201 from here, the stock trades through 189 and 195 first, so both rungs have already filled. That clause belonged to the old plan, where a tranche sat waiting at 199.68. Since we all dropped that rung, I'd take it off the ticket instead of pretending it's a safeguard. The change-my-mind conditions that are still live concern the residual: what the 10-Q says, what the print says, and the weekly close at 159.83.

Third, the 169 rule. I'll take Conservative's midday amendment if it's mechanical, as Neutral asked. If the stock is below 169 at 1 p.m. Eastern and session volume is already above 10 million shares, sell the remainder. Otherwise test at 3:30 and sell into the close if it's below. The 10 million is a judgment call, anchored to today's 12.4 million, the heaviest since 9/18. It's one line, and today's low was ten points above 169, so it won't sell every wick.

Fourth, on the 10-Q, I'd tell the trader what to look for, because that's where the residual's money is. On the 410 basis points of margin, is it mix from a new product ramp that should fade with volume, a price concession to a large customer that would be structural, or a one-time charge? For cash conversion near 70%, are receivables or inventory growing faster than revenue? For the $664M equity jump against $129M of net income and negative $27M of financing flow, was it stock issued for a deal or a stake, an unrealized gain, or a data error, and does it connect to the margin line? A ramp or mix explanation with a clean balance sheet puts the residual at the top of its range. A pricing explanation or an unexplained balance sheet puts it at half that. The call transcript from the early-September print, if that's when it was, may answer most of this in ten minutes.

Fifth, the basket, which I'd put first on the checklist. Retail groups CRDO with NBIS, CRWV, IREN, ALAB and MRVL. If the book holds even two of those, the 0.2 to 0.3 percent loss on a 2% CRDO residual is the small number and the factor exposure is the big one. A cap on one ticker and not the factor is a cap in name only. If CRDO is the only name held, this whole debate is worth a few basis points and the answer is to sell to target. If the caller is flat, don't initiate at 179.

I still disagree with "err toward being done." The plan is done in the way that matters, since most of the reduction executes tomorrow morning on a written rule and the rest has a timer. I'd keep the residual because the cap already says how much unpaid risk we tolerate. Taking it to zero tomorrow would be a verdict on a company whose filing we haven't read. If $530M is the next quarter, revenue is still up roughly 98% against $268M a year ago.

I'll also say plainly that I've retired most of my directional claims. I'm not saying CRDO goes to 160. What's left is a sizing argument: a 0.74% free cash flow yield, a bull case near 4.4% a year, a beta of 3.25, and a position above its budget. That needs no forecast, and it's why I'd rather be early and small than late and large. If TSMC reports this week, as the macro report guessed but didn't confirm, the window is short and the rungs shrink to one, and I'm fine with that. It can move the group either way, so work the first clip after the open settles, not at it.

Here is the ticket I'd hand over. Tonight, do three lookups: peer closes, the 10-Q and transcript, and the dates. Tomorrow, sell two-thirds of the reduction through the morning, or three-quarters if any check fails. Put the rest on rungs at 189 and 195, or only 189 if the window is two sessions, expiring before the first event we wouldn't hold above target through. Whatever is unfilled goes at the 1 p.m. or 3:30 test below 169, or at the expiry. The residual is the lower of 2% or half of standard size, halved if the filing doesn't explain the margin, and zero on a weekly close below 159.83. The cap covers the whole basket, with no add-back before the print. The entry field should read 189 to 195 for the rungs, with a blended expected fill near today's price, and the stop field should read as a weekly-close exit.

It's still a Sell, sized as a trim. Aggressive Analyst: Aggressive Analyst: Most of this is settled, so I'll take the easy points first and then spend my time on the three things that still matter.

Conservative, you're right about the volume clause. Ten million shares by 1 p.m. implies a day bigger than today's 12.4 million, so the rule only fires on a 9/2-type gap, when we'd be selling wherever it lands anyway. Drop it. A plain check below 169 at 1 p.m. and again at 3:30 is enough, and the first clip should be done by noon so the two don't overlap. "Confirmed long" also needed a number, and five sessions is as good as any. Neutral, you're right that this is a three-quarters plan and two-thirds is the upgrade. My defense is that the position is above its budget, so the burden belongs on whatever lets you hold more. And I'll take the ticket correction. A rung is a fair bet, not an edge, so the entry field should read market tomorrow, last quarter at 189, with a blended fill near 179.

Now the three things.

First, the filing. Neutral, your balance sheet reading is the most useful thing said this round, and I'd tighten one piece of it. The $664M equity jump already includes the $129M of earnings, so earnings don't explain about $535M, less stock comp. The cash flows tie out almost exactly: 90 from operations, minus 762 investing, minus 27 financing is minus 699, against a 698 drop in cash. So there's roughly $500M of asset growth with no cash flow behind it and roughly $500M of equity growth that earnings don't explain. That's back-of-envelope, but it supports your inference. Something non-cash, likely stock or warrants issued to a deal partner or a customer, or a revaluation, or a data error.

I draw a different conclusion from it. If it's one story, with a deal or a customer arrangement, purchase accounting in cost of sales and an opex step-up, then part of the margin and opex hit may be a one-time lump and not a trend. That helps the residual. If it's bought revenue, then my "still doubling" line leaned on a number I hadn't earned, and I concede that. Either way, the filing is worth more than any rung on this ticket.

Second, the residual test. I concede I led with the 410 basis points when the sequential opex step was three times the drag, $59M against about $20M. Conservative is also right that EPS grew 97% against 115% for revenue. So make it one test: guided operating margin, compared like for like. The tool tables look GAAP and company guidance is usually non-GAAP. Comparing a non-GAAP guide to the GAAP 64.5% would manufacture a recovery that isn't there, so pull the same-basis Q1 figure from the release. If the guide implies operating margin back at or above last year's 27%, the residual sits at the top of its range. If it's at or below Q1's 25.3%, halve it.

Third, the basket. I agree it's a gate, but it cuts both ways. If the book holds CRDO plus ALAB, MRVL, CRWV or NBIS, the answer isn't automatically to sell more CRDO. From general knowledge, not the tools, CRDO is the one with no debt, a 7.4 current ratio and positive free cash flow in every disclosed quarter, while several of the others are leveraged or capex-funded. If the factor exposure has to come down, cut the weakest balance sheets first and let the CRDO reduction shrink. A basket cap shouldn't be filled by selling the name you'd least want to be forced out of.

A calendar note, unverified. If TSMC reports tomorrow, it prints in Taiwan hours, before the US open. That's the best argument for starting at 10 and not at the bell. One retail post also cites a chip-equipment beat-and-raise on a day CRDO fell 9.6%. If that's true, good demand news didn't hold the stock up, which fits the peer test and tells me the upside on good news is capped by rates right now. Both tails are fine under this plan. If it opens above 189, the rung fills and we sold into strength. If it opens below 169, the clip still goes and the remainder meets the 1 p.m. test. That's what a plan that doesn't need a forecast looks like.

Since we've been accused of nine conditions, here's the short version. Before anything is sold, check the book for basket overlap and read the filing and guide on the same basis. Tomorrow, sell three-quarters of the reduction between ten and noon, or two-thirds only if the peers, the filing and a five-session calendar all pass. Put the last quarter on a rung at 189, with 195 only in the long case. Anything unfilled goes at the 1 p.m. or 3:30 check below 169, or at expiry. The residual is the lower of 2% or half of standard size, halved unless the guided operating margin clears 25.3%, and it goes to zero on a weekly close below 159.83. The cap covers the whole basket, with no add-back before the print.

The bold part of this plan is that it doesn't need to be right about direction. We sell the piece we can't justify at a 0.74% free cash flow yield and a 3.25 beta. We keep the piece that pays if we're wrong, and we hold the freed capital as an option for after the print, inside the same cap. It's still a Sell, sized as a trim.