Marvell Q3 FY26 Earnings Postmortem: Beat, Strong Guide, and a 7–8% Pop vs 12% Implied

MRVLReport Date: 2025-12-02After Market Close

Read original prediction

Results

Model:✔ Correct

Outcome (Actual/Expected)

Beat / Beat

Guidance (Actual/Expected)

Strong / Inline

Predicted Move

up +9.0%

Confidence

60%

Earnings Gap

+7.7%

Session Return

+7.9%

Final crowd results:

No votes recorded

Postmortem

TL;DR

Marvell’s Q3 FY26 print broadly matched the bullish earnings preview:

The company delivered a clean beat on EPS and revenue, with AI data-center strength front and center.
Guidance came in better than “just inline”, with management sounding confident about the next couple of years of AI-driven demand.
At the stock level, MRVL jumped about 7–8% from the pre-earnings close to the next session’s open/close, a solid move but smaller than the ~12% one-day move implied by the front-week straddle.
The concurrent announcement of the Celestial AI acquisition added a new strategic pillar but also injected some deal-risk debate.

Directionally, the call for a beat and an upside reaction was correct. The main miss was on magnitude: realized upside was respectable, but did not fully match the 9% “central case” and came in well below the 12% implied move.

What Actually Happened

Results vs expectations

Q3 FY26 landed clearly in “beat” territory:

Non-GAAP EPS printed above consensus, with operating leverage from the AI data-center ramp.
Revenue came in slightly ahead of Street estimates, driven by data-center and AI custom silicon while other segments remained more mixed.
Year-on-year revenue growth stayed in the mid-30s percent range, consistent with the AI-acceleration narrative rather than a one-off spike.

Relative to the preview, this lined up with the expectation that headline numbers would not be the problem; the Street was more focused on guidance, AI trajectory, and deal optics.

Guidance tone

Guidance was the real swing factor and it came in better than the “inline” stance baked into the preview:

Management pointed to continued strength in AI-related data-center demand over the next couple of years, with commentary that implied a healthy runway rather than near-term exhaustion.
The outlook for the next quarter and beyond was framed with constructive language on AI ramps, while acknowledging lumpiness in legacy or non-AI areas.
The overall message felt closer to “strong” guidance than a cautious “just enough” update.

This was the key area where reality modestly exceeded the preview’s “inline” guidance expectation and helped support the upside gap.

Celestial AI deal

The other major development was the definitive agreement to acquire Celestial AI, a photonics startup focused on scale-up optical interconnects for next-gen AI and cloud data centers.

Key implications:

Strategically, the deal deepens Marvell’s AI/data-center stack, positioning the company as a more integrated provider of high-bandwidth optical links into custom accelerators and XPUs.
Financially, the size and structure of the transaction framed Celestial AI as a meaningful long-term growth driver, with earn-out milestones tied to ambitious revenue targets.
Near-term, investors had to digest a multi-billion-dollar deal on top of an already rich AI narrative, which likely tempered how far the stock could run on the earnings beat alone.

In the preview, the M&A chatter was flagged as a risk: the fear was that expensive or fuzzy deal terms could overshadow a good quarter. In practice, the market’s initial reaction appeared to be cautiously positive—seeing the deal as strategically logical, even if the final verdict will depend on execution.

Price Action vs the Call

Using the canonical amc framework:

Prev close (report day): about $92.89
Next regular-session open: about $100.04
Next regular-session close: about $100.20

That translates to:

Earnings gap return: roughly +7.7% from report-date close to next open.
Earnings session return: roughly +7.9% from report-date close to next close.

The preview’s directional call was:

Direction: up
Expected move magnitude: about +9%
Directional confidence: 0.60

In practice:

Direction was correct: the stock gapped and closed higher in the post-earnings session.
The magnitude came in a bit lighter than the 9% central case and significantly below the ~12% move priced by the front-week straddle.

From an options perspective, that outcome favored:

Traders who sold event vol (iron condors, calendars, diagonals) over those who paid full freight for the front-week straddle.
Directional bulls who used defined-risk call spreads rather than naked calls or lottos.

How the Options & Tape Setup Aged

The preview highlighted a few key features of the tape:

Front-week ATM IV around the mid-100s, with a 12% implied move into the event.
A call-tilted front-week and near-term OI profile, signaling hedged bullishness.
Meaningful downside put demand at strikes just below spot, reflecting respect for another “August-style” disappointment.
Strong call interest and OI further out (e.g., December and mid-’26 strikes), consistent with a structurally bullish AI story.

Postmortem, that read holds up reasonably well:

The directional bias embedded in the chain — modestly bullish but hedged — matched the realized outcome: a solid upside gap but no blow-off move.
The front-loaded IV did prove expensive in hindsight: realized volatility around the event was lower than what the straddle implied, so sellers of pure event vol had the edge.
Longer-dated call interest still looks reasonable given:
- A confirmed AI revenue ramp,
- Strong guidance tone,
- And a strategic optical interconnect acquisition that can leverage hyperscaler demand.

The main takeaway: the options market correctly sniffed out an upside bias, but overpaid for the tail relative to the actual move.

Where the Preview Was Right (and Wrong)

What held up

Beat vs consensus
The preview framed Q3 as a likely fundamental beat on both EPS and revenue, driven by AI data-center strength. That’s exactly what played out.
Directional call (up)
The view that the stock would move higher on a decent quarter and non-disastrous guidance was validated by the 7–8% post-earnings jump.
Characterization of positioning
The chain was described as hedged-bullish — call-heavy, but with substantial downside protection. The post-earnings reaction fits that setup: an upside move big enough to reward bulls, but not so violent that it looked like everyone was offsides.
Framing of risks
The preview emphasized:
- Guidance tone,
- The possibility of another “August-style” sentiment shock,
- And the new Celestial AI M&A overhang.
  In reality, those were exactly the areas investors focused on — they just landed more positively than feared.

Where it missed

Move magnitude
The central case of ~9% upside vs a 12% implied move ended up slightly overstating realized volatility. The stock did rally, but the move was closer to 8% at the open/close rather than 9–12%. This matters for traders sizing straddles, strangles, or wings.
Guidance characterization
Guidance was tagged as “inline expected” going in, with the Street not counting on a heroic raise. Actual guidance and tone came through as stronger than that baseline, which helped reinforce the upside despite already-elevated expectations.
M&A drag vs lift
The preview leaned a bit more toward the risk that Celestial AI deal terms could weigh on the reaction. Early market response suggests investors saw the transaction as more of a strategic accelerator than a reckless overpay, at least for now.

How the Suggested Trade Framework Would Have Fared

The preview sketched out a few broad trade ideas. In light of what actually happened:

1. Defined-risk upside via front-week call spreads

Conceptually: long 12/05 near-ATM call, short higher-strike call (e.g., 93/100 or 95/105).

With a ~7–8% post-earnings pop, bull call spreads anchored near spot would have worked reasonably well, especially if upper strikes were around the high-90s/low-100s.
Because the move undershot the implied 12%, spreads likely captured a good chunk of intrinsic value without being overly exposed to event IV crush.

Verdict: Directionally effective, and more efficient than naked long calls, but still exposed to rich vol pricing.

2. Calendars/diagonals to sell rich event vol and buy time

Structures that sold 12/05 premium and owned later-dated calls fared well in this outcome:

The front-week IV collapsed as expected after earnings, and because the move was significant but not explosive, short front-week options likely decayed nicely.
Longer-dated calls retained decent value thanks to the upward price shift and stronger medium-term AI story.

Verdict: One of the cleaner ways to monetize the “rich event vol / moderate move” mismatch.

3. Downside hedges and put spreads

Protective put spreads around the mid-80s or lower were not needed this time:

They provided insurance that ultimately expired with little or no intrinsic value, consistent with the view that downside risk was real but not realized this quarter.
For long-only holders, the “wasted” hedge premium can still be justified by the size of the implied tail and the August playbook.

Verdict: Costly but not irrational insurance in a name that still has a history of double-digit earnings gaps both ways.

Lessons for Future MRVL- and AI-Style Setups

Strong AI narrative + rich implied move doesn’t guarantee a blowout
Even with a beat, strong guide, and strategically sound M&A, the stock only moved high single digits. Once AI optimism is widely held, there’s a limit to how much incremental good news can move the needle in one shot.
Event vol can stay too high even when direction is right
The chain correctly flagged upside odds, but still overpriced the magnitude. For similar cases — AI leaders with strong positioning and big implied moves — calendars, diagonals, and spreads may be better than naked long gamma.
M&A overhangs are nuanced, not purely bearish
The Celestial AI deal shows that a sizeable acquisition in the AI stack can be taken as validation of a long-term strategy rather than just a near-term EPS drag, especially when structured with performance-based earn-outs.
Guidance “inline vs strong” matters more than small EPS beats
The Street largely assumes Marvell will beat the quarter; the swing factor is how management frames the next 1–2 years of AI demand. Small differences between “inline” and “strong” guidance can be the difference between a flat close and a high-single-digit pop.

Bottom Line

For Q3 FY26, Marvell delivered exactly the kind of AI-driven beat that the preview anticipated, coupled with better-than-feared guidance and a high-profile acquisition that deepens its AI data-center story.

The directional call (beat and up) was validated, and the 7–8% post-earnings jump was a respectable outcome in a name that had a 12% implied move baked into the front-week options.

For traders, the quarter reinforces a familiar pattern in high-expectation AI leaders:

Directional skew can be right,
Event vol can still be too expensive, and
The best risk/reward often sits in defined-risk spreads and vol-selling overlays, not in paying full price for the straddle.

Published: Dec 5, 2025, 8:03 AM

TL;DR — Headline Signal

Marvell reports Q3 FY26 after the close on Tuesday, December 2, 2025. Consensus is looking for a strong AI-driven quarter: EPS up more than 70% year over year and revenue up mid-30s percent, with the call set for the late afternoon.

From a late-afternoon options snapshot (spot at $92.81), the front weekly 12/05 ATM 93 straddle is trading around $11.25, implying roughly a 12% one-day move. Implied vol is massively front-loaded into this earnings week.

Read of the setup:

A fundamental beat on EPS and revenue versus consensus is the base case.
Guidance is likely to land basically in line with what the Street already models, with a constructive tone on AI but still some lumpiness elsewhere and a new M&A overhang.
The stock reaction is expected to be higher, with a base-case move in the high single digits (roughly 9%) versus the ~12% implied move.
The main risk to this view is guidance or deal terms (Celestial AI) landing badly enough to trigger another “August-style” de-rating.

Directionally bullish, but not convinced the stock comfortably outruns the implied move unless guidance is surprisingly clean.

Setup: What the Street Is Braced For

Across Zacks, TipRanks, Yahoo Finance, and other preview pieces, the setup for Q3 FY26 is very consistent:

Non-GAAP EPS: around $0.74–0.75, up roughly 70–75% year over year.
Revenue: around $2.06–2.07B, up about 36–37% year over year.
Company guide coming into the quarter: revenue guidance of $2.06B at the midpoint and EPS guidance of $0.69–0.79, so consensus is sitting comfortably inside the guided range.

This is clearly framed as an AI-driven growth quarter: the Street expects big year-on-year EPS acceleration and mid-30s revenue growth, but not a “blowout” relative to what management itself already guided.

The wrinkle is the recent history:

On August 28, 2025, Marvell reported Q2 FY26 with revenue of about $2.01B and adjusted EPS of $0.67, roughly in line with expectations — yet the stock dropped ~18–19% on a softer-than-hoped data-center forecast and guidance language that spooked AI investors.
Since then, the stock’s 1-year return is slightly negative and year-to-date performance is down around 18%, even after a strong rally off the August lows in the run-up to this print.

So the market is giving Marvell another chance to prove the AI story is intact, but everyone remembers that guidance tone — not the headline numbers — can drive a very large move.

Fundamentals & Filings: Strong Topline, Messy Bottom Line

Recent fundamentals paint Marvell as a classic “AI infrastructure growth with accounting noise” story:

Market cap: about $77B.
Revenue growth TTM: ~37% (TTM) and ~5% year-over-year on the latest annual comp — a clear acceleration phase.
Gross margin TTM: just under 50%, solid for a fabless semi name.
Net margin and ROE: slightly negative on a TTM basis, reflecting write-downs and investment — GAAP remains noisy even as non-GAAP earnings ramp.
Valuation: price-to-sales near 8.9x, P/E TTM not meaningful (negative EPS), so the stock still trades like a high-expectation AI growth name.
Risk profile: beta close to 1.9, making MRVL more of a high-beta AI proxy than a defensive chip name.
52-week range: high around $127.48, low around $47.09. At $92.81, shares are roughly 27% below the 52-week high but almost 2x the 52-week low, consistent with a big drawdown and partial recovery.

On the last call, management highlighted:

Data center as the key growth engine, with strong AI custom silicon ramps offset by some digestion in traditional cloud/networking.
Solid sequential revenue growth and strong demand for AI accelerators, but data-center guidance that came in a bit soft versus exuberant expectations, which is what triggered the August wipeout.

Net-net, the fundamental trend is still strong: AI revenue is climbing rapidly, and overall sales growth is robust. The risk is more about how much of that is already in the price and whether management can avoid another guidance misstep.

Options & Tape Diagnostics

Now the fun part: what the tape is actually saying.

1. Implied move and term structure

From the MRVL chain snapshot:

Spot price: $92.81.
Front weekly expiry: 2025-12-05 (3 DTE).
Nearest-to-the-money strike: 93.
93C mid: $5.55.
93P mid: $5.70.
ATM straddle: $11.25.

That implies a 12.1% move ($11.25 / $92.81) priced into this week’s expiry.

Term structure around the event:

3 DTE (12/05) ATM IV ≈ 165%.
10 DTE (12/12) ATM IV ≈ 104%.
17 DTE (12/19) ATM IV ≈ 87%.
24 DTE (12/26) ATM IV ≈ 78%.
31 DTE (01/02) ATM IV ≈ 69%.

Realized vol context:

30-day realized (HV30) is about 57–58% annualized.
90-day realized (HV90) is about 61% annualized.

So front-week IV is almost 3x realized volatility, and there’s a very steep IV step-down as the chain moves away from earnings. This is classic “all the juice is in the earnings week” term structure, which fits a name with double-digit historical earnings gaps — but it still means the straddle is very expensive.

2. Flow, open interest, and skew via positioning

Chain-level totals from the snapshot:

Total options volume: ~142k contracts.
Total OI: ~1.24M contracts.
Volume put/call: ~0.77 (more call volume than put volume on the day).
OI put/call: ~1.20 (more puts than calls outstanding overall).

Drilling into the key front week (12/05):

Front-week volume:
- Calls: ~37.2k.
- Puts: ~35.3k.
- Volume P/C ≈ 0.95 — very balanced.
Front-week open interest:
- Calls: ~71.5k.
- Puts: ~31.9k.
- OI P/C ≈ 0.45 — call-heavy positioning into the event.

Unusual-activity highlights:

100 calls (12/05): ~5.7k volume vs ~4.3k OI — classic short-dated upside speculation.
86 puts (12/05): ~5.3k volume vs a few hundred OI — chunky downside hedge / protection just below spot.
Additional large flows in 87P, 80P and various 90–101 calls.
A notable longer-dated May 2026 90 put block (~2k contracts) suggests a serious effort to insure the tail further out.

Skew measured purely via IV is fairly symmetric at the strike level (call IV ≈ put IV at a given strike in this dataset), but the distribution of OI tells the real story:

Downside strikes (80–88) are well-owned in puts, especially in front-week and mid-term expiries.
Upside strikes (95–115) are crowded in calls, particularly short-dated calls at 100 and 101.

Put differently: upside participation is popular, but crash risk is very much on the radar. This looks like “hedged bullishness,” not complacent euphoria.

3. How the tape colors the signal

The options structure lines up well with a “beat but still fragile” narrative:

The implied move is big but not outrageous given MRVL’s history of >10% earnings gaps.
Front-loaded IV and heavy call OI say traders are leaning to the upside.
The chunky downside put trades and long-dated protection say larger accounts still don’t fully trust the story and fear another August.

Taken together, the tape supports a directional bullish lean, but with enough hedging that a modest beat plus inline guide might result in a squeeze-then-fade pattern rather than a smooth trend day.

Sentiment: News, Analysts, and Social Buzz

1. News and sell-side tone

Heading into this print, the news flow is busy:

Celestial AI deal chatter: Multiple outlets (The Information, Reuters, Benzinga, Seeking Alpha) report Marvell is in advanced talks to acquire photonics startup Celestial AI in a cash-and-stock deal that could exceed $5B. The technology — light-based interconnects for AI compute/memory — sits squarely in Marvell’s AI/data-center sweet spot. Strategically bullish, yes; but near-term investors are understandably nervous about price tag, dilution, and execution risk.
Q3 preview pieces: Trefis, Yahoo, Zacks, and MarketBeat all frame this as a strong EPS/revenue quarter with consensus around $0.74–0.75 EPS and $2.06B revenue, but they explicitly call out August’s weak data-center forecast and stock crash as the key overhang.
Ratings and technicals: IBD has pushed Marvell’s Composite Rating into the mid-90s, with EPS growth >100% and sales growth near 60%. At the same time, the stock is extended above prior buy points and no longer early-stage cheap.

Taken together, institutional sentiment looks like this:

Medium-term: Bullish on the AI/data-center story.
Near-term: Wary of guidance and M&A terms, and not willing to chase the stock without a clean message.

2. Social sentiment and retail positioning

On X and Reddit, MRVL is visible but not in meme-stock territory:

The Celestial AI rumor is generating noticeable buzz, with bulls hyping the strategic angle (“real AI plumbing, not hype”) and skeptics focusing on “overpaying at the top of the AI cycle.”
There are recurring posts along the lines of “MRVL always dumps on earnings, I’ll buy the dip” — a behavior pattern that primes dip-buyers and can shorten the duration of any selloff if the numbers aren’t catastrophic.

Overall, the social tone is constructively skeptical:

The long-term AI angle is widely respected,
but there is low trust in MRVL’s ability to communicate cleanly around guidance, especially after the August debacle and with a fresh M&A storyline.

That psychology argues in favor of asymmetric upside if management over-delivers on clarity, but it also means disappointment could travel far and fast.

Guidance Scenarios

Given the chain, the news, and the recent history, guidance is the real swing factor.

Base case (central scenario)

Q3 prints modestly above consensus EPS and revenue.
Q4/FY guidance is broadly in line with current Street numbers — perhaps slightly more upbeat AI/data-center commentary but not a sweeping upgrade.
Management acknowledges prior concerns but frames data-center lumpiness as a near-term digestion, not a structural problem.
Celestial AI, if mentioned or announced, is positioned as strategically important but disciplined on valuation and synergies.

In this scenario, the stock gaps up mid- to high-single digits and trades choppy as shorts scramble, call buyers take profits, and hedges blunt the move. That is consistent with the ~9% expected move against a 12% implied move.

Bull case (smaller probability, big payoff)

Strong beat on both lines.
Guidance moves meaningfully above consensus, particularly in AI/data center, with clear visibility into hyperscaler ramps.
Celestial AI is either not yet announced or is announced with investor-friendly terms and a tight, credible synergy narrative.

That is how a 15–20% upside print happens, potentially retracing a good chunk of the August damage in one go.

Bear case (fat left tail)

Numbers come in only in line or slightly below.
Guidance remains cautious on data-center growth or flags further digestion.
Celestial AI is framed (or perceived) as expensive and dilutive with long-dated payback.

In that world, MRVL can absolutely revisit its August playbook, with a 15%+ drawdown and a possible break of the recent uptrend. The presence of large downside puts and longer-dated 2026 hedges shows this scenario is front-of-mind for many players.

Taken together, this justifies setting guidance expectations at “inline”: the market is looking for reassurance and no new disasters, not a moonshot guide.

Trade Framework (Non-Advice)

These are framing examples for earnings-focused traders, not recommendations.

1. Defined-risk upside vs. rich implied move

For traders who are bullish on a beat and constructive guidance but wary of 165% front-week IV:

One approach is to sell some of that front-week IV and buy time via December monthly call spreads:
- For example, buying Dec 19 calls just above spot (e.g., 95) and selling higher strikes (e.g., 105 or 110) to cheapen the structure.
This keeps exposure to a post-earnings drift and reduces the impact of IV crush relative to straight 12/05 call lottos.

It lines up with the existing call interest around 95–105, but in a more risk-defined, premium-conscious way.

2. Calendar/diagonal structures around the implied move

Given how hard the vol curve front-loads earnings:

A calendar or diagonal spread around 93–95 (long a slightly later expiry, short 12/05) tries to exploit:
- Front-week implied move (~12%) being a bit rich versus a base-case realized move closer to high single digits.
Risk comes if the stock explodes 15–20% in either direction, so sizing and strike selection matter.

This framework is aligned with the tape: very high front-week IV, still-elevated but lower vol further out.

3. Hedge overlays for existing longs

For holders who like the long-term AI story but don’t want another August, the chain practically begs for downside definition:

Protective put spreads centered around 80–88 in the near-term expiries align with where a lot of institutional hedging already sits.
Alternatively, a collar (selling some upside calls near 100–105 to fund protective puts) can make sense for those content to cap upside in exchange for protection through earnings.

The overarching idea is to respect that this is a high-beta AI name with a history of double-digit earnings gaps, not a sleepy dividend stock.

Risks & What Would Change the View

Ways this call could be wrong:

Guidance / data-center outlook: If management again guides data-center growth below what the market quietly hopes for, or leans too hard on “lumpiness,” the downside could exceed the implied move and invalidate the bullish stance.
Celestial AI deal optics: A perceived overpay or fuzzy integration story could overshadow a solid quarter, especially if it looks like MRVL is “chasing AI headlines” at the wrong time in the cycle.
Macro or AI-capex narrative shift: Any hint from hyperscalers of slower AI capex growth or design wins trending elsewhere (e.g., to rivals) would hit the long-term thesis, not just this quarter’s numbers.
Positioning more one-sided than it looks: If dealers are net short gamma into the event and the crowd is much more call-heavy than raw volumes imply, a modest disappointment might trigger a sharp, reflexive downdraft beyond what current hedging suggests.

If a soft guide, rough M&A terms, or clear signals that AI capex is structurally shifting away from Marvell emerge, this quickly shifts from “beat / up” toward “inline or miss with downside skew”, and hedged or outright bearish structures become more appealing.

For now, with a 12% move implied, hedged-bullish positioning, and fundamentals that still look strong beneath the GAAP noise, the base case remains a beat, an “okay” guide, and an upside reaction that probably falls a bit short of what the straddle is pricing in.