Skip to content
MRVL
~8 min read · 1,795 words ·updated 2026-04-28 · confidence 39%

Valuation Framework — Marvell Technology (Sensitivity Analysis)

Executive Summary

This framework builds a sensitivity grid showing how Marvell’s valuation changes across different FY28 AI/datacenter revenue scenarios and forward P/E multiples. NO price target is provided; this is purely illustrative to show how changes in growth assumptions or investor sentiment drive valuation.

Current Market Context (April 28, 2026):

  • Market cap: ~$130–140B (range due to intra-month volatility).
  • Shares outstanding: ~874.5M (diluted basis, including SBC and contingent equity from Celestial AI).
  • Implied stock price: ~$150–160 per share.
  • Consensus forward P/E: 35–40× on FY27E EPS of $3.55.

Key Assumptions & Confidence Levels

Revenue Projection Framework

FY27 Revenue: Management guidance $9.93B (+25% YoY). ✓ (Guided)

FY28 Revenue Scenarios:

  • Base case: $17.5B (+76% YoY, or 40% normalized growth from FY27 baseline). ◐

    • Assumes 1.6T optical DSP ramps to 40%+ of revenue; custom silicon doubles; Celestial AI begins $100M+ revenue contribution.
    • Probability: Medium (aligns with management “FY28 double YoY custom silicon” and analyst consensus on optical ramp).

  • Bull case: $19.5B (+96% YoY, or 50% normalized growth). ⚠

    • Assumes hyperscaler capex outperformance; 3.2T DSP sampling begins (early revenue); Celestial AI tracking to $200M+ ARR; Polariton early product revenue.
    • Probability: Low–Medium (requires Celestial AI overperformance and CPO adoption delay; aggressive but possible).

  • Bear case: $15.0B (+51% YoY, or 24% normalized growth). ⚠

    • Assumes hyperscaler capex reprieve; 1.6T DSP ramp disappoints (only 20% of revenue); Celestial AI slips 1–2 quarters; custom silicon growth moderates to 15% YoY.
    • Probability: Medium (capex cycle risk is real; DSP cannibalization by LPO is visible).

Gross Margin Assumptions (FY28)

  • Baseline: 52.0% gross margin (Q1 FY27 guidance: 51.4–52.4%). ✓

    • Optical DSP at 55%+ GM, custom silicon at 40–45% GM, legacy at 45%+.
    • Weighted average at $9.93B FY27 revenue: 51–52% realistic.

  • Bull case: 53.0% GM (Celestial AI mix benefit; 3.2T DSP at 58%+ margin due to modulator integration).

    • Probability: Low–Medium (integration could take 2–3 quarters to optimize).

  • Bear case: 50.0% GM (Celestial AI integration drag; Polariton amortization; DSP mix shift to lower-margin LPO/CPO integration).

    • Probability: Medium (near-term integration risk is real).

Operating Expense Assumptions (FY28)

Current SBC + R&D + SG&A: ~$1.5B annually (~15% of revenue at $10B run rate). ✓

FY28 OpEx estimate:

  • SBC: $200–250M (Celestial AI, Polariton, stock issuance dilution) ⚠
  • R&D: $450–500M (3.2T DSP, CPO/LPO risk mitigation, integration support) ◐
  • SG&A: $250–300M (scaling HQ, integration staff) ◐
  • Total OpEx: ~$900–1050M (~5–6% of $17.5B revenue, or 6–7% at base case).

Assumed operating leverage: OpEx grows at 5–7% YoY (leverage building as revenue scales).

EBITDA Bridge (FY28 Base Case)

Line ItemFY27EFY28E (Base)Comment
Revenue$9.93B$17.5B+76% YoY
Gross profit (52%)$5.16B$9.1B52% GM assumption
OpEx$(1.0B)$(1.0B)Slight leverage; SBC elevated
EBITDA$4.16B$8.1B~46% EBITDA margin at scale
EBITDA margin42%46%Expanding margin
Net income (tax ~13% of EBT)$3.1B$6.0B~34% net margin
Diluted shares885M910MDilution from SBC + acquisitions
EPS (diluted, GAAP)$3.50$6.60Implied earnings growth

Non-GAAP adjustments: Marvell typically excludes SBC, acquisition-related amortization, and one-time charges from non-GAAP EPS. Assume $200–250M annual SBC reduces non-GAAP EPS by ~$0.25–0.30. Non-GAAP EPS estimate: $6.30–6.35 (vs. $6.60 GAAP).

Forward P/E Multiple Assumptions

Marvell’s forward P/E is driven by:

  • Growth rate: 25%+ YoY → 30–40× P/E. 15–20% YoY → 20–25× P/E.
  • AI sentiment: Strong AI enthusiasm (2024–2026) supports 35–45× P/E; if sentiment shifts (CPO ramp, capex reprieve), multiple compresses to 20–30×.
  • Sector comparables: Broadcom trades 25–30× (broader portfolio); Nvidia trades 35–50× (peak sentiment); semiconductor peer average ~22× (ex-AI).

Forward P/E scenarios for FY28:

  • Bull (45%): Sustained AI momentum, Celestial AI overperforming, modulator IP moat expanding. ⚠
  • Base (35%): Moderate growth deceleration, Celestial AI on plan, DSP pressured by CPO but moat intact. ◐
  • Bear (25%): Growth miss, capex reprieve, Celestial AI slip, DSP cannibalization visible. ⚠

Valuation Sensitivity Matrix

Scenario 1: Base Case Valuation Grid

Axis A (Y): FY28 Revenue Scenarios Axis B (X): Forward P/E Multiple (on FY28E EPS)

Assumptions:

  • FY28 Diluted Shares: 910M
  • Tax rate: 13% (non-GAAP adjustments)
  • Gross margin: 52% (base)
  • OpEx: $1.0B
FY28 RevenueFY28 EPS (diluted, GAAP)25× P/E35× P/E (Base)45× P/E (Bull)
$15.0B (Bear)$5.25$131B / $144$184B / $202$237B / $261
$17.5B (Base)$6.60$165B / $181$231B / $254$297B / $326
$19.5B (Bull)$7.40$185B / $203$259B / $285$334B / $367

Implied Per-Share Values (right column of each cell):

  • Bear case (25× multiple): $144–203 per share (downside scenario).
  • Base case (35× multiple): $202–285 per share (most likely range given current consensus).
  • Bull case (45× multiple): $261–367 per share (upside on sustained AI momentum).

Current price context: Stock trading at ~$150–160 implies the market is pricing in the “Bear/Low-End Base” scenario. This suggests:

  • Limited upside if Base case holds (30–50% upside to $200–260). ◐
  • Significant downside (20–30%) if Bear case materializes (capex reprieve, DSP cannibalization). ⚠

Scenario 2: Sensitivity to Gross Margin Variance

Gross margin is critical because optical DSP and custom ASIC have divergent margins. If Celestial AI/Polariton integration is margin-dilutive, EPS miss is ~$0.30–0.50 per $1pt margin decline.

FY28 RevenueGross MarginFY28 EPS35× P/E Value
$17.5B50% (bear)$5.90$207B / $227
$17.5B52% (base)$6.60$231B / $254
$17.5B54% (bull)$7.30$256B / $281

Margin sensitivity: Every 1% of margin change = ~$0.35 EPS swing at $17.5B revenue. Precision on Celestial AI + Polariton margin integration is crucial.

Scenario 3: Multiple Re-Rating Risk (Growth Deceleration)

If FY28 revenue reaches $17.5B but hyperscaler capex disappoints and DSP cannibalization accelerates, the forward multiple may compress from 35× to 25× despite EPS being on plan.

FY28 EPS25× P/E30× P/E35× P/E (Current)
$6.60$165B / $181$198B / $218$231B / $254

Scenario: If market reprices Marvell from 35× to 30× (on deceleration narrative, not earnings miss), valuation drops $33B (14% downside) despite EPS on target. This is a multiple compression risk, not a fundamental risk.

Discount Rate & DCF Sanity Check (Optional)

For reference, a discounted cash flow (DCF) valuation can be cross-checked against the P/E sensitivity matrix:

Simplified assumptions:

  • FY28 net income: $6.0B (GAAP)
  • FCF conversion: 70% of net income (~$4.2B FCF in FY28)
  • Terminal FCF growth: 3–5% (mature semi)
  • WACC: 6–7% (low cost of capital, AI momentum premium)

Implied FCF yield: $4.2B FCF / $231B market cap = 1.8% (very low, implies market is pricing growth, not yield). This is consistent with 35× P/E on a 3%+ net margin business.

DCF sensitivity: If terminal growth is 3% and WACC is 7%, a 40% FCF miss in FY28–FY29 (due to capex cycle) would justify only $150–180B valuation. This aligns with the “Bear case” scenario above.

Dilution & Share Count Risks

Dilution factors:

  1. SBC: ~$200–250M annually (~0.25–0.30 EPS dilution). ✓ (Priced into non-GAAP guidance)
  2. Acquisitions: Celestial AI issued ~24.5M shares (2.8% dilution). Polariton (undisclosed) may issue 5–10M more shares. ⚠
  3. Buyback offset: Marvell authorized $5B buyback (at current $150–160 stock = ~33–38M shares). Only partially executed; net dilution expected 1–2% annually. ◐

Revised share count: 874.5M (current) → 910M (FY28) = 4% net dilution over 2 years. Included in EPS calculations above.

Comparison to Current Market Price

April 28, 2026 market price: ~$150–160 per share. Implied market cap: $131–140B.

Market is implicitly valuing Marvell at:

  • FY28E EPS: ~$5.00–5.50 (assumes GAAP EPS at lower end of bear case).
  • P/E multiple: 28–32× (below consensus 35×, suggesting market is pricing in modest growth miss or multiple compression).
  • Or: Market is pricing in a blend of Base and Bear scenarios, with significant weight on capex cycle risk or DSP cannibalization.

Valuation verdict:

  • At $150–160, stock offers limited upside if Base case holds (20–30% to $200–210). ◐
  • Stock offers downside risk (20–30%) if Bear case materializes. ⚠
  • Upside catalyst: Positive Celestial AI ramp (Q2 FY28), Polariton tape-out success (Q1 FY27), or hyperscaler capex surprise (2026–2027 earnings calls).
  • Downside catalyst: DSP revenue disappointment (Q2 FY27), Celestial AI integration slip (Q1 FY28), or capex reprieve memo (2026–2027 hyperscaler earnings).

Framework Limitations & Caveats

  1. No price target: This is purely illustrative. Valuation is subjective and market-dependent. Earnings, multiples, and sentiment shift quarterly.

  2. SBC dilution: Marvell’s non-GAAP EPS (used in analyst consensus) strips out $200–250M SBC annually. GAAP EPS is ~$0.25–0.30 lower. Framework uses GAAP; compare vs. sell-side non-GAAP guidance with care.

  3. Acquisition integration: Celestial AI ($3.25B) and Polariton (undisclosed) assumptions on timing and margin are speculative. Slips of 1–2 quarters compound EPS drag across FY28–FY29.

  4. Capex cycle risk: Framework assumes capex continues at current levels. A “DeepSeek efficiency” memo or ROI scrutiny could cut capex 15–25%, collapsing 1.6T ramp. No scenario explicitly models 50%+ capex cut (tail risk).

  5. Competitive loss: Framework assumes Marvell retains AWS/Microsoft/Meta. Loss of any single customer (to Broadcom or in-house design) would erase $300M–500M revenue. Not modeled separately.

  6. CPO/LPO cannibalization: Framework assumes gradual DSP mix shift (2028–2030). If CPO adoption accelerates 1–2 years ahead, DSP revenue cliff is steeper, margin compression is larger.

Summary: How to Use This Framework

For bull case: Use $17.5B–$19.5B revenue range and 35–45× P/E multiples. Implies $230B–$330B market cap ($250–360 per share). Catalysts: Celestial AI ramp, Polariton success, hyperscaler capex surprise.

For base case: Use $17.5B revenue and 30–35× P/E multiples. Implies $200B–$230B market cap ($220–250 per share). Assumes Celestial AI on plan, Polariton on track, capex moderate growth.

For bear case: Use $15.0B revenue and 25× P/E multiple. Implies $130B–$165B market cap ($145–180 per share). Assumes capex reprieve, DSP cannibalization, integration missteps.

Decision rule:

  • If risk/reward is 1:1 or better: Bull case dominates; accumulate at current $150–160 prices (1.3–1.7× upside, 1:1 downside).
  • If risk/reward is <1:1: Wait for clarity on Celestial AI ramp (Q2 FY28) or Polariton tape-out (Q4 FY27) before adding exposure.

Cross-references