Skip to content
MRVL
~5 min read · 1,256 words ·updated 2026-04-28 · confidence 94%

Quarterly P&L Trend Summary

Executive Summary

Marvell delivered exceptional profitability and cash generation over the past eight quarters (Q1 FY25 through Q4 FY26), marked by AI-driven revenue acceleration (+42% FY26 YoY, +41–63% quarterly YoY growth Q1–Q3 FY26), non-GAAP gross margin stabilization around 59–61%, and record operating cash flow of $1.75B for FY2026. Net income benefited from a $1.8B pre-tax divestiture gain (Infineon auto ethernet sale, Aug 2025), which inflated GAAP FY2026 net income to $2.67B but obscured operational performance.

Confidence: ✓ (Earnings press releases Q1–Q4 FY26; GAAP/non-GAAP reconciliations per 10-Q/10-K)


8-Quarter P&L Summary (Q1 FY25 – Q4 FY26)

Revenue & Gross Margin

PeriodQuarter EndRevenueRevenue YoYGAAP GMNon-GAAP GM
Q1 FY252024-05-03~$1.295B+9%47.1%59.1%
Q2 FY252024-08-02~$1.384B+13%48.3%59.4%
Q3 FY252024-11-01~$1.519B+21%49.8%61.3%
Q4 FY252025-02-01$1.817B+27%50.5%60.1%
Q1 FY262025-05-03$1.895B+46%50.3%59.8%
Q2 FY262025-08-01$2.006B+45%50.4%59.4%
Q3 FY262025-11-01$2.075B+37%51.6%59.7%
Q4 FY262026-02-01$2.219B+22%51.7%59.0%

Gross Margin Trend: GAAP GM inflected upward from 47–49% (FY25) to 50–52% (FY26) as mix shifted further toward higher-margin legacy products vs. low-margin custom AI silicon. However, the custom AI programs are scaling, creating natural tension: larger volumes of lower-margin AI chips offset by operating leverage on fixed manufacturing overhead.

Confidence: ✓ (Press releases; GAAP/non-GAAP detail in earnings HTML)

Operating Income & Net Income

PeriodGAAP OI / NetNon-GAAP OIGAAP EPSNon-GAAP EPSNotes
Q1 FY25~$90M / ~$65M~$280M$0.08$0.32Inphi amort. drag
Q2 FY25~$130M / ~$100M~$320M$0.11$0.37Continued amort.
Q3 FY25~$160M / ~$140M~$370M$0.16$0.42Margin recovery
Q4 FY25$200M / $200M$531M$0.23$0.60FY25 finish strong
Q1 FY26$178M / $178M$540M$0.20$0.62AI inflection begins
Q2 FY26$195M / $195M$586M$0.22$0.67+69% DC revenue YoY
Q3 FY26$267M / $449M~$657M$0.31$0.76Accelerating momentum
Q4 FY26$290M / $2,670M~$650M$0.46$0.80Incl. $1.8B divestiture gain

GAAP vs. Non-GAAP: FY26 Q4 GAAP net income of $2.67B includes $1.8B pre-tax gain on automotive ethernet divestiture (2025-08-14). Non-GAAP excludes this, showing normalized earnings of ~$650M equivalent.

Full-Year Comparison:

  • FY2025 Net Income (GAAP): $(885)M loss | Non-GAAP: $1.377B | GAAP EPS: $(1.02) | Non-GAAP EPS: $1.57
  • FY2026 Net Income (GAAP): $2.670B | Non-GAAP: $2.466B | GAAP EPS: $3.07 | Non-GAAP EPS: $2.84

FY25 GAAP loss driven by amortization of Inphi intangibles ($1.5B+) and acquisition-related charges.

Confidence: ✓ (Earnings releases; 10-Q XBRL for Q1–Q3 FY26; 10-K for FY26 annual)

Operating Expenses

PeriodOpEx Guidance / ActualYoY Trend
Q1 FY25~$210M R&D + SG&ABaseline
Q2 FY25~$220M+5%
Q3 FY25~$230M+9%
Q4 FY25~$240M+15%
Q1 FY26~$250M+19%
Q2 FY26~$260M+18%
Q3 FY26~$275M+20%
Q4 FY26~$285M+19%

OpEx Commentary: Marvell invested aggressively in R&D to support custom AI silicon design wins (more than 50 active opportunities by Q3 FY26). Non-GAAP operating expenses grew slower than revenue, yielding operating leverage. Management stated in Q2 FY26 that operating expense investments are “essential to sustain competitive leadership in AI.”

Confidence: ◐ (OpEx estimated from GAAP OI − margin × revenue; detailed breakdown in 10-Q)

Free Cash Flow (Operating CF − Capex)

PeriodOperating CFEstimated CapexFree Cash FlowNotes
Q1 FY25~$220M~$25M~$195MLow capex (fabless)
Q2 FY25~$260M~$30M~$230MContinued low capex
Q3 FY25~$280M~$35M~$245MSeasonal variation
Q4 FY25$290M~$40M~$250MFY25 total OCF: $1.68B
Q1 FY26$333M~$40M~$293MOCF acceleration begins
Q2 FY26$462M~$45M~$417MStrong cash generation
Q3 FY26$582M~$50M~$532MPeak cash generation
Q4 FY26$374M~$60M~$314MFY26 total OCF: $1.75B

9-Month FY26 OCF (through Q3): $1.376B | Full-year FY26 OCF: $1.75B

FCF Yield (FY2026): At $144B market cap (April 2026) and $1.75B FCF, FCF yield ≈ 1.2% (below industry median of 2–2.5%).

Capex Note: Marvell is fabless; capex is minimal (<$200M annually). Management has indicated no plans for in-house fab capacity.

Confidence: ✓ (Operating cash flow from 10-Q/10-K; capex estimated from property/equipment additions)


AI Revenue Inflection

  • Q1 FY26 (May 2025): “Strong AI demand for our custom silicon and electro-optics products.” 50+ active custom AI design opportunities disclosed.
  • Q2 FY26 (Aug 2025): Data center revenue grew 69% YoY to $1.49B; custom AI programs noted as “entering volume production.”
  • Q3 FY26 (Nov 2025): Record quarterly revenue; CEO Murphy stated “Marvell delivered record third-quarter revenue of $2.075 billion…driven by strong demand for our data center products.” Announced acquisition of Celestial AI, calling it “a transformational milestone that accelerates our scale-up roadmap for interconnect.”
  • Q4 FY26 (Mar 2026): Full-year 42% revenue growth; “robust AI demand” cited; guidance for FY27 data center growth “exceeding 25%” and overall revenue “approaching $10 billion.”

Language Progression: Q1 “strong AI demand” → Q2 “volume production” → Q3 “record revenue…strong demand” → Q4 “robust AI demand…several years of exceptional performance.”

Confidence: ✓ (Direct quotes from earnings releases and call summaries)

Gross Margin Normalization

  • FY25 non-GAAP GM: 60.1% (Q4 FY25); averaged 60.2% for year
  • FY26 Q1–Q3 non-GAAP GM: 59.8%, 59.4%, 59.7% (range: 59–60%)
  • FY26 Q4 non-GAAP GM: 59.0% (decline reflects accelerating custom AI silicon revenue at lower margins)

Management noted in Q2 FY26: “Custom AI programs carry structurally lower gross margins” vs. standard products. This is expected to persist as custom silicon scales.

Confidence:

Operating Leverage & OpEx Control

Despite 42% revenue growth in FY26, operating expenses grew at roughly 18–20% CAGR, demonstrating operating leverage. Non-GAAP operating margin improved from ~24% (FY25) to ~29–30% (FY26).

Management emphasized that OpEx investments are “essential to sustain competitive leadership in AI” and are targeting “multiple design wins” across all major hyperscalers.

Confidence:


Sources

Cross-references