MRVL Earnings Preview: 63% Margin Compression Risk

The Core Question

After a 35% correction from $127 to ~$80, the market has repriced Marvell from STRETCHED to DEMANDING expectations. The ensemble assigns 63% probability that gross margins break below 59% — the clearest bearish signal in the set. Tomorrow's Q4 results will show whether custom silicon mix shift is compressing margins faster than revenue can grow.

For the full four-lens analysis covering moat durability, revenue conditionality, and the Celestial AI acquisition, read the deep dive here.

Ensemble Forecast

Key Market

Gross Margin Below 59%

Probability

63%

Model Agreement

0.94

Strong consensus on margin pressure

Our nine-model prediction ensemble assigns 63% probability that Marvell's non-GAAP gross margin will break below 59% in a quarter through Q2 FY2027. Margins have already compressed from 60.1% to 59.7% as custom silicon — the fastest-growing but lowest-margin segment — accelerates its revenue share. Despite only 9% probability of a Q4 revenue miss, the ensemble identifies margin mix as the most likely source of a negative earnings surprise. See all eight active markets on the MRVL forecasting page.

Earnings Scorecard — March 5

Revenue

Bull: >$2.2BBear: <$2.09B

Marvell has beaten guidance every quarter during the AI ramp. Consensus expects ~$2.2B (+21% YoY). A miss below $2.09B (9% probability) would be the first material guidance failure, challenging the FY2027 $10B target.

Non-GAAP Gross Margin

Bull: >60%Bear: <59%

The ensemble's top bearish signal. Margins have compressed from 60.1% to 59.7% as custom silicon scales. Below 59% would be the first concrete evidence of structural mix-driven compression, validating concerns that FY2028 margin targets are unachievable.

Data Center Revenue

Bull: >$1.8BBear: <$1.5B

At 73% revenue concentration, data center is the entire thesis. DC revenue growth has decelerated from 78% to 38% YoY. Only 28% probability of exceeding $2B quarterly by Q2 FY2027 — the growth trajectory may converge toward ~21% CAGR rather than management's guided ~30%.

Custom Silicon Pipeline

Bull: 3rd programBear: Still only 2

Only 2 of 18+ custom XPU programs are in production today. The ensemble assigns 58% probability to a 3rd reaching volume by Q2 FY2027. Any announcement of a 3rd production program would narrow the DIVERGING narrative-reality gap. The market prices 3-5+ by FY2028.

FY2027 Revenue Outlook

Bull: $10B+ trajectoryBear: Below consensus

Management's $10B FY2027 target implies ~30% CAGR. The ensemble suggests the normalized trajectory may be closer to 21%. Q1 FY2027 guidance will reveal whether the gap between narrative and reality is narrowing or widening.

Current Assessment

Marvell is classified as “Price at Value” at medium confidence. At ~$80, after a 35% correction, the price appears to approximately reflect the probability-weighted center of outcomes — near-term execution is strong (9% miss probability), but margin compression (63%), uncertain DC revenue trajectory (28% probability of $2B), and the Celestial AI coin flip (47%) create material medium-term uncertainty. Read the full thesis assessment. We will update within 24 hours of the March 5 earnings release.

Full four-lens analysis covering moat durability, revenue conditionality, the Celestial AI acquisition, and all eight active prediction markets

MRVL Full Analysis MRVL Forecasts