Will Marvell's quarterly Data Center revenue exceed $2B by Q2 FY2027?

The math is demanding: reaching $2B from $1.52B (Q3 FY26) requires +$480M over 3 quarters, averaging ~$160M sequential adds. The observed trend is sharply declining: +$70M, +$50M, +$30M. Even if Q4 FY26 comes in around $1.61B (implied by total company guide), that still leaves ~$390M needed over Q1-Q2 FY2027. At the Myth Meter's base-normalized ~21% CAGR, DC revenue would reach approximately $1.84B by Q2 FY2027 — well short of $2B. The critical swing is custom silicon program timing: if a 3rd program enters volume production, that could provide a step-function $100-150M add. But program timing is lumpy and uncertain. Management's >25% FY2027 DC growth guide gets to roughly $1.9B quarterly run rate by year-end FY2027, still suggesting $2B is more of an H2 FY2027 event than Q2 FY2027. The CFO stock purchase is directionally positive but doesn't change the arithmetic.

I weight two competing dynamics. Against: sequential additions have declined from $70M to $30M, and the Myth Meter's 21% base-normalized CAGR would put DC at ~$1.84B by Q2 FY2027. For: management has beaten guidance every quarter during the AI ramp (per Gravy Gauge), switch revenue is guided to >$500M in FY2027 (up from >$300M), and custom silicon targeting a doubling by FY2028 implies meaningful ramp in FY2027. The question asks for any of three quarters to exceed $2B — the third quarter (Q2 FY2027) provides the most runway. If custom silicon ramps on schedule and switch revenue accelerates as guided, Q2 FY2027 could approach $1.9-2.0B. But 'approaching $2B' is not 'exceeding $2B.' The bar is crossing $2.0B, which requires the optimistic end of the guided trajectory. I estimate ~35% probability, acknowledging the declining sequential trend is the dominant signal but leaving room for the custom pipeline step-function.

Let me construct the arithmetic pathway. Q4 FY2026: total company guided $2.2B, DC at ~73% = ~$1.61B. For FY2027, management guides DC >25% growth. If Q3 FY26 DC is $1.52B annualized at ~$6.08B, then >25% growth implies >$7.6B in FY2027, or >$1.9B/quarter average. That makes $2B possible but not certain by Q2 FY2027 — it's more consistent with reaching $2B in the back half. The critical unresolved debate matters enormously: if the Moat Mapper's 'normal maturation' view is correct, then the guided >25% is achievable and $2B is borderline by Q2 FY2027. If the Myth Meter's 'narrative disconnect' view is correct, actual growth runs closer to 21%, and $2B is pushed to H2 FY2027 at earliest. The declining sequential adds ($70M -> $50M -> $30M) are the strongest empirical signal, and they favor the Myth Meter's view. I lean toward 32% — recognizing the possibility but respecting the declining trend.

The numbers don't lie: sequential DC revenue additions went $70M, $50M, $30M. That's a clear decelerating trend. Reaching $2B from $1.52B requires an average of $160M/quarter in sequential adds over 3 quarters — more than 5x the most recent quarter's $30M add. Even the most optimistic Q4 FY2026 estimate of ~$1.61B still leaves a $390M gap over 2 quarters. Management's >25% FY2027 guide is an annual figure that gets to ~$1.9B average quarterly, not $2B+. Custom silicon ramp could provide a step-function but timing is inherently uncertain and lumpy. The Myth Meter's base-normalized CAGR of 21% is the more grounded estimate. At 21% growth applied to Q3's $1.52B, Q2 FY2027 DC revenue would be roughly $1.76B — nowhere close. I give this 25%.

The question is fundamentally about whether custom silicon and switch revenue can inflect fast enough to overcome the decelerating electro-optics trend. Electro-optics is ~50% of DC revenue and growing more slowly; custom XPUs at ~25% are the growth engine. Switch revenue going from >$300M to >$500M in FY2027 adds ~$50M/quarter if evenly distributed. Custom silicon doubling by FY2028 implies meaningful but back-loaded ramp. The question gives 3 quarters, with Q2 FY2027 being the latest possible resolution date. Even with optimistic sub-segment growth — custom silicon +50% ($187M -> $280M/quarter) and switches adding $50M/quarter — you get to roughly $1.85B by Q2 FY2027. Getting to $2B requires everything firing simultaneously plus electro-optics re-accelerating. Possible but improbable within the timeframe. 30%.

I'm giving slightly more weight to the management beat pattern documented by the Gravy Gauge — Marvell has beaten guidance every quarter during the AI ramp. If Q4 FY2026 beats by even 3-5%, DC could come in at $1.65-1.70B rather than $1.61B. That narrows the gap meaningfully. And Q4 FY2026 earnings (expected March 2026) may come with an aggressive Q1 FY2027 guide if the custom programs are ramping. Still, even at $1.70B in Q4, reaching $2B by Q2 FY2027 requires +$300M over 2 quarters ($150M/quarter sequential). That's 3x the recent trend. The CFO purchase at $78 is a mild positive. I settle at 33% — beat pattern provides some upside but the magnitude of acceleration required is substantial.

Sequential DC adds declining: $70M -> $50M -> $30M. Need ~$160M/quarter average to reach $2B in 3 quarters. The gap is enormous. Even management's >25% FY2027 guide doesn't get to $2B by Q2 FY2027. The Myth Meter's 21% normalized CAGR makes this unlikely within the timeframe. Custom silicon ramp is the wildcard but timing is uncertain.

Simple math: $1.52B to $2B is +32% growth needed over 3 quarters. Recent YoY DC growth decelerated from +78% to +38%. At 38% YoY from Q3 FY26's $1.52B, you'd get to $2.10B — but only 4 quarters later (Q3 FY27), which is after the resolution window. By Q2 FY2027 at 38% YoY pace, DC would be approximately $1.90-1.95B. Below $2B. At the 21% normalized CAGR it's even further away.

The three-quarter window (Q4 FY26, Q1 FY27, Q2 FY27) provides some runway but the trend is clearly against. Q4 FY2026 likely around $1.61B based on guidance. Switch revenue ramp ($300M to $500M) and custom silicon inflection could accelerate things, but $2B by Q2 FY2027 requires near-perfect execution plus favorable timing on lumpy programs. More likely a H2 FY2027 event.