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

Two of three resolution quarters are now effectively known: Q4 FY2026 DC was $1.5B (miss), Q1 FY2027 is guided around $1.65B (very likely miss). The entire question now hinges on Q2 FY2027. To resolve YES, Q2 DC revenue must exceed $2.0B — a $350M sequential increase from ~$1.65B, or approximately 21% sequential growth. This is an extraordinarily large single-quarter jump. However, the new FY2027 DC growth guidance of 40% YoY (raised from >25%) implies an average quarterly DC revenue of ~$2.08B, and the exit rate >$3B by Q4 FY2027 requires steep sequential ramps through the year. The implied trajectory is roughly $1.5B -> $1.65B -> $1.95B -> $2.35B -> $3.0B+. Under this trajectory, Q2 would be ~$1.95B — tantalizingly close to but below $2B. Custom silicon programs ramping in H2 (3rd and 4th programs entering volume) could pull some revenue forward, but the management commentary emphasizes second-half strength. Interconnect growth raised to >50% YoY and switch revenue to >$600M both support faster DC growth, but the sequential math is demanding. I estimate ~22% probability — the path exists but requires Q2 to be the quarter where the custom silicon step-function arrives rather than Q3/Q4.

The new data has fundamentally reshaped this question from 'will DC reach $2B in one of three quarters' to 'will DC reach $2B specifically in Q2 FY2027.' This is a much harder bar. The Q4 actual of $1.5B was actually slightly below Q3's $1.52B — the first sequential decline in DC revenue during the AI era. While management attributes this to seasonality and guides Q1 to ~$1.65B (a $150M sequential rebound), the fact remains that DC took a step back before stepping forward. The 40% FY2027 DC growth guidance is extremely bullish, but management explicitly describes an exit rate trajectory pointing to Q4 FY2027 as the big quarter. The words 'exit rate' and 'stronger second half' dominate the forward-looking commentary. If we model a linear ramp from $1.65B (Q1) to $3.0B+ (Q4), we get Q2 at approximately $2.1B — which would clear $2B. But ramps are rarely linear; they're typically S-curved with custom silicon, meaning Q2 could be more like $1.85-$1.95B with the steep ramp hitting in Q3-Q4. The switch revenue raise to >$600M and interconnect >50% growth are positive but likely have the same H2 weighting. I put this at 18%.

I want to steelman the bull case more than the other Opus runs. The 40% FY2027 DC growth guidance implies $8.3B in annual DC revenue. If Q1 is $1.65B and Q4 is $3.0B+, the remaining $3.65B must come from Q2+Q3. A roughly even split gives Q2 at ~$1.85B and Q3 at ~$1.8B — wait, that doesn't work. Let me recalculate: FY2027 total DC = $8.3B. Q1 = $1.65B. Q4 = $3.0B. Remaining for Q2+Q3 = $3.65B. If growth accelerates each quarter: Q2 = $1.75B, Q3 = $1.9B — no, that only sums to $3.65B if Q2 = $1.75B and Q3 = $1.9B. That puts Q2 well below $2B. But what if the custom silicon programs ramp faster? The 3rd custom program entering volume in H1 FY2027 could create a step function. Also, Marvell has beaten guidance every quarter — if Q1 comes in at $1.75B instead of $1.65B, and the acceleration continues, Q2 could reach $2B. The interconnect segment alone at >50% growth contributes ~$900M to Q2 if evenly distributed. There's also the possibility that management is sandbagging to maintain beat-and-raise — their actual trajectory could be steeper than guided. I put this at 25%, reflecting the plausible bull case where custom silicon conversion is earlier than H2-weighted guidance implies.

The question has narrowed to a single quarter: Q2 FY2027 (reporting ~August 2026). With Q4 FY2026 at $1.5B and Q1 FY2027 guided to ~$1.65B, DC revenue needs to jump from approximately $1.65B to above $2.0B in one quarter — a 21% sequential increase. For context, the best sequential DC growth in the recent history was the Q1 FY2026 jump from $1.37B to $1.44B (+5.1%). Even the guided Q1 FY2027 rebound of ~$150M (+9.1% sequential) would be the strongest in recent quarters. A $350M jump would be unprecedented. However, several factors make this less impossible than the historical sequential pattern suggests: (1) FY2027 DC growth of 40% YoY is a massive raise from >25%, (2) interconnect growth raised to >50% means this ~50% DC sub-segment alone could add $150M+ sequentially, (3) custom silicon programs transitioning to volume production create lumpy step-functions rather than smooth ramps. The question is whether the step-function lands in Q2 or Q3. Management's emphasis on exit rates and second-half strength suggests Q3-Q4, but custom program timing is inherently uncertain. I estimate 20% — the step-function path exists but is more likely to materialize one quarter later.

I'm taking a more conservative stance. Let me work through the math carefully. Q4 FY2026 DC = $1.5B. Q1 FY2027 DC estimate = $1.65B. For the 40% FY2027 DC growth: FY2026 DC was approximately $5.96B ($1.44B + $1.49B + $1.52B + $1.5B = $5.95B). So FY2027 DC = $5.95B * 1.40 = $8.33B. Q1 = $1.65B. If Q4 exit rate is >$3B, say $3.1B, and Q3 is somewhere between: Q2 + Q3 = $8.33B - $1.65B - $3.1B = $3.58B. If Q3 = $2.2B (a reasonable ramp point), then Q2 = $1.38B — which can't be right because Q1 is already $1.65B. So let me try: Q3 = $2.5B, Q2 = $1.08B — that's worse. The math requires a more gradual ramp. Let me try: Q1 = $1.65B, Q2 = $1.9B, Q3 = $2.3B, Q4 = $2.48B — sum = $8.33B but Q4 is below $3B exit rate. Try: Q1 = $1.65B, Q2 = $1.85B, Q3 = $2.35B, Q4 = $2.48B — sum = $8.33B, still Q4 below $3B. The exit rate >$3B constrains the math to require very steep H2 growth: Q1 = $1.65B, Q2 = $1.78B, Q3 = $2.4B, Q4 = $3.0B, sum = $8.83B (above 40% target). Or Q1 = $1.65B, Q2 = $1.78B, Q3 = $2.2B, Q4 = $3.0B, sum = $8.63B (45% growth — plausible given beat tendency). Under these scenarios, Q2 is around $1.78-$1.85B — meaningfully below $2B. I put this at 15%.

There's genuine uncertainty in this estimate because the quarterly distribution of DC revenue within FY2027 is unknowable with precision. The 40% annual growth guidance and >$3B exit rate establish boundary conditions but leave wide ranges for Q2. I model three scenarios: (1) H2-loaded ramp (most consistent with management language): Q2 = $1.78B — NO. Probability-weighted at 50%. (2) Gradual acceleration: Q2 = $1.95B — NO but close. Weighted at 30%. (3) Early custom silicon inflection: Q2 = $2.1B — YES. Weighted at 20%. Expected value: 0.50 * 0 + 0.30 * 0 + 0.20 * 1 = 20%. But I should also account for a tail scenario where the ramp is even more back-loaded than scenario 1 (5% weight) and an optimistic scenario where beats push Q2 to $2.2B+ (5% weight). Adjusted: ~23%. The LOW confidence reflects that the quarterly distribution within an annual guidance is inherently noisy — custom silicon program timing is the key swing variable and it's genuinely unpredictable at this resolution.

Two quarters failed: Q4 FY2026 ($1.5B) and Q1 FY2027 (~$1.65B). Q2 FY2027 must exceed $2B — needing $350M+ sequential growth from ~$1.65B. The 40% FY2027 DC growth and >$3B exit rate support a steep ramp, but management's language points to H2 (Q3-Q4) as the inflection. Implied Q2 is $1.8-$1.9B under most ramp models. Beat history could add $50-100M upside but not enough to bridge the full gap. Probability: 17%.

I want to push back on the consensus forming around 15-22%. The 40% FY2027 DC growth guidance was a MASSIVE raise — from >25% to 40%. Management doesn't raise guidance by 15 percentage points without strong conviction. If Q1 comes in at $1.75B (beating the ~$1.65B estimate), and Q2 growth accelerates further, $2B is reachable. The interconnect segment at >50% growth is roughly $800M/quarter growing to $1.2B+ — that's $400M of incremental revenue across the year from one sub-segment alone. Switch revenue at >$600M (from >$300M FY2026) adds another $75M/quarter. Custom silicon doubling adds further. The pieces exist for Q2 to hit $2B if the ramp is even slightly less H2-loaded than the base case. Management has beaten every quarter — their 40% guide could easily mean 45%+ actual. I go higher at 28%.

Balancing the bull and bear cases. The strongest bear argument: management explicitly emphasizes exit rate and second-half strength, suggesting they themselves expect Q2 to be below $2B. If $2B were expected in Q2, they would likely signal it — it's a major milestone. The strongest bull argument: 40% growth with >$3B exit rate means enormous sequential jumps are coming, and custom silicon timing is inherently lumpy enough to pull forward. The question is essentially whether the custom silicon step-function arrives in Q2 (August 2026) or Q3 (November 2026). Given management's emphasis on H2 and the program transition timing, Q3 is more likely. But there's meaningful probability of Q2 given program timing uncertainty. I settle at 19%.