Summary

  • Marvell targets US$70–90bn of FY31 revenue; its business-level midpoint assumptions add to roughly US$80bn.
  • The bridge is a management forecast, not backlog. Custom contributes about US$30bn at midpoint and carries a 75–85% FY26–FY31 revenue CAGR assumption.
  • FY27’s US$12bn target is the first nearer-term test of whether the growth path is converting into reported sales.

A forecast bridge, not an order book

At the October 6 Investor Day, Marvell raised its FY28 revenue outlook to US$20bn, or about 67% above the US$12bn FY27 estimate. It also put FY31 revenue in a US$70–90bn range. The midpoint is US$80bn, but the company did more than publish a single headline: its slide 40 assigns approximate FY31 midpoints of US$37.5bn to Interconnect, US$10bn to Switching and Storage, US$30bn to Custom, and US$2.5bn to Communications and Other. Those figures add to US$80bn.

That makes the long-range ambition legible as a model. It does not make it contracted. The slide pairs market-growth assumptions with revenue growth rates; it does not list annual customer programs, committed purchase volumes, shipment schedules or backlog that independently supports each line. A useful midpoint sum can explain management’s arithmetic without validating its probability.

Custom carries the steepest climb

The company’s FY31 revenue CAGR assumptions from FY26 are 60–70% for Interconnect, 40–50% for Switching and Storage, and 75–85% for Custom. The approximate US$30bn Custom midpoint would represent 37.5% of an US$80bn total. Marvell separately targets more than US$12bn of Custom revenue in FY29 and about US$30bn by FY31. The presentation points to design-win diversity, but does not publish a customer-by-customer bridge from wins to qualified products, shipments and recognized revenue.

That gap is not evidence the targets will fail. It is the part of the forecast that investors cannot independently reconstruct from the presentation. “Design win” is an earlier commercial milestone than production shipment; qualification, customer deployment, product mix and timing still intervene. The slides do not say how much of the forecast is already under purchase commitment or how the company would revise individual program assumptions.

The business labels also need care. Marvell reports two broad end markets in its filings—Data Center and Communications and Other. Interconnect, Switching and Storage, and Custom are additional business views within Data Center, not separate audited revenue lines in the filing. A reader should not treat the illustrative bridge as a GAAP segment reconciliation.

FY27 offers an observable checkpoint

Marvell reported US$8.195bn of FY26 revenue, including US$6.100bn in Data Center. In FY27’s first half, revenue reached US$5.157bn, of which US$4.004bn was Data Center. August guidance put Q3 revenue at US$3.150bn, plus or minus 5%. Against the US$12bn FY27 estimate shown on Investor Day, H1 plus the Q3 midpoint would leave about US$3.693bn for Q4. That is arithmetic conditional on guidance landing at midpoint—not a company forecast for Q4.

The sequence matters more than a straight-line extrapolation. FY27 at US$12bn implies roughly 46% growth over FY26; FY28 at US$20bn implies another 67%. The next results can show whether the nearer step is arriving, while subsequent quarters reveal whether that pace is sustainable. Marvell’s FY31 midpoint is consistent with roughly 58% annualized growth from FY26, but consistency inside one company model is not external confirmation.

For monitoring, separate reported end-market revenue from addressable-market estimates and design-win commentary. The useful updates are actual Data Center growth, Custom shipment or revenue disclosures, customer qualification and deployment milestones, and revisions to the FY27/FY28 outlook. If the long-range target changes, that would be evidence of a changed forecast—not by itself proof of weak demand or execution failure.

Sources