Marvell’s Hyperscaler Deal Lands In A Fiscal Year Management Will Not Size Yet

The programs the deal covers in the near term were already inside the targets management had published, and it is connectivity rather than custom silicon that management points to for the raise that did arrive.

Marvell Technology (MRVL) reported a record quarter, raised its revenue outlook for fiscal 2027 (by about $500 million) and fiscal 2028, and lost 10.3% of its value in the first session after the report. Coverage of the session named two drivers: softer-than-hoped fiscal 2028 guidance despite the raise, and the timing of an expanded commercial agreement and warrant with a key hyperscaler. The second turns on when that money arrives rather than whether it does.

The Warrant Is Already Inside The Targets Through Fiscal 2028

Revenue of $2.74 billion beat the consensus estimate of $2.71 billion, and adjusted earnings of $0.94 a share cleared consensus too. AVGO, NVDA and AMD fell between 0.7% and 4.6% over the same span, so most of the selling was Marvell's own. The agreement covers inference accelerators, storage controllers and network interface controllers attached to the TPU ecosystem, and an analyst on the call sized it at roughly $120 billion of cumulative revenue over six years if every milestone is hit, about $18.5 billion a year annualized, from that one customer's custom-attach programs alone. The guidance does not carry that pace: the same analyst put custom at around $5 billion to $6 billion in calendar 2027. Management's answer: the programs the agreement covers through fiscal 2028 are already inside the custom revenue target the company had published earlier, and the meaningful contribution starts in fiscal 2029.

Connectivity, More Than Custom, Paid For The Fiscal 2028 Raise

Fiscal 2028 revenue is now guided to roughly $18 billion, up $1.5 billion from the $16.5 billion outlook given one quarter earlier. Custom silicon is part of that, but management calls connectivity the bigger bucket and probably the largest driver, pointing to scale-up optics and switching rather than at the programs the warrant covers. Both sit inside one end market: data center revenue made up 79% of the total, and the company's own risk disclosures flag a rising dependence on a few customers. Concentration of that kind is a different proposition from the Trefis High Quality Portfolio, which does not depend on the handful of largest technology names to produce its returns.

The Custom Ramp Reaches Gross Margin First

Non-GAAP gross margin was 58.9% in fiscal Q2 2027, and the fiscal Q3 2027 guidance midpoint is 58.0%, a 90 basis point step down that management attributes to the forecast acceleration of its custom business. The CFO's preliminary view is that fiscal 2028 gross margin sits in the same band as the back half of fiscal 2027. The lower-margin part arrives first, though management guides operating margin into its 38% to 40% target range exiting fiscal 2027.

October Is Where Fiscal 2029 Gets Its Number

Management's fiscal 2029 custom target of more than $10 billion is not a new number: management said it had flagged the figure as far back as its last call and as early as June 2025, after having put custom revenue at $8 billion to $10 billion in April 2024. Management says the warrant biases that target higher but has declined to size the increase before the October 6 investor day. Until then, the case for owning Marvell rests on connectivity execution that is already inside the guide, not on a warrant increment nobody has put a number on. Options currently price Marvell's implied volatility near 60%, in the 56th percentile of its trailing one-year range, a middling reading for how much movement the market is pricing over a year.

A Bet On One Build Cycle Is Still One Bet

Marvell's story from here turns on how fast a handful of hyperscaler programs move from design to volume, a schedule set by its customers. Investors who would rather hold a rules-based system than a single build cycle can start with the Trefis High Quality Portfolio. That portfolio has a track record of outpacing the three major indices - the S&P 500, S&P Mid-cap, and Russell 2000.