Marvell (MRVL) Falls 6% Two Days After the Google Warrant: The Vesting Schedule Explains the Round Trip
Marvell traded down roughly 6% Friday, two sessions after the Google disclosure closed it up 9.85% on Wednesday and another 1.6% Thursday. Nothing negative arrived in between. The 8-K is the same 8-K, the sell-side response has been uniformly constructive, and Reuters reported analysts reading the agreement as capable of adding more than 60% to the longer-term revenue outlook. What changed is who owns the shares.
The mechanics of the announcement explain most of the round trip. Marvell issued Google a warrant on 18 August covering 58,970,907 shares at an exercise price of $206.58, about 7% of shares outstanding and roughly $12.2 billion at the strike. Only about 1.36 million of those shares vest on a clock, in equal quarterly installments through the first year. The remaining 57.6 million vest as Google spends: 240 tranches of roughly 240,042 shares each, one for every $500 million of custom-products revenue Marvell recognizes from Google and its affiliates, measured from 1 August 2026 through 29 January 2033. Two hundred forty times $500 million is $120 billion. That is where the headline number comes from, and it is a ceiling on vesting rather than a revenue forecast.
The distinction is the entire trade. Marvell’s record fiscal Q1 revenue was $2.4 billion counting every customer it has, and consensus for the quarter reporting 27 August sits near $2.71 billion. A fully vested warrant implies something close to $18 to $19 billion a year from one account, sustained for six and a half years, against a company currently annualising under $11 billion in total. The market spent Wednesday pricing the ceiling and has spent the two sessions since discovering that the ceiling is not the guide. That is not a fundamental deterioration. It is the ordinary process by which a structured incentive gets re-read as a structured incentive.
The exercise price deserves more attention than it has received. At $206.58 the warrant was struck comfortably below the roughly $234 the stock fetched on Wednesday morning, and further below where it traded Thursday. Google’s option was in the money at issue. The economic transfer is therefore not a contingent future event dependent on appreciation; it is live now, scaling with purchases. Marvell has moved a customer-acquisition cost off the income statement and onto the capitalisation table, where it does not touch gross margin and does not appear in the guide. It also means the dilution schedule and the revenue schedule are the same schedule. Every $500 million of Google revenue that validates the thesis simultaneously issues 240,042 shares against it. Investors who bought Wednesday’s pop bought the numerator; the denominator arrived with it.
Positioning did the rest. The stock carries a trailing multiple in the mid-70s and a street average target near $273 against a spread that runs from the low $200s to the high $300s, which is a disagreement about the terminal number rather than the quarter. A name priced that way converts good news into an exit rather than an entry, because the marginal holder is short-term and the marginal buyer needs a reason that has not already been announced.
The rate backdrop compounds it, though the sequencing matters and has been muddled in most of the coverage. Marvell fell 8.3% on Tuesday 18 August as the 30-year Treasury yield hit its highest level since 2007 and the semiconductor complex sold off alongside Micron and the rest of the group. That move preceded the Google disclosure rather than accompanying it. It is context for why the stock was jumpy going in, not a second cause of Friday’s decline. What it does establish is duration sensitivity: a valuation built almost entirely on data-centre revenue arriving between 2028 and 2033 discounts at the long end, and the long end has been moving against it.
On competitive position, the more durable read is that Broadcom fell more than 5% on the same news for a reason. Google diversifying its custom-silicon supply is a structural loss for the incumbent and a structural gain for Marvell, but the agreement’s scope is the tell: AI inference accelerators, storage controllers, network interface controllers, memory interface controllers and near-memory compute. Most of that list is attach rather than compute. The SerDes and optical DSP franchise survives a hyperscaler deciding to design its own accelerator; the accelerator socket does not. Marvell has been paid in warrants for a relationship whose most defensible revenue is the part nobody writes headlines about.
Fiscal Q2 lands 27 August. The print itself is unlikely to move the model, since guidance for FY27 and FY28 was pre-spent in May. The disclosure worth watching is whether management breaks out custom-products revenue attributable to Google as its own series. It now has a contractual reason to track it to the half-billion, and once that number is public every quarter, the $120 billion stops being a headline and starts being a run rate that either compounds toward it or does not.