Oppenheimer raises Marvell stock price target on AI growth outlook

Oppenheimer raised its Marvell Technology (MRVL) price target to $325 from $300 while keeping an Outperform rating, citing management’s upbeat AI-driven outlook. Marvell lifted its calendar-year 2026 revenue guidance to $12B (from $11.5B) and set a 2027 outlook of $18B, with Trainium/Maia programs and interconnect strength supporting growth (custom chip revenue >$2B this year, >2x next year; Maia ~ $700M next year). Second-quarter results also slightly beat expectations ($2.739B revenue, EPS $0.94), and analysts largely reiterated Buy/Overweight ratings despite at least one Neutral call on valuation—keeping overall sentiment moderately positive.

Analysis

MRVL is being priced less like a one-quarter beat and more like a long-dated claim on hyperscaler ASIC budgets. That is attractive if the Google relationship converts into repeat volume, but it also raises concentration risk: once a customer is large enough to matter, the bargaining power shifts back to the buyer, which caps the durability of any premium multiple. The market is likely overvaluing the visibility of 2028-2033 revenue before the milestone economics are independently validated.

The nearer-term catalyst is the October analyst day, where the stock will trade on incremental proof rather than the broad AI narrative. If management cannot show that interconnect and custom silicon growth is already pulling through into firm order cadence, the setup is vulnerable to a fast 15-20% multiple reset because expectations are now too far ahead of confirmed backlog. NVDA is only a marginal loser here; the more relevant second-order effect is that each dollar of custom silicon spend shifts AI capex share toward efficiency-oriented solutions, which can improve cloud margins at GOOGL.

Contrarian view: the consensus is treating every announced program as recurring, linear demand, when much of this is milestone-based and subject to customer redesign risk. The cleaner expression is to own the platform beneficiary with lower execution sensitivity, not the supplier whose valuation is already discounting an uninterrupted climb in revenue and margin. If analyst day merely reiterates the current path without accelerating the 2027+ bridge, the current move is likely too extended to chase.

Key Decisions for Investors

• Long GOOGL / short MRVL for 1-3 months into the Oct. 6 analyst day; express the view that Google captures the margin benefit from custom silicon while MRVL has already priced in near-perfect execution.
• If already long MRVL, trim 25-50% into strength before analyst day and retain only a core position for the 6-18 month optionality; the upside remains, but the multiple is now vulnerable to any wobble in order visibility.
• Do not short NVDA outright on this print; use it only as a relative hedge if follow-up commentary suggests custom ASIC adoption is taking share from merchant accelerators faster than expected.
• Set an alert on MRVL guidance cadence: if October commentary does not add credible 2027-2028 revenue detail or implies slower custom-chip conversion, use that as the trigger to de-risk further or reverse into a short.