Prediction: Marvell Technology Stock Will Jump 150% in 5 Years. Here’s the Math.

Marvell reported fiscal Q2 revenue growth of 37% year over year to $2.74 billion and non-GAAP EPS growth of 40% to $0.94, but its shares sold off and are now 26% below their mid-June 52-week high amid investor disappointment over fiscal 2028 expectations. Management raised fiscal 2028 revenue guidance by $1.5 billion to $18 billion and expects fiscal 2027 revenue to increase 45% to roughly $12 billion, supported by custom AI processors and networking demand. Consensus projects EPS to rise 48% to $4.20 in fiscal 2027, while the article argues continued AI and optical-networking expansion could support substantially higher long-term earnings and valuation.

Analysis

MRVL’s selloff creates an asymmetric setup only if investors distinguish design-win value from near-term shipment timing. Custom AI silicon carries materially higher customer-concentration and program-execution risk than merchant networking, but successful ramps can convert a single hyperscaler engagement into multi-year revenue visibility; the market is likely applying a discount until volume commitments, yields, and gross-margin progression are independently visible. The key competitive read-through is that incremental ASIC adoption is more disruptive to NVDA’s premium accelerator economics than to AVGO’s existing custom-silicon franchise, while MRVL remains the higher-beta, less-proven beneficiary.

Over the next 1-3 months, estimates can move higher on additional customer disclosures or optical/switch design wins, but the stock will remain vulnerable if revenue upside is purchased through lower pricing, expedited supply-chain costs, or mix-driven gross-margin dilution. The relevant confirmation is not another TAM statistic: it is sequential data-center revenue growth alongside stable or expanding gross margin and evidence that receivables/inventory are not rising faster than sales. A weak capex signal from GOOG or other hyperscalers would disproportionately impair MRVL because a smaller customer base leaves less ability to offset a delayed program.

The five-year bull case embeds both unusually high earnings compounding and a still-growth-like terminal multiple; this is not a conservative valuation framework. Even modest misses to ramp timing or a normalization in AI infrastructure multiples can sharply reduce implied equity value, particularly versus AVGO, which has broader earnings support. Consensus may be underestimating the strategic value of scale-up optical connectivity, but it is also underpricing the possibility that customers vertically integrate or dual-source critical ASIC and interconnect functions.

Key Decisions for Investors

• Initiate a staged long MRVL position over the next 2-4 weeks, capped at half normal size until the next earnings report confirms data-center growth and gross-margin resilience. Target a 6-12 month rerating on estimate revisions; exit if management cuts its next-year revenue outlook or quarterly non-GAAP gross margin declines more than 150 bps sequentially without a credible mix explanation.
• Prefer a relative-value expression: long MRVL / short AVGO in equal dollar amounts for 3-6 months only after MRVL shows a technical base. This isolates potential upside from incremental custom-AI and optical design wins while reducing broad AI-capex and semiconductor-multiple risk; cover the pair if AVGO announces comparable incremental ASIC wins or MRVL fails to secure a named scale-up customer.
• Use MRVL call spreads rather than outright long-dated calls ahead of the next earnings release: buy 6-9 month at-the-money calls and sell calls 20-25% above spot. The structure captures an upside guidance revision while limiting premium exposure to implied-volatility compression; avoid if implied volatility is already above the prior four-quarter earnings range.
• Set a GOOG capex and AI-infrastructure procurement alert. Any evidence of deferred accelerator/networking deployments is a near-term reason to reduce MRVL exposure, while sustained capex acceleration plus MRVL customer/program disclosure would justify moving from a tactical position to a 12-18 month core long.