Is Marvell Stock a Buy on the Dip as AI Revenue Soars?

Marvell reported fiscal Q2 revenue up 37% YoY to $2.74B and adjusted EPS up 40% to $0.94, landing just above the outlook midpoint (EPS $0.93; revenue $2.7B). Data center revenue jumped 46% YoY to $2.17B, and management guided fiscal 2027 Q3 revenue to $3.15B (+/-5%) with adjusted EPS of $1.05–$1.15; it projected data center revenue +75% YoY. The article also highlights an Alphabet partnership that won’t be meaningful until fiscal 2029 (calendar 2028) and an updated growth outlook to $12B revenue in fiscal 2027 (+45% growth) with data center growth expected to accelerate to 60%.

Analysis

MRVL’s setup is increasingly a valuation-vs-duration story, not a pure growth story. The market is paying for multiple years of optical and custom-silicon growth already, while the most bullish incremental revenue from the Alphabet design win is still too far out to support near-term rerating. That creates a classic “good news, lower upside” dynamic: if hyperscaler capex merely stays strong instead of accelerating, the stock can de-rate even with double-digit revenue growth.

The second-order winner is the AI infrastructure supply chain tied to high-speed interconnect, but the economic rent may accrue more to the hyperscalers than to MRVL. AMZN, GOOGL, and MSFT can absorb more custom silicon to lower their own inference costs, which is strategically positive for them but caps merchant supplier margin power over time. A broader implication is that optical networking spend should benefit the adjacent ecosystem, while copper/networking legacy content is structurally losing share.

Near term, the risk is a mismatch between consensus enthusiasm and the actual cadence of design-win monetization. The thesis is vulnerable if data-center growth decelerates even modestly over the next 1-2 quarters or if FY28 guidance stops moving higher; at ~34x FY28, the stock has less room for “story” premium expansion than investors assume. The contrarian view is that the dip may not be a buy-the-dip setup unless you have high conviction that hyperscaler capex and share gains both continue; otherwise this is a high-quality franchise priced like a near-perfect execution story.

Key Decisions for Investors

• Do not chase MRVL on post-earnings weakness reversal; wait for a 10-15% pullback or evidence of a second consecutive guide raise before adding. Falsify the bearish view if data-center growth reaccelerates above the current implied run-rate for two straight quarters.
• Use MRVL as a tactical short vs SOXX/SMH into strength if you want semiconductor exposure with less multiple risk: MRVL’s longer-dated upside is more dependent on flawless execution than the broad index.
• For a relative-value AI infrastructure pair, prefer long GOOGL vs short MRVL over the next 6-12 months: GOOGL retains core cash-flow support while its AI custom-silicon optionality is earlier and less fully capitalized than MRVL’s current growth narrative.
• If you need upside participation, buy a 3-6 month MRVL put spread only on rallies back toward recent highs; the risk/reward is attractive because downside can come from multiple compression even without an earnings miss.
• Set an alert on FY2028 revenue or gross-margin revisions: another upward revision supports the long, but any pause in guidance momentum is a clear signal to reduce exposure.