Marvell Crashed Below $200: This Wall Street Firm Thinks It Doubles From Here

Marvell (MRVL) trades below $200 versus a $252.56 consensus target (~22.4% implied upside) after a sharp selloff—shares are down 33.21% over the past month and 10.96% in the past week—driven by hyperscaler capex anxiety and competitive ASIC pressure. Despite the price decline, fundamentals held: Q1 FY2027 revenue hit a record $2.418B (+27.6% YoY) and non-GAAP EPS of $0.80 beat consensus; guidance calls for Q2 revenue of $2.70B (~35% growth). KeyBanc reiterates a highly bullish $400 target (upside ~93.9%) on a custom AI accelerator design-win pipeline, including an Amazon Trainium 3 volume ramp and Google “Merope” LPU program potentially generating up to $12B over its lifecycle, with management citing exceptional AI bookings.

Analysis

This looks more like a duration reset than a fundamental break. When the market discounts 12-24 month AI ramp assumptions, the stock can trade like a project-finance asset: small changes in customer timing compress the multiple far more than the underlying revenue miss would justify. That means the next clean confirmation point matters more than the recent selloff itself.

The bigger second-order issue is substitution inside the AI spend stack. If hyperscalers pause custom ASICs, that capital does not vanish uniformly: some of it shifts toward merchant accelerators and networking, which is relatively better for NVDA and less bad for AVGO than for MRVL. MRVL is the most levered name here because its upside depends on several large programs converting on schedule rather than just overall AI capex staying elevated.

The bear case is a delay cascade, not a one-quarter disappointment. Falsifiers would be another downward guide revision, weaker data-center backlog conversion, or evidence that large custom programs are slipping into 2027. Over 6-18 months, the setup is constructive only if the market starts treating these wins as durable platforms rather than timing noise.

Contrarian view: consensus may be underestimating how much bad news is already in the tape after a sharp de-rating. But the bull crowd is also glossing over valuation fragility; at this multiple, MRVL needs execution cadence, not just headline design wins, to outperform.

Key Decisions for Investors

• Buy MRVL on weakness into the next earnings print via a 3-6 month call spread; target a move back toward consensus valuation if management reaffirms the 2H ramp. Risk: exit if the data-center outlook is cut again.
• Relative-value: long MRVL / short AVGO as a beta-neutral custom-AI recovery trade only after a confirming guide. MRVL has more upside torque if timing stabilizes; AVGO is the hedge if the market keeps de-risking semis.
• Use NVDA as the upside hedge to a failed custom-silicon thesis: if hyperscaler budgets are reallocated away from ASICs, NVDA should outperform MRVL on incremental GPU substitution. This is a better hedge than shorting broad semis outright.
• Set a hard alert on the next revenue guide and backlog commentary; if data-center growth is revised down, cut any MRVL long immediately. The thesis is invalidated by schedule slip, not by headline volatility.