BMO raises Marvell stock price target on AI data center outlook

Marvell raised its FY2028 revenue target to $20 billion from $18 billion, including data center revenue of $18 billion versus about $15 billion previously, and outlined FY2031 revenue of $70 billion–$90 billion with EPS above $30. BMO raised its price target to $325 from $250 but maintained Market Perform; the stock was quoted at $284.39 after rising 231% over the past year. Other firms also lifted targets to $370–$425, citing AI, custom silicon and interconnect growth, while InvestingPro characterized the shares as overvalued relative to its Fair Value estimate.

Analysis

The key risk is not whether AI spending continues, but whether Marvell converts that spending into durable, profitable design wins. A data-center growth rate that outpaces customer capex implies share gains, richer content per system, or both; if growth instead tracks capex, the long-range revenue framework becomes harder to underwrite. Custom silicon also has a two-sided effect: hyperscalers’ in-house chip programs create demand for Marvell’s design and connectivity work, while increasing customer bargaining power and the risk that programs are delayed, redesigned, or sourced elsewhere. Broadcom is a relevant competitor for custom silicon, while in-house efforts can also limit the addressable share for merchant suppliers.

The price-target dispersion is more informative than the headline: BMO’s $325 target is only modestly above the quoted share price, while higher targets depend on successful execution of multi-year assumptions. The cited “nearly 200%” framing is not supported by the article’s quoted current price and BMO target. After a large prior rerating and upward estimate revisions, execution misses or slower hyperscaler spending could compress the multiple before reported revenue visibly weakens. Near term, monitor customer program commentary and quarterly data-center growth; over 6–18 months, verify design-win conversion, revenue concentration, and whether growth yields corresponding earnings and cash generation. The thesis weakens if guidance or design-win timing slips, or data-center growth materially trails the company’s trajectory.

Key Decisions for Investors

• Do not chase on the analyst-target headlines alone. Treat the long-range revenue and EPS framework as management targets, not independently verified demand; the current evidence supports a watchlist, not a high-conviction valuation call.
• For investors seeking exposure, consider staged entry after a pullback or after a quarterly report confirms data-center growth and earnings conversion. Define a thesis break around a guidance reduction, delayed major programs, or sustained deceleration in data-center growth—not a single volatile trading session.
• Prefer a relative-value expression over an unhedged momentum purchase if conviction is specifically in Marvell’s custom-silicon and interconnect execution: long Marvell against a diversified semiconductor basket, sized modestly. Reassess if customer concentration rises or growth fails to outpace the relevant spending pool.
• Track hyperscaler capex commentary, disclosed design-win ramps, segment growth, and cash conversion over the next several quarters. These are the missing checks needed to distinguish durable share gains from a target-driven expectations cycle.