Marvell's Long-term Outlook Upgraded on AI Strengths; Price Target Raised to $195

Overview

Morgan Stanley issued a research report on Marvell Technology (MRVL), upgrading the price target due to improved long-term AI advantage prospects. While quarterly results were largely in line with expectations, the company significantly raised its full-year FY27/28 outlook. Data Center growth expectations increased to 50% for FY27 and 55% for FY28, with AI Interconnect maintaining >50% growth levels, and confidence in Custom Chip/XPU business notably improving. Morgan Stanley accordingly raised the price target to $195 (from $172) while maintaining an Equal Weight rating, noting that while Marvell has broad long-term AI opportunities, the current valuation fully reflects these prospects.

Core views

For the quarter (FY27 Q1), Marvell reported revenue of $2.42 billion, up 9% sequentially and 28% year-over-year, beating the midpoint of guidance and slightly exceeding market and Morgan Stanley estimates. Data Center revenue was $1.83 billion, representing ~76% of the total, up 11% sequentially and 27% year-over-year; Communication & Other stood at $585 million, up 3% sequentially and 29% year-over-year. Non-GAAP gross margin of 58.9% was slightly below expectations, while EPS of $0.80 was in line with expectations. Guidance for the next quarter (FY27 Q2) set revenue midpoint at $2.7 billion (up 12% sequentially and 35% year-over-year), higher than street consensus of $2.601 billion; gross margin expected at 58.75%, and EPS at $0.93, both better than street expectations. More importantly, Marvell significantly upgraded its long-term outlook: Full-year FY27 revenue expectation increased by ~40% to nearly $11.5 billion, with further growth of 45% in FY28 to $16.5 billion. Morgan Stanley adjusted its model accordingly: FY27 forecasted revenue/gross margin/EPS are $11.5 billion/58.5%/$3.98 (previous: $10.87 billion/58.3%/$3.82); FY28 forecasts $16.2 billion/57.2%/$5.93 (previous: $14.29 billion/57.0%/$5.16). Interconnect has become the core growth driver. Morgan Stanley believes the prior ~50% FY27 growth expectation was too conservative, with new guidance of >70% more aligned with AI networking trends. The company is expanding in multiple directions: TIA drivers are expected to reach a $1 billion annualized run-rate within the next few quarters; scale-up optics have risen from a previous ~$150 million outlook to over $300 million; scale-across emerging segment represents a significant future driver. Morgan Stanley now models ~80% Interconnect growth for FY27 and ~45% for FY28, continuing to view network business as Marvell's primary growth source. Confidence in XPU custom chips is also improving. For FY27, the company maintains a ~25% growth expectation, but the FY28 framework is significantly strengthened with expectations for custom revenue to double. Morgan Stanley holds stronger conviction in three main growth vectors: existing XPU business, XPU attach opportunities, and new Tier-1 customer XPU ramp expected to enter in FY28. Although XPU business remains highly selective, positive signals regarding customer demand and project expansion give Morgan Stanley greater confidence in modeling. On the financial front, Marvell continues to increase investments in optical, switching, and custom chips to build a more complete AI connectivity platform. Despite high investment levels, operating leverage remains strong, with FY28 opex growth expected in the mid-to-high teens relative to ~45% revenue growth. Share dilution has increased slightly, moving from 856 million shares to 893 million this quarter, with an additional 22 million shares expected next quarter.

Analysis framework

Morgan Stanley's analysis logic revolves around multi-layered demand for AI infrastructure investment. First, from the demand perspective, AI interconnect represents an increasingly critical component of AI infrastructure, with rapidly rising internal data center network density and cross-datacenter interconnect needs. Second, regarding Marvell's product positioning, the PAM4 optical and 400ZR high-speed interconnect products acquired through the Inphi acquisition align directly with this demand, creating an intrinsic competitive advantage. Morgan Stanley analyzes by business line layering: On Interconnect, focusing on three growth vectors—TIA/driver, scale-up optics, and scale-across—treating their market size and penetration rates as leading indicators. For Data Centers overall, using Data Center capex growth as a benchmark, expecting Marvell's addressable content growth to exceed overall capex growth rates. For Custom Chips, although transparency is lower, tracking progress signals from three major customer projects frames customer demand intensity. In terms of valuation, Morgan Stanley uses a ~40x CY27 ModelWare EPS multiple (including SBC). This is a common level for high-growth AI semiconductors but remains high relative to peers like NVIDIA and AMD—especially considering the significant gap between Marvell's expected FY27 EPS of ~$6 versus NVIDIA's ~$13. Morgan Stanley concludes that while the long-term outlook is positive, recent stock prices have already fully priced in these growth expectations, hence maintaining an Equal Weight rating.

Methodology notes

• Industry/Industrial Analysis FrameworkSupply-demand framework

  Continued expansion of interconnect demand in AI infrastructure investment, where Marvell possesses supply-side competitiveness in high-speed optical interconnect products

  The report analyzes from both demand (AI networks becoming increasingly network-intensive) and supply (Marvell's Inphi optical products) perspectives, arguing that this supply-demand mismatch creates a business opportunity window for Marvell, driving Interconnect growth to outpace overall Data Center capex growth.

• Company Fundamentals and Financial FrameworkWorking capital cycle

  During rapid growth phases, inventory and accounts receivable management significantly impact cash flow

  The report mentions Days Inventory Outstanding (DOI) and changes in accounts receivable. In a high-growth environment, managing fluctuations in these items directly affects free cash flow and operational efficiency, making them key indicators for tracking business health.

• Valuation MethodologyPE/PEG valuation

  In high-growth industries, PE multiples relative to growth rates serve as the core valuation reference

  The report utilizes a ~40x CY27 EPS multiple, but compares it against peer growth rates (comparing Marvell's growth speed against the implied growth expectation embedded in the valuation multiple) to argue that current stock prices either fully price in or slightly overvalue long-term growth.

• Quantitative/Factor/Portfolio TheoryBeta/alpha analysis

  The report mentions beta risk associated with Marvell's 1.6 trillion dollar market cap, which amplifies drawdowns in volatile markets

  High beta implies that during general market pullbacks, Marvell's decline would be larger relative to the market, but it would also see larger gains in strong markets; the report considers this risk relatively controllable when supply is tight and demand is strong.

• Industry/Industrial Analysis FrameworkUpstream-Midstream-Downstream Chain Transmission

  AI infrastructure investment cascades from cloud giants' GPU procurement decisions down to the entire chain including network chips and storage chips

  By tracking cloud capex moderating to +30% while interconnect growth remains above 50%, the report suggests AI network investment is becoming an independent, faster-growing driver, reflecting structural relative growth differences within the industry chain.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Marvell Technology (MRVL.US)

  Primary subject of the report; directly benefits from accelerated AI investment as an AI infrastructure Interconnect and Custom Chip solution provider

  Strengths

  Leading Inphi optical product portfolio; high-speed interconnect products like PAM4/400ZR align with AI network demands; three major XPU custom clients have clear ramp plans; TIA/scale-up new growth vectors offer vast space

  Weaknesses

  Custom chip project low transparency and high risk; increased share dilution (356m -> 393m); R&D investments remain high suppressing margins; valuation still high relative to competitors (e.g., NVIDIA) with massive EPS gap

  Comparison

  Relative to NVIDIA: Marvell is a peripheral supplier rather than a core GPU provider; growth of 40-45% is still lower than NVIDIA's; EPS of ~$6 vs NVIDIA's $13, yet stock price ($198.7) is comparable, suggesting relative overvaluation; Relative to Broadcom and other interconnect suppliers, Marvell offers superior professionalism and product coverage in AI optics

  Risks

  Concentration risk in Custom Chip Tier-1 clients; if AI infrastructure investment corrections occur, Interconnect demand faces pressure; continuous share dilution creates EPS pressure; valuation fragility—if growth expectations miss, downside potential is significant

Key data

• FY27 Data Center Revenue Growth Rate50%Year-over-year growth, new guidance, previously more conservative
• FY28 Data Center Revenue Growth Rate55%Year-over-year growth, accelerating further based on larger revenue base
• FY27 Interconnect Growth Rate>70%Company guidance; Morgan Stanley models 80% increase
• FY28 Interconnect Growth Rate~45%Expected growth slowdown but still above cloud capex growth of 30%+
• TIA/Drivers Annualized Run-rate$1 BillionExpected to reach within next few quarters; previously not a focus
• Scale-up Optical Market Size> $300 MillionFY28 outlook; previously only ~$150 million
• FY27 Custom Chip Growth Rate~25%Consistent with previous expectations
• FY28 Custom Chip Growth ExpectationDoublei.e., >100% year-over-year growth from three growth vectors
• FY27 Full Year Revenue Expectation$11.5 Billion~40% growth, Morgan Stanley adjusted forecast
• FY28 Full Year Revenue Expectation$16.2 Billion~45% growth, Morgan Stanley adjusted forecast
• FY27 EPS Expectation$3.98Previous $3.82
• FY28 EPS Expectation$5.93Previous $5.16
• Price Target$195Raised from $172; based on ~40x CY27 EPS multiple
• Current Stock Price$198.70As of May 27, 2026
• Implied Upside/Downside-1.86%Price Target vs Current Price
• CY27 Valuation Multiple~40xModelWare EPS multiple (includes SBC), approx. 34x non-GAAP EPS
• Data Center Revenue Share~76%FY27 Q1 actual
• Non-GAAP Gross Margin58.9%FY27 Q1 actual, slightly below expected 58.7%
• Q1 EPS$0.80In line with expectations
• Q2 Guidance Revenue Midpoint$2.7 BillionUp 12% sequentially, 35% year-over-year

Impact & implications

The report posits that Marvell's long-term AI opportunities are improving, primarily reflected in two core areas: Interconnect and Custom Chips. For Marvell, this means enhancing its market position as a provider of AI infrastructure network layer solutions; as AI infrastructure investment deepens, network density and cross-domain interconnect demands will continue to release. For investors, Marvell's long-term growth ceiling has been raised, but current stock prices have significantly priced in these prospects. Thus, there are no obvious upward catalysts unless: (1) estimates continue to be revised upwards; (2) Interconnect share/content continues to be captured; or (3) Custom Chip Tier-1 customer ramps exceed expectations. Short-term, Marvell still needs to prove whether it can convert guidance into actual results and maintain margin resilience amidst high investment backgrounds, which will be a key follow-up focus. Relatively speaking, current stock prices remain at a premium compared to peers, particularly similar pricing to NVIDIA despite a huge EPS gap, so Morgan Stanley prefers selecting other targets in AI chip allocations.

Risks

• Custom Chip (XPU) Project Risk: Although confidence has improved, it remains an all-or-nothing business; any delay or reduction in major client projects would significantly impact full-year results
• AI Infrastructure Investment Rhythm Risk: If cloud giants' capital expenditure growth slows or network investment priority declines, Interconnect demand may face pressure
• Storage and Enterprise Storage Market Weakness Continues: Although the company provided no clear recovery guidance, continued downturn in these sub-segments could drag overall growth
• Inventory Risk: Client or channel inventory buildup could offset growth momentum from Interconnect business
• Continuous Share Dilution: High SBC portion in executive compensation causes continuous dilution, pressuring EPS growth
• Valuation Risk: Current 40x CY27 EPS multiple fully reflects long-term AI growth expectations; any growth below expectations or decline in market risk appetite could lead to multiple contraction

What to watch

• Whether subsequent quarters' Interconnect actual growth can maintain >70% levels, specifically if TIA/scale-up accelerate as expected
• Whether FY28 Custom Chip revenue doubling materializes as expected, along with progress on new Tier-1 customer XPU ramps
• Gross margin trajectory under high investment background, and ability to control opex
• Whether Storage and Communication businesses show signs of recovery or continue as drag factors
• Flow of Interconnect orders to major cloud clients and status of long-term contracts
• Memory and margin management status, and whether inventory cycle enters a downward phase