Marvell Technology (MRVL) has recently experienced a strong rally fueled by the AI infrastructure narrative, with its share price reclaiming the $300 mark and now just shy of its all-time high. As of June 16, 2026, according to Gate stock market data, MRVL is currently trading at $308.88, up 10.43% today, marking its return above $300 after a week.
Since the beginning of 2026, MRVL shares have surged by nearly 140%—158%, with several single-day gains exceeding 30%. However, this rapid climb has sparked intense debate over valuation bubbles and potential technical corrections. Whether MRVL can reach new highs in June depends on whether structural demand for AI data center construction continues to translate into verifiable performance growth, and whether its current lofty valuation has already overdrawn future growth expectations.
What’s Driving MRVL’s Return to $300?
The primary catalyst behind MRVL’s latest rally came from NVIDIA CEO Jensen Huang’s public remarks at Computex 2026 in early June. Huang stated that Marvell could become "the next trillion-dollar company," which sent MRVL shares soaring over 25% pre-market on June 2, 2026 (Eastern Time). Following this, Stifel analysts quickly raised their 12-month price target from $230 to $321 and reiterated their buy rating, citing increasing market recognition of Marvell’s position in the data center and AI investment cycle. Raymond James also raised its target from $105 to $235, maintaining a strong buy rating. These upgrades reinforced bullish consensus on MRVL, propelling its price from around $220 to near $300.
From a business perspective, Huang’s comments are not mere hype. As AI data centers scale up to tens of thousands or even millions of GPU/XPU clusters, high-speed interconnect communication among thousands of chips becomes as critical as computing power itself. Marvell is a global leader in both DSP (digital signal processing) chips and switching networks, with products spanning coherent optical DSPs, PAM DSPs, and Teralynx series switching chips—covering mainstream technology routes. Its switching chips are already widely deployed in the world’s largest cloud data centers, with millions of ports installed. Additionally, NVIDIA invested $2 billion in Marvell and integrated it into the NVLink Fusion ecosystem as early as March 2026, a strategic move made before Huang’s public endorsement, signaling that industry insiders had already recognized Marvell’s technical value. Thus, the $300 breakout reflects a concentrated market repricing of established facts, rather than a sentiment-driven surge.
How Strong Is the Fundamental Support for AI Data Center Business?
Marvell’s AI data center business has moved from the "growth story" stage to delivering quantifiable results. According to official financial reports, for Q1 of fiscal year 2027 ending May 2, 2026, total revenue reached $2.418 billion, up 28% year-over-year and beating Wall Street consensus. Data center revenue hit $1.833 billion, now accounting for 76% of total revenue, marking Marvell’s complete transformation from a legacy storage and networking chipmaker into a pure-play AI infrastructure giant.
Profitability remains robust. Non-GAAP gross margin held at a high 58.9%, and Non-GAAP EPS reached $0.80, perfectly matching Wall Street expectations. More importantly, management issued guidance far above estimates: median Q2 FY2027 revenue is projected at $2.7 billion, implying roughly 35% year-over-year growth. At the Evercore TMT Summit, Marvell further raised its full-year FY2027 revenue guidance to about $11.5 billion, with interconnect business growth expectations lifted from around 50% to over 70%. The FY2028 revenue target is set at $16.5 billion, with custom ASIC business aiming for over $4 billion in 2028.
Looking further ahead, management expects that if Marvell captures roughly 20% of the custom silicon market, annual custom silicon revenue could reach $10–$11 billion in FY2029. These projections are backed by concrete execution paths: each generation of PAM DSP maintains a lead in mass production (1.6T was widely deployed in 2026, with 3.2T samples planned for 2027), capacity for key components like lasers is secured, and orders have already translated into visible growth trajectories.
Why Are Wall Street Analysts So Divided on MRVL’s Price Target?
Despite strong fundamentals, the average 12-month price target from 38 Wall Street analysts is about $208.64, with the highest at $300 (HSBC) and the lowest at just $180 (Goldman Sachs)—a $120 gap that reflects deep valuation disagreements rather than doubts about AI infrastructure’s growth trajectory.
Bulls base their arguments on the outlook for AI data center construction and Marvell’s strategic positioning in the industry chain. Stifel raised its target to $321 using a 55x calendar year 2027 P/E, believing Marvell’s role in the data center and AI supercycle is increasingly recognized. Bank of America bumped its target from $200 to $240, forecasting Marvell’s FY2029 Non-GAAP EPS at about $10.02—roughly triple current trailing earnings.
Bears focus on valuation and customer concentration. Goldman Sachs maintains a neutral rating with a 12-month target of just $125, implying about 40% downside from the report’s $208.26 price. Their core view uses a 28x P/E on normalized EPS of $4.50, arguing that the long-term trajectory depends on data center growth targets, Google ASIC partnership progress, and demand shifts from Agentic AI. MRVL’s trailing GAAP P/E is about 66x, Beta is around 2.25, and custom chip revenue is highly concentrated among a few hyperscale customers—all factors that amplify price volatility and downside risk.
What Technical and Capital Variables Could Impact New Highs in June?
Technically, MRVL is in a high-volatility consolidation range between $280 and $320, following a strong AI-driven expansion. Key support lies in the $280–$290 accumulation zone, with resistance at $320–$330, and a breakout above $350 would confirm trend continuation. As the price nears the $300 level, options traders will fiercely battle over breakout versus pullback: if opening volume holds above $300, trend-following bulls will keep entering; repeated failures to break through could trigger profit-taking and increased risk from declining implied volatility.
Capital dynamics are equally important. After a rapid rally in early June, trading volume surged, and market focus has shifted from "finding AI beneficiaries" to "how much further can it go" and correction risks. Notably, while the price jumped after Huang’s comments, market pricing already reflects many long-term optimistic assumptions, not just short-term order support. In other words, even if Marvell’s fundamental narrative remains intact, technical correction pressure is building.
For long-term investors, a key variable is the commercialization pace of next-gen optical technologies like CPO. Current growth drivers are mature product lines such as PAM DSPs, TIA drivers, and 400G/800G/1.6T pluggable optical modules, while higher-value optical scale and CPO are preparing for mass production in 2027. This means much of the AI infrastructure narrative priced in today still relies on next-gen interconnect technologies not yet widely commercialized. The actual progress of CPO commercialization will be crucial for validating sustained growth post-2027.
Has Valuation Overdrawn Future Growth Expectations?
Marvell’s valuation is at the heart of current market disagreements. Based on FY2027 revenue guidance of $11.5 billion, its price-to-sales ratio is already high. If FY2028 revenue hits $16.5 billion, there will be some digestion—provided the company delivers on growth promises. Rosenblatt analysts note that while MRVL’s trailing P/E is about 65.64x, its PEG ratio is just 0.16, making adjusted-for-growth valuation attractive. They expect Marvell to achieve compound revenue and earnings growth rates of 40%–50%+ over the next few years.
Goldman Sachs, however, is cautious, emphasizing that GAAP-based valuation metrics remain elevated, and GAAP net income is significantly affected by rising operating expenses and equity dilution. Q1 FY2027 GAAP net income was only $34.5 million, while Non-GAAP net income reached $718 million—a huge gap mainly due to stock-based compensation, acquisition-related amortization, and integration costs. This means Marvell’s actual earnings quality is still diluted by M&A and option expenses, and when the market uses Non-GAAP metrics for valuation, it is essentially assuming these costs will gradually fade or be offset by growth.
Another angle worth considering: Marvell’s current market cap is close to $264 billion. Assuming a 30% annual growth rate, it would take about five years to reach $1 trillion. Huang’s "trillion-dollar company" comment should be seen as a statement of long-term potential, not a quantified short-term upside.
How Does the Semiconductor Industry’s Macro Environment Affect MRVL?
From an industry cycle perspective, semiconductors are entering a new structural uptrend driven by AI. UBS forecasts global wafer fabrication equipment spending will reach about $147 billion in 2026, climb to $198 billion in 2027, and could hit $247.5 billion in 2028—doubling the market in just three years. More notably, some customers are now providing equipment suppliers with up to eight quarters of demand visibility, something the UBS team has never seen in nearly 30 years covering the industry, suggesting the equipment cycle may last much longer than expected.
In memory, the DRAM supply gap is about 8% in 2026, and NAND about 5%, with shortages likely to persist through 2027. Meanwhile, SEMI has raised its annual outlook, and SK Hynix announced a five-year capacity doubling plan, all pointing to a sustained global semiconductor upcycle.
For Marvell, continued expansion in semiconductor equipment spending means downstream hyperscalers have extended capital expenditure visibility. These hyperscale customers are the core buyers of Marvell’s custom ASICs and optical interconnect products. As long as the global AI infrastructure capex cycle remains intact, Marvell’s end-market demand environment is macro-supported. However, MRVL’s Beta is about 2.25, meaning it will swing much more than the broader market during systemic adjustments. When the industry is booming, returns are amplified, but if macro or sector expectations shift, downside moves are equally magnified.
Conclusion
MRVL’s return above $300 is essentially a value re-rating driven by the AI infrastructure narrative and industry leader endorsements. The record $2.418 billion revenue in Q1 FY2027 and 76% share from data center business validate its transformation into a pure AI infrastructure company. Whether it can hit new highs in June depends on three core variables: first, whether the $280–$290 support zone holds during consolidation; second, whether hyperscale customers’ capex pace and ASIC order fulfillment align with guidance; and third, whether profit-taking at current high valuations can be digested through range-bound trading rather than sharp corrections.
From a long-term structural perspective, Marvell’s product matrix is complete across custom ASICs, optical interconnects, and switching networks, and NVIDIA’s strategic investment and ecosystem integration further cement its irreplaceable role in AI interconnects. In the short term, however, mass production progress of next-gen interconnect technologies like CPO, improvements in GAAP-level profitability, and wide analyst target price divergence mean the $300 threshold remains a fiercely contested battleground for bulls and bears alike.
FAQ
Q1: What is the main reason behind MRVL’s latest rally?
A1: The immediate catalyst was NVIDIA CEO Jensen Huang’s public statement at Computex 2026 that Marvell could become the next trillion-dollar company, along with multiple Wall Street institutions sharply raising their price targets. On the fundamentals side, Q1 FY2027 revenue was $2.418 billion, up 28% year-over-year, with data center business accounting for 76%—the financials support the narrative of Marvell’s transformation into a core AI infrastructure supplier.
Q2: What is Marvell’s core positioning in the AI industry chain?
A2: Marvell’s core role is in the AI data center interconnect layer, mainly covering three areas: custom ASIC chips (designing dedicated AI chips for cloud service providers), optical interconnect products (including DSPs, retimers, optical modules, etc.), and switching network chips (Teralynx series). As AI clusters scale from thousands to millions of cards, inter-chip communication becomes as important as computing power.
Q3: Is MRVL’s current valuation level too high?
A3: There are clear differences depending on the metric. By trailing GAAP P/E, MRVL is about 66x, which is indeed high; but its PEG ratio is around 0.16, making adjusted-for-growth valuation somewhat attractive. The gap between Goldman Sachs’ 12-month target of $125 (neutral rating) and Stifel’s $321 (buy rating) reflects fundamental market disagreements on valuation.
Q4: What are the main risks facing MRVL?
A4: Key risks include: high volatility with a Beta of about 2.25; custom chip revenue highly concentrated among a few hyperscale customers, posing customer concentration risk; next-gen interconnect technologies like CPO are not yet in mass production, so current growth is driven by mature pluggable optical module product lines; GAAP-level profitability is weighed down by stock-based compensation and M&A amortization, leading to a significant gap with Non-GAAP performance.
Q5: What is the long-term significance of CPO technology for Marvell?
A5: CPO (co-packaged optics) integrates the optical engine with the XPU or switch, significantly boosting bandwidth density and reducing power consumption for AI clusters. Marvell has been deeply involved in this field for over a decade, and CPO is seen as the key solution for scaling AI clusters from tens of thousands to millions of cards. However, CPO is expected to enter mass production in 2027, and it will take a lengthy execution phase from early deployment to generating hundreds of millions in annualized revenue by 2028.
