As of early June 2026, Marvell Technology (MRVL) reached an intraday high of $324.20, setting a new all-time record and briefly pushing its market capitalization above $250 billion. The immediate catalyst for this surge was NVIDIA CEO Jensen Huang’s public statement at Computex Taipei on June 2, 2026, where he said Marvell could become the next tech giant to surpass a $1 trillion market cap. This remark quickly ignited market sentiment, sending the stock soaring in a short period.
Prior to this, Marvell released a record-breaking FY2026 annual report in March 2026: total revenue reached $8.195 billion, up 42% year-over-year; GAAP net income was $2.67 billion. The Q1 FY2027 results, announced at the end of May 2026, showed revenue climbing further to $2.418 billion, a 28% year-over-year increase—another all-time high. As of the June 10, 2026 close, MRVL traded at $252.59, marking a significant pullback from its record high.
However, after the initial euphoria, the core question investors should focus on is this: Is Marvell’s growth driven by short-term catalysts like NVIDIA’s endorsement, or does it rest on the sustained industrial logic of the AI infrastructure supercycle?
Assessing Growth Quality Through Financials: AI Demand Is No Longer Just a "Story"
Marvell’s FY2026 financial performance (fiscal year ending February 1, 2026) provides a direct anchor for understanding its current market position. Annual revenue hit $8.195 billion, up 42% year-over-year and slightly above the FactSet consensus estimate of $8.185 billion. Net income reached $2.67 billion, and adjusted EPS was $2.84, up 81% year-over-year. Fourth-quarter revenue alone was $2.22 billion, a 22% year-over-year increase.
Marvell Technology FY2026 Key Financial & AI Metrics
| Metric | Figure |
|---|---|
| FY2026 Total Revenue | $8.195 billion (record high) |
| Year-over-Year Growth | 42% |
| Net Income | $2.67 billion |
| Gross Margin | 51.0% (up 9.7 percentage points YoY) |
| Q4 Data Center Revenue | $1.65 billion (75% of total) |
| FY2027 Revenue Guidance | $11 billion (up 34% YoY) |
Data center operations are the clear growth engine. In Q4, data center revenue reached $1.65 billion, up 9% sequentially and 21% year-over-year. Chairman and CEO Matt Murphy highlighted "strong AI demand" as the primary driver of these results.
From a business mix perspective, management disclosed in September 2025 that optical interconnect products account for about 50% of Marvell’s data center revenue, custom silicon about 25%, and the remaining 25% comes from data center storage, switching, and emerging security solutions. This structure means Marvell’s growth logic follows two main tracks: the ramp-up of AI custom chip production and the ongoing need for higher interconnect bandwidth within AI data centers.
Custom AI Chips: The Next Compute Wave After GPUs
Marvell’s position in the AI ASIC arena is shifting from "potential beneficiary" to "tangible winner." Public data shows Marvell holds roughly 20%–25% of the global custom AI ASIC design services market, forming a de facto duopoly with Broadcom—together, they control about 95% of the market. Marvell’s client roster includes Amazon’s Trainium and Inferentia chip programs, and it also supports Microsoft, Meta, and Google’s Arm-based CPU businesses.
This segment’s market is expanding rapidly. According to TrendForce, ASIC’s share of the AI server market will jump from 20.9% in 2025 to 27.8% in 2026, while GPUs will drop from 75.9% to 69.7%. The rising share of ASICs doesn’t mean GPU demand is shrinking; rather, as the overall AI accelerator market grows, custom solutions are gaining priority among cloud providers due to their energy efficiency and cost advantages for specific workloads.
Evolution of AI Server Accelerator Market Structure (ASIC vs GPU)
From this perspective, Marvell’s current data center custom silicon revenue—about one-quarter of the total in FY2026—represents only the early stage of its growth curve. The company targets $900 million to $1.1 billion in AI custom chip revenue. This growth is validated on two fronts: first, large-scale client deals already secured (such as Amazon Trainium/Inferentia); second, the expectation that cloud providers’ capital spending on ASICs will rise from 2% in 2023 to 13% by 2027.
Optical Interconnect: The "Main Artery" of Data Centers
Optical interconnect products make up roughly 50% of Marvell’s data center revenue, forming the company’s core business foundation. This growth is rooted in the technical evolution of AI data center architectures: as AI training clusters scale from thousands to hundreds of thousands of GPUs, chip-to-chip bandwidth and data transfer efficiency have become the main bottlenecks to scaling compute power.
Marvell offers a comprehensive optical interconnect portfolio, including DSPs, retimers, and silicon photonics engines. Its silicon photonics engine supports speeds up to 6.4T, integrating multiple components in a compact module and supporting both pluggable optics and co-packaged optical solutions. Management expects double-digit sequential growth for electro-optical products in Q3 FY2026.
In December 2025, Marvell announced the $3.25 billion acquisition of optical interconnect startup Celestial AI, including $1 billion in cash and about 27.2 million shares of Marvell common stock. The deal is expected to close in Q1 2026. Celestial AI specializes in photonic transmission, enabling high-speed connections between AI chips and memory using light signals instead of electrical ones. CEO Matt Murphy described the acquisition as "transformational" for Marvell, expanding its leadership in AI connectivity as "scaling" becomes the next frontier for AI infrastructure.
Strategically, the rationale is clear: bottlenecks in AI infrastructure emerge and are addressed sequentially. Once compute power nears its limits, data throughput becomes the next constraint. By acquiring Celestial AI, Marvell has secured a strategic early-mover position in optical interconnect technology.
NVIDIA CEO Jensen Huang echoed this view at Computex, calling Marvell’s optical partnership a "match made in heaven" and predicting Marvell would become the next trillion-dollar tech company. While this is an industry leader’s opinion rather than investment advice, it sends a clear signal: the value of optical interconnect in the AI infrastructure chain is being re-rated.
Competitive Landscape and Valuation: A Two-Sided Assessment
No analysis of Marvell is complete without comparing it to Broadcom. Broadcom holds about 70% of the AI ASIC market and boasts a $2.3 trillion market cap, leading both in scale and client coverage. Broadcom expects its AI ASIC addressable market to grow from roughly $12.2 billion in 2024 to $60–90 billion by 2027, and it secured a key OpenAI order in 2025.
By contrast, some analysts believe Marvell’s growth outlook is less robust than management suggests, mainly because Broadcom’s ASIC roadmap is more predictable. Other risks include Marvell’s heavy reliance on Asian markets for over 70% of revenue and potential margin pressure in custom chip operations. These issues require ongoing monitoring and can’t be resolved with a single solution—they must be reassessed each earnings cycle.
Marvell vs. Broadcom—Custom AI ASIC Duopoly
| Dimension | Broadcom | Marvell Technology |
|---|---|---|
| Combined AI ASIC Custom Design Market Share | About 95% (duopoly) | About 95% (duopoly) |
| Individual Market Share | 1st, ~55%–70% | 2nd, ~13%–25% |
| Core Clients | Google (7th-gen TPU), OpenAI (10GW partnership) | Amazon (Trainium), Microsoft (Maia) |
| Key Differentiators | AI network switch chips | Optical interconnect, data center DSP |
Analyst ratings for Marvell have trended upward, but target prices vary widely. As of late May 2026, the 12-month average target from 38 analysts was about $208.64, with 32 buy ratings and 6 holds. Following the stock’s sharp rally, several firms raised their targets: Stifel hiked its target from $230 to $321; Benchmark raised to $275; KeyBanc to $260; Raymond James from $105 to $235. Citi made the largest jump, from $118 to $215. The broad range of targets reflects divergent views on Marvell’s valuation rather than a strengthening consensus.
After COMPUTEX, MRVL shares experienced significant volatility, plunging over 16% in a single day from a record $316 to around $263. This wasn’t due to a fundamental shift but rather a natural correction after a rapid run-up—shares jumped nearly 50% in just three days, from $219.43 on June 2 to an intraday high of $324.15 on June 4. Such swings underscore that price moves driven by extreme sentiment are rarely sustainable.
Gaining Exposure to Global Tech Assets via Gate
For investors interested in Marvell and other US tech stocks, Gate’s recently launched US stock trading feature offers a compelling access point. Users can trade over 10,000 stocks and ETFs listed on the NYSE and NASDAQ directly with USDT in their Gate accounts—no platform switching or fund transfers required—enabling seamless cross-market allocation within a crypto account.
Gate’s US stock trading is based on compliant broker channels and involves real spot stock transactions, not CFDs. Funds are managed separately, with zero holding costs, no swap or overnight fees, and dividends paid directly in USDT. For users already familiar with crypto trading, Gate’s stock interface mirrors its crypto spot trading logic, lowering the learning curve for cross-market investing.
With Marvell as an example, investors can use Gate to trade MRVL shares directly and benefit from its AI growth story. At the same time, they can flexibly adjust allocations when crypto market opportunities arise, optimizing asset management across markets. In today’s interconnected capital markets and shifting tech stock valuation paradigms, this trading approach offers long-term strategic value beyond short-term arbitrage.
Conclusion
In FY2026, Marvell Technology delivered a record $8.195 billion in revenue and 42% year-over-year growth, demonstrating that AI demand is now a quantifiable driver for its business. Its three core pillars—custom AI chips, optical interconnect, and data center networking—align with different segments of the AI compute value chain, complementing rather than replacing each other.
However, high returns often come with high volatility, especially for tech companies whose market caps have multiplied rapidly. Marvell faces ongoing challenges from competitive pressures (notably from Broadcom), client concentration risks, and the long-term margin outlook for its custom chip business. These variables require continuous monitoring and reassessment each earnings cycle, rather than betting on a single event-driven rally or one industry leader’s comment.
