'ASIC Giant' Marvell Sets Record Quarterly Revenue, Raises Guidance Again, CEO Says Data Center Business Is 'On Fire'

marsbitPublicado em 2026-05-28Última atualização em 2026-05-28

Resumo

Marvell Technology, a leading player in custom AI chips and data center connectivity, reported record revenue for its fiscal Q1 2027, driven by explosive demand in its data center business. Revenue reached $2.418 billion, slightly surpassing expectations, though GAAP net income fell year-over-year due to acquisition-related costs. Crucially, data center revenue hit $1.83 billion, making up 76% of the total and growing 27% YoY. The company significantly raised its full-year and next-year guidance, citing "exceptionally strong AI-related orders." Revenue is now projected at ~$11.5 billion for FY2027 and ~$16.5 billion for FY2028. CEO Matt Murphy emphasized that growth in the data center segment is accelerating. The AI Interconnect business, now expected to grow over 70% annually, saw its forecast lifted again due to rising network demands in complex AI models. Additionally, Marvell's custom chip (XPU) business is on a steep growth path, with FY2028 revenue anticipated to double and a target of over $10 billion by FY2029. The company also expanded its strategic collaboration with NVIDIA, focusing on silicon photonics, system integration, and AI-RAN solutions. To secure supply for surging demand, Marvell plans about $1 billion in supplier prepayments this fiscal year, highlighting its long-term capacity planning. Despite the strong results, the stock dipped slightly in after-hours trading.

Marvell delivered a seemingly brilliant report card with record revenue and a significant guidance raise.

After the U.S. market closed on May 27th, AI custom chip, optical communication, and data center interconnect leader Marvell released its Q1 FY2027 financial results and held an earnings call. The data center business continued its explosive growth, leading the company to once again significantly raise its full-year guidance. CEO Matt Murphy directly stated on the call, "Our data center business is on fire", and "orders are exceptionally strong."

Marvell's Q1 FY2027 revenue was $2.418 billion, up 28% year-over-year and 9% quarter-over-quarter, slightly exceeding analyst expectations of $2.41 billion. Non-GAAP earnings per share were $0.80, in line with analyst estimates. However, GAAP net profit was $34.5 million, a significant decline from $177.9 million in the same period last year, primarily due to one-time costs and non-cash amortization related to the acquisitions of Celestial AI and XConn.

The data center business contributed $1.83 billion in revenue, accounting for 76% of total revenue, up 27% year-over-year and 11% quarter-over-quarter.

Following the earnings report and call, the company's stock price fell slightly by about 1%. Year-to-date, the stock had already more than doubled ahead of the earnings report. Against high expectations, merely "meeting expectations" may no longer be enough to impress the market.

01

"AI-Related Orders Are Exceptionally Strong," Marvell Raises Guidance Again

This marks Marvell raising its guidance for multiple consecutive quarters.

For the Q2 FY2027 guidance, the company expects revenue of approximately $2.7 billion (plus or minus 5%), representing year-over-year growth of about 35%, higher than the analyst consensus estimate of $2.6 billion. The non-GAAP EPS guidance range is $0.88 to $0.98, compared to the analyst consensus of $0.90.

For the full-year guidance, Marvell raised its FY2027 revenue expectation to approximately $11.5 billion, representing growth of about 40% year-over-year. Three months ago, the company's guidance was "approaching $11 billion."

More attention-grabbing is the outlook for FY2028. Marvell raised its FY2028 revenue target to approximately $16.5 billion, about $1.5 billion higher than the guidance provided last quarter, representing year-over-year growth of about 45%.

CEO Matt Murphy stated in the earnings release: "We are seeing exceptionally strong AI-related orders, and as a result, we are significantly raising our revenue expectations for Marvell's FY2027 and FY2028, showing substantial improvement compared to the guidance provided last quarter."

02

Data Center: Accounts for 76% of Revenue, Growth Still Accelerating

Q1 data center revenue was $1.83 billion, up 27% year-over-year and 11% quarter-over-quarter, accounting for 76% of total revenue.

Marvell's growth forecasts for this business are:

FY2026: +46% (already achieved)

FY2027: approximately +50%

FY2028: approximately +55%

Murphy stated:

The data center business is on fire, and we expect growth rates to accelerate this year and next, starting from an already high base.

03

Interconnect Business: Growth Rate from 30%→50%→70%, CEO Says "There's Room for Upside"

The AI Data Center Interconnect business is the largest segment within Marvell's data center portfolio, covering optical interconnect, DCI modules, coherent optics, and other product lines.

The expected annual growth rate for this business has been raised consecutively over the past few quarters: around 30% around September last year, later raised to 50%, and now raised again to over 70%.

When pressed by analysts, Murphy said directly:

I think there is a lot of potential for upside here. Our traditional DSP business will see a significant jump next year, the 1.6T product line has higher value content, DCI is accelerating, along with new businesses like retimers, AEC, and scale-up optics... This is the beginning of a major growth cycle for us.

Why has the interconnect business suddenly become so important? Murphy provided a clear logic:

Early generative AI primarily addressed compute and memory bottlenecks, with network interconnect being a secondary concern. However, with the deployment of more complex architectures like inference models and Mixture of Experts (MoE) models, the volume of data transmission within AI clusters has increased dramatically, significantly elevating the importance of network interconnect.

Some key numbers:

TIA and Driver Chips: Expected to surpass an annualized revenue run rate of $1 billion in the coming quarters.

DCI Module Business: Already shipping to all five major U.S. hyperscale cloud providers. Expected to surpass an annualized revenue run rate of $1 billion in FY2028, approximately double that of FY2026 (~$500 million).

Scale-up optics (NPO/CPO Optical Interconnect): Previously expected to be around $150 million; now raised, expected to exceed $300 million in FY2028.

04

Custom Chips (XPU): Doubling Next Year, Targeting Over $10 Billion by FY2029

Marvell's custom chip (Custom/XPU) business is another important growth line and a major focus for the market.

Current Progress:

FY2027 Custom Chip Revenue: Growth exceeding 20% year-over-year.

FY2028 Custom Chip Revenue: Expected to double year-over-year, higher than last quarter's expectation.

FY2029 Target: Exceed $10 billion (previous target was approximately $8 billion).

Analyst Vivek Arya (Bank of America Securities) pressed during the call: Does this mean FY2028 custom chip revenue will be over $4 billion, then jump to $10 billion in FY2029, implying a single-year incremental revenue of $5-6 billion?

Murphy's response was: Yes, you heard that correctly.

Three drivers for FY2028 custom chip growth:

Continued growth from existing flagship XPU projects.

More than ten XPU satellite projects (NIC, CXL, etc.) entering higher-volume production stages, with demand consistently exceeding expectations.

A new leading XPU project entering production — Murphy stated, "The project is progressing smoothly, and the full-year production plan is in place."

Murphy also revealed that while newly won design projects typically require about a two-year development cycle before contributing revenue, the significance of these projects lies in providing security for longer-term growth, which he referred to as "insurance policies."

05

Expanding Collaboration with NVIDIA, Three Key Directions Materializing

This quarter, Marvell announced an expanded strategic collaboration with NVIDIA. Murphy detailed three core directions on the call:

1. Optical Interconnect Collaboration: Marvell has long provided DSP, TIA, and drivers to NVIDIA. The two companies are now further collaborating on developing silicon photonics technology, considered a key enabling technology for scale-up networks.

2. NVLink Fusion Integration: Allows Marvell to build custom chips and network semiconductors that seamlessly interface with NVIDIA's infrastructure. Murphy stated this provides hyperscale cloud providers with greater flexibility to freely mix and match between custom chips and NVIDIA chips. "Marvell uniquely provides the bridge between these two architectures," creating new market opportunities for both companies.

3. AI-RAN: Marvell will enhance its Octeon base station processors to enable them to work directly in tandem with NVIDIA GPUs, running both 5G/6G wireless workloads and AI applications on the same hardware platform.

06

Supply Chain: Locking in Capacity Early, ~$1 Billion in Prepayments Planned This Year

Facing continuously climbing demand, supply chain management has become a critical variable.

CFO Willem Meintjes revealed on the call that the company plans to make approximately $1 billion in supplier prepayments during FY2027, with the first payments starting in Q2. These prepayments will be credited against future material purchases.

COO Chris Koopmans explained Marvell's supply chain strategy in response to analyst questions:

Everything related to AI has been supply-constrained from day one. Our approach is to build extremely close relationships with a small number of core suppliers, provide them with five-year demand forecasts, deliver on our promises every time, and back up our forecasts with action and prepayments.

Financially, Q1 operating cash flow reached a record $639 million. The company repurchased $200 million worth of stock in the quarter and paid $54 million in dividends. As of the end of Q1, total debt stood at $4.96 billion, with net debt/EBITDA at 0.32x.

This article is from the WeChat public account "Wall Street Insights Max," author: Long Yue

Perguntas relacionadas

QWhat was the key factor behind Marvell's significant increase in its revenue guidance for fiscal years 2027 and 2028?

AThe key factor was exceptionally strong AI-related orders, which led the company to significantly raise its revenue outlook compared to the guidance provided the previous quarter.

QWhat percentage of Marvell's total Q1 revenue came from its Data Center segment, and what was its growth rate?

AMarvell's Data Center segment contributed 76% of its total Q1 revenue, with a year-over-year growth of 27% and a sequential growth of 11%.

QHow did Marvell describe the growth trajectory for its AI Data Center Interconnect (Interconnect) business?

AMarvell described the growth trajectory as accelerating significantly, with the annual growth rate expectation being raised from around 30% to 50%, and now to over 70%. The CEO indicated there is substantial upside potential remaining.

QWhat is Marvell's stated revenue target for its Custom Chip (XPU) business by fiscal year 2029, and what major driver was mentioned for growth in fiscal year 2028?

AMarvell's stated revenue target for its Custom Chip (XPU) business by fiscal year 2029 is over $10 billion. A major driver for growth in fiscal year 2028 is the expected doubling of its custom chip revenue, aided by a new major XPU project entering production.

QWhat are the three core areas of Marvell's expanded strategic collaboration with NVIDIA, as detailed in the earnings call?

AThe three core areas are: 1) Collaboration on optical interconnect and silicon photonics technology. 2) NVLink Fusion Integration, allowing Marvell to build custom chips and networking semiconductors that interface seamlessly with NVIDIA infrastructure. 3) AI-RAN, enhancing Marvell's Octeon base station processors to work directly with NVIDIA GPUs for simultaneous 5G/6G and AI workloads.

Leituras Relacionadas

How to Define "Real U.S. Stocks": Differences Between On-Chain Tokens, Price Contracts, and Direct Broker Connections

**Title:** Defining "Real US Stocks": Differences Among On-Chain Tokens, Price Contracts, and Broker-Direct Access **Summary:** In 2026, using stablecoins to purchase US stocks is mainstream, but products marketed as "buying US stocks with USDT" offer fundamentally different assets. This article analyzes three primary models. **1. Tokenized Stocks:** These are on-chain tokens representing economic exposure to underlying stocks, held by an issuer or custodian. They offer benefits like 24/7 trading and DeFi composability (e.g., use as loan collateral). However, users lack direct legal shareholder status; dividends may not be paid in cash, and voting rights are typically non-binding advisory expressions. Examples include platforms like Ondo Finance. **2. Stock Futures / Equity Perpetuals:** These are derivative contracts tracking a stock's price, allowing leveraged long/short positions 24/7, similar to crypto perpetuals. They offer high efficiency and flexibility but involve funding fees, which can be a significant long-term cost, especially during strong trends. Crucially, they confer no ownership rights (dividends, voting) to the holder. **3. Broker-Direct Model:** This model provides access to real securities via licensed broker-dealers. Stocks/ETFs are bought and held within the US clearing and custodial system (e.g., DTCC), making it the only path to genuine stock ownership. Users receive cash dividends and formal proxy voting rights (where applicable). It supports thousands of stocks and ETFs, far exceeding the coverage of the other two models. Key advantages include no funding fees, a clean cost structure for long-term holds, and the potential to transfer holdings to other brokers. Some platforms facilitate stablecoin (USDT/USDC) deposits, reducing reliance on traditional banking. A critical distinction exists *within* the broker-direct model: the underlying brokerage architecture (e.g., Fully Disclosed IB, Omnibus IB, Self-Clearing) determines how client assets are held, protected, and how safeguards like SIPC insurance are conveyed. Users should verify the specific clearing structure and regulatory compliance of any platform. In conclusion, "buying US stocks with USDT" can mean holding an on-chain economic proxy (Tokenized Stocks), trading a price derivative (Stock Futures), or owning the actual security (Broker-Direct). For users seeking full ownership rights and long-term investment, the broker-direct model is the definitive choice, though its implementation details require careful scrutiny.

marsbitHá 59m

How to Define "Real U.S. Stocks": Differences Between On-Chain Tokens, Price Contracts, and Direct Broker Connections

marsbitHá 59m

NVIDIA Launches DSX Platform, Expanding into AI Factory Infrastructure

NVIDIA has unveiled the DSX platform at its GTC Taipei event, marking a strategic expansion from GPU sales into comprehensive AI factory infrastructure solutions. The platform addresses challenges like power supply, cooling, and resource orchestration as AI models scale, shifting the industry focus from single-chip performance to overall infrastructure efficiency. DSX integrates NVIDIA's chips, systems, software, and partner technologies to cover the entire AI factory lifecycle—from design and simulation to deployment and operations. It aims to accelerate deployment, improve reliability and operational efficiency, and reduce the cost per generated token in AI inference. The software suite includes DSX MaxLPS, which uses 45°C liquid cooling and rack-level optimization to allow up to 40% more GPUs per megawatt, and DSX OS, an open-source platform for AI factory operations. The platform also encompasses reference designs, digital twin simulation (DSX Sim), dynamic workload adjustment based on grid conditions (DSX Flex), and data exchange between systems. Early adopters include cloud providers like CoreWeave and Lambda. Major hardware partners, including Dell, HPE, Lenovo, and Supermicro, are developing DSX-ready systems. Pilot projects for DSX Flex are underway with energy providers. Strategically, DSX represents NVIDIA's ongoing transition from an AI chip supplier to a full-stack AI infrastructure platform provider, aiming to set industry standards and solidify its market leadership.

marsbitHá 1h

NVIDIA Launches DSX Platform, Expanding into AI Factory Infrastructure

marsbitHá 1h

After Burning Tens of Billions of Dollars in Tokens, Silicon Valley Giants Start Limiting Employee Token Usage

After burning tens of billions of dollars on AI tokens, major Silicon Valley firms are now restricting employee usage. Companies like Microsoft, Uber, and Salesforce, which heavily promoted AI for "efficiency," are facing a cost crisis. The practice of "tokenmaxxing"—pushing employees to maximize AI tool usage—led to wasteful spending on trivial tasks like checking the weather or writing birthday messages, with studies showing significant hidden costs for bug fixes and code rewrites. The core issue is a misalignment between individual productivity gains and actual business value. While employees use AI to automate tasks they dislike, such as writing reports, this often doesn't translate to increased company revenue or improved core business outcomes. For instance, AI-generated code speeds up development but also sees an 800% increase in "code churn" (code being discarded or rewritten). As a result, only 14% of CFOs report seeing a clear, measurable return on AI investments. Firms are now shifting strategies. Microsoft has revoked most internal licenses for Claude Code, while others are implementing monitoring and cost controls. New tools from companies like Harness and CloudZero aim to track AI spending and tie costs to business results. Some AI vendors, like HubSpot, are moving from token-based pricing to charging based on outcomes, such as "resolved conversations" or "leads generated." This represents a necessary correction in the AI adoption cycle. The challenge now is for companies to move beyond using AI merely to speed up old tasks and instead rethink their workflows and business models fundamentally. The future of enterprise AI depends on proving its value, not just its usage.

marsbitHá 1h

After Burning Tens of Billions of Dollars in Tokens, Silicon Valley Giants Start Limiting Employee Token Usage

marsbitHá 1h

Trading

Spot
Futuros
活动图片