Jensen Huang’s Seal of Approval: Marvell Guidance Raised Again, Targeting $16.5 Billion in Revenue by 2028

TL;DR

• Marvell raised its fiscal 2027 revenue guidance to $11.5 billion, with interconnect business growth forecast accelerated to over 70%.
• Current growth is primarily driven by mature, proven out-of-scale businesses, while in-scale technologies like CPO still await post-2027 validation.
• Related tickers: MRVL (US Stock), AVGO (US Stock), COHR (US Stock)

Following Jensen Huang's enthusiastic promotion, Marvell reaffirmed exceptionally strong AI-related orders at the Evercore TMT Conference, raising its fiscal 2027 revenue guidance to approximately $11.5 billion and lifting interconnect business growth expectations from around 50% to over 70%. Custom ASIC is expected to contribute roughly $2 billion this year, with a target to double next year. The company's stock has surged nearly 140%-158% since the start of 2026, significantly outperforming peers and the broader market, with multiple single-day gains exceeding 30% and noticeably higher trading volumes.

Many reports interpret this series of signals as Marvell solidifying its position as the AI interconnect leader, with the data center business becoming its absolute core. However, the current growth driver remains its proven out-of-scale businesses, leveraging the mass production advantages of PAM DSPs, TIA drivers, and pluggable optical modules. Meanwhile, the higher per-cluster value and strategically more significant in-scale optics and CPO (Co-Packaged Optics) are still being prepared for mass production in 2027. The execution certainty from early deployment to generating hundreds of millions in annualized revenue by 2028 is the key to determining if growth momentum can extend to 2028, whether the company can transition from a component supplier to a strategic partner for cloud providers, and if the current high valuation already fully reflects optimistic expectations.

Summit Update Confirms Strong AI Orders

Marvell's guidance upgrade at the summit was based on realized AI orders and supply chain commitments, not long-term visions. Fiscal 2027 revenue guidance was raised to approximately $11.5 billion, with a further upper limit increase to roughly $16.5 billion for fiscal 2028. Specifically, the interconnect business's year-over-year growth for 2027 was raised to over 70%. Custom ASIC is expected to contribute around $2 billion this year, targeting over $4 billion next year. The off-chip CXL, combined with custom NIC, targets over $3 billion by 2028.

These figures are supported by concrete details. Management repeatedly highlighted robust, capex-driven demand, with orders translating into a visible growth trajectory. While lead times for critical components like lasers remain tight, production capacity has been secured. Marvell continues to achieve first-to-market mass production with each generation of PAM DSP (1.6T mass deployment this year, 3.2T sampling next year). This aligns with the path outlined in previous earnings calls, pointing to genuine demand rather than mere optimistic forecasting.

However, the stock price has already priced in much of this narrative with significant gains. The year-to-date return far exceeds peers and indices. Volatility and implied options volatility have also increased following the surge, shifting market discussion from identifying beneficiaries to questioning "how long the rally can last" and the risk of a pullback. This creates a clear contrast: while short-term momentum from orders is visible, market pricing already incorporates a significant degree of medium-to-long-term optimistic assumptions.

Following the strong guidance, investors naturally question which specific businesses are driving this growth and the respective contributions from out-of-scale and in-scale segments.

Current Growth Driver is the Mature Out-of-Scale Business

The current mainstay of the interconnect business remains the out-of-scale segment. PAM DSPs, TIA drivers, and 400G, 800G, 1.6T pluggable optical modules form a mature product line. First-to-market production with each generation helps the company maintain market share, with the TIA and driver business alone reaching an annual revenue scale of roughly $1 billion. Comprehensive coverage from long-haul coherent to lighter-weight high-speed PAM positions the company favorably in the market for horizontal scaling between data centers.

In contrast, the in-scale business is transitioning from startup to acceleration phase. As AI clusters scale from tens of thousands of GPUs or XPUs to millions, the front-end computing is like upgrading an engine, but the back-end interconnect is like expanding a highway from single lanes to multiple lanes. The bandwidth demand *within* a cluster is roughly 10 times that of the front-end. Power and density bottlenecks are driving the industry from traditional pluggable optical modules toward near-package optics and CPO. This shift significantly increases chip loading per rack and expands the total addressable market.

CPO's core lies in integrating the optical engine directly with the XPU or switch, reducing electrical-to-optical conversion losses, thereby enhancing bandwidth density and lowering overall power consumption. This is not a simple component swap but a system-level solution for the new bottlenecks in large-scale clusters. Leveraging SerDes IP and DSP capabilities accumulated through past acquisitions, combined with recently integrated Celestial AI technology, Marvell aims to build an integration advantage in this direction. However, in-scale business contributions remain limited currently, with meaningful volume generation not expected until mass production launches post-2027.

Since in-scale represents higher content per system and market expansion opportunities, how solid is Marvell's technology roadmap, production readiness, and execution capability in this emerging field?

What is the CPO Mass Production Path After the Celestial Acquisition?

The Celestial AI acquisition was integrated in early 2026, with its photonic fabric technology incorporated into Marvell's in-scale roadmap, set to be co-packaged with custom XPUs and in-scale switches. Management expects to enter formal mass production preparation phase in fiscal 2027, targeting annualized revenue of approximately $500 million from the photonic fabric by the end of fiscal 2028, with a subsequent push towards doubling that figure, contributing a cumulative total of roughly $1 billion over 15 months (on a single product basis). This path represents an optimization over earlier outlooks but remains reliant on successful customer validation and manufacturing ramp-up.

Supply chain details suggest increased certainty. While laser lead times are tight, Marvell has locked in relevant production capacity. The company focuses on silicon photonics chip design and integration, not manufacturing lasers itself. The collaboration with NVIDIA on NVLink Fusion further covers silicon photonics, network interoperability between custom XPUs and NVIDIA clusters, and OCTEON baseband support for AI-RAN. This enhances technical compatibility and strengthens credibility in hyperscaler RFPs, making Marvell's heterogeneous computing solutions more likely to be included in next-generation AI factory plans.

If million-scale XPU clusters widely adopt CPO, power consumption decreases and bandwidth density increases, directly raising Marvell's content value per rack and supporting revenue and gross margin expansion. Conversely, if the 2027 mass production ramp-up is slower than expected, customer validation cycles lengthen, or yield and cost curves underperform, the 2028 contribution will be lower than current market expectations, increasing pressure to meet overall guidance. All current statements remain forward-looking; actual shipment and market share data for a more solid verification won't be available until 2027.

With a full-stack layout from SerDes and DSP to optical engines and custom XPUs, backed by ecosystem partners, can Marvell maintain differentiation and convert it into actual market share gains against competition?

Can Marvell Maintain its Differentiation Against Broadcom?

Marvell's leadership in PAM DSP market, comprehensive coverage from long-haul coherent to 1.6T and 3.2T, and integration advantages with custom ASICs and OCTEON are tangible achievements. The partnership with NVIDIA also provides an additional option for heterogeneous computing. These elements make the company competitive in certain hyperscaler projects, particularly in scenarios demanding rapid mass production and ecosystem compatibility.

However, Broadcom maintains a clear edge in network ASICs, the overall scale of optics, and maturity. Broadcom's custom silicon projects are larger in scale with stronger customer stickiness. In overlapping areas of out-of-scale and in-scale, Marvell still needs to rely on higher integration and specific technology pathways to capture incremental market share. The positioning as a "strategic interconnect partner" remains largely an ambition; its eventual realization depends on actual RFP results and shipment breakdown in 2027-2028, not the current breadth of the portfolio.

For investors, this implies Marvell's content value per rack in next-generation hyperscaler tenders could potentially increase by 20-50% from current levels, supporting revenue visibility and gross margin expansion potential. But if CPO ramps slower than guidance, or if Broadcom dominates more custom projects, 2028 growth could fall short of the optimal scenarios embedded in current valuations, increasing downward pressure on the stock. The current high-beta nature amplifies gains driven by narratives while also magnifying volatility if execution falls short of expectations.

Large-Scale CPO Deployment Remains the Biggest Unverified Variable

Despite strong AI orders, leading PAM production, and NVIDIA compatibility providing real medium-term momentum and optionality, CPO as a new technology faces constraints. Yield curves, cost curves, thermal management, supply chain stability during the transition from preparation to mass deployment, and whether hyperscaler capex can sustain its current pace are core constraints for growth sustainability into 2028. Historical experience shows that new technologies typically require several iterations from early production to generating meaningful revenue, and CPO is currently in this early stage.

Marvell has a decent track record of meeting guidance, but the current ramp-up targets are relatively aggressive. The valuation has significantly expanded in 2026, incorporating many best-case scenario assumptions. Increased market discussion and crowding have also pushed up volatility. Broadcom's scale advantage will not disappear soon, so net market share gains require continuous proof. The current phase represents acceleration, but also shows early signs of late-stage crowding—sharp price swings and a shift in discussion focus towards delivery pressure are characteristic of this dynamic.

Investors should look beyond the visions presented at the summit and focus on tracking actual revenue breakdown in 2027, shipment milestones, specific CPO customer validation progress, and hyperscaler capital expenditure rhythms. These variables are most likely to alter current assessments. If 2027 milestones are met strongly, the trend of structurally increasing interconnect content will gain firmer support. If delays occur or market share falls short, the high valuation faces correction pressure. The structural opportunity in AI interconnects is real, and Marvell's full-stack layout provides differentiated options. However, the final outcome hinges on execution details, making fiscal 2027 a critical validation window.