Marvell Technology's Ambitious Growth Trajectory: A Closer Look at Revised Targets

Marvell Technology upgrades its financial targets, forecasting $11B in FY27 and $15B in FY28, driven by custom AI chips and CXL memory solutions.

Marvell Technology has significantly upgraded its medium-term financial outlook for the second time in just six months. The company's confidence is rooted in two primary growth engines: custom-designed artificial intelligence (AI) chips for major cloud providers and advanced interconnect solutions aimed at solving memory bottlenecks in modern AI computing clusters.

Record Performance and Surging Ambitions

The foundation for these bold projections is a robust fiscal year 2026. Marvell reported record net revenue of $8.2 billion, with GAAP net income reaching $2.67 billion. Building on this momentum, the company has set increasingly aggressive targets.

For fiscal year 2027, Marvell is now targeting revenue of approximately $11 billion. This represents a growth rate exceeding 30% compared to the prior year and marks a notable increase from previous guidance of $9.5 billion in September 2025 and $10 billion in December. Looking further ahead to FY28, the goalpost has been moved to $15 billion in revenue, implying an annual growth rate of roughly 35%.

A Three-Pillar Strategy for Expansion

Company leadership attributes the projected FY28 growth to three core pillars. The first is the continuation of existing custom ASIC (Application-Specific Integrated Circuit) projects. The second involves the ramp-up of several XPU programs, which are experiencing particularly strong demand for custom network interface cards and Compute Express Link (CXL) applications. The third pillar is a new flagship XPU project entering mass production.

Marvell anticipates that CXL expanders and tailored NICs alone will contribute over $2 billion to revenue by fiscal year 2029.

Introducing Structera S: A Key Innovation

Central to the CXL growth strategy is the Structera S 30260, a 260-lane CXL switch enabling memory pooling at the rack level. This device allows for the dynamic allocation of memory resources across CPUs, GPUs, XPUs, and other accelerators without requiring costly platform replacements.

According to Marvell, CXL pooling technology can improve inference throughput by a factor of 4.8 and reduce the time to first token by 82.7%. Customer sampling for the Structera S is scheduled to commence in the third calendar quarter of 2026.

Market Sentiment and Underlying Risks

Despite this strong operational backdrop, Marvell's shares have faced pressure, declining approximately 11% since the start of the year amid a broader semiconductor sector sell-off. Analyst sentiment, however, remains generally positive. The median price target among 28 covering analysts sits at $120, with a range spanning from below $90 to over $150 per share.

A significant risk factor remains Marvell's customer concentration. A substantial portion of its revenue is tied to a handful of large cloud and AI clients. The company's ability to hit its ambitious FY27 and FY28 targets will be largely determined by the ongoing investment appetite of these key customers in AI infrastructure. For the first quarter of FY27, Marvell expects revenue of about $2.4 billion.