Marvell Technology Inc. shares have fallen roughly 37% from their prior peak even as the chipmaker projects substantial revenue gains tied to networking and optics for artificial intelligence data centers. The disconnect between market sentiment and company guidance highlights tensions over valuation and the pace of AI infrastructure investment.
Company outlook versus market reaction
Marvell, a provider of optical interconnects, Ethernet switches and custom silicon, reported recent annual revenue growth of about 28% year over year. Management now expects full-year revenue to increase roughly 40%, targeting $11.5 billion in total sales.
Much of that growth is expected to come from Marvell’s data center business, which the company projects will expand about 50% in fiscal 2027 and accelerate to roughly 55% in fiscal 2028. The company has also raised near-term forecasts for certain optics products: revenue for near-packaged optics and co-packaged optics was previously projected at $150 million next fiscal year and now is expected to be about $300 million.
What investors are weighing
Despite the bullish outlook on demand, the stock trades at a forward price-to-earnings multiple near 52, a level that many investors view as high for a company outside the top-tier semiconductor names. Recent market behavior has reflected broader caution toward companies positioned to benefit from AI, with investors tempering enthusiasm for aggressive infrastructure spending.
Analysts and market participants are parsing whether the drop represents a buying opportunity or a signal that risks are being underpriced. Factors complicating that analysis include execution risk as customers deploy new optics and switch architectures, the timeline for large-scale AI deployments, and competition within networking components.
- Revenue momentum: Company expects total revenue growth of about 40% to $11.5 billion.
- Data center strength: Data center revenue forecast to rise ~50% in fiscal 2027 and ~55% in fiscal 2028.
- Optics upgrade: Near-packaged and co-packaged optics revenue outlook doubled to $300 million.
Context: why networking matters for AI
As AI models grow in size and complexity, the demand for higher bandwidth and lower latency across data centers increases. Networking components such as high-speed switches and optical interconnects play a central role in enabling parallel processing and reducing communication bottlenecks between accelerators. Marvell aims to position itself as a core supplier in that layered infrastructure.
| Metric | Company Guidance / Current |
|---|---|
| Stock decline from peak | 37% |
| Recent revenue growth (YoY) | 28% |
| Full-year revenue growth target | 40% to $11.5B |
| Data center revenue growth | ~50% (FY2027); ~55% (FY2028) |
| NPO / CPO revenue forecast | Raised from $150M to $300M |
| Forward P/E | ~52 |
The company’s roadmap for near-packaged optics (NPO) and co-packaged optics (CPO) reflects a broader industry push to reduce latency and improve throughput by bringing optical components closer to processing units. That trend can yield sizable demand if hyperscalers and large cloud providers accelerate their adoption, but timing and scale remain uncertain.
Investor considerations and risks
For investors, the decision to buy on the dip hinges on confidence in Marvell’s execution and the durability of AI-driven networking spend. A high forward multiple implies elevated expectations; if revenue growth slows or margins compress, the stock could face further pressure.
Conversely, if the company achieves its forecasts and optics adoption accelerates, current prices could be seen as a discounted entry point. The balance between those outcomes will depend on customer adoption timelines, competitive dynamics, and broader macroeconomic sentiment toward technology capital expenditures.
Marvell’s trajectory will be a test case for how market participants value suppliers that sit behind the more visible AI chipmakers. The company’s guidance signals confidence in demand, but the market has already priced in caution—leaving investors to decide whether the potential upside justifies the premium multiple.
