Marvell Technology in AI’s Expanding Multiverse

Key Points

• Marvell Technology is gaining fresh market attention as investors look beyond NVIDIA in the AI trade.
• The company’s data centre business is now the core growth engine, supported by demand for custom silicon, optical networking, and interconnects.
• NVIDIA CEO Jensen Huang’s bullish comments have added a strong halo effect to Marvell’s AI narrative.
• The main risk is valuation, as the stock now needs strong execution to justify rising expectations.

Marvell Technology may be one letter away from the superhero universe, but its role in the AI boom is becoming harder to ignore.

NVIDIA still plays the lead in the AI blockbuster. It owns the strongest GPU story, the deepest software ecosystem, and the clearest claim on the first wave of AI infrastructure spending. Yet the AI trade is starting to look less like a one-hero story and more like a full multiverse.

Data centres don’t run on GPUs alone. They need custom silicon, optical connectivity, switching, storage, high-speed interconnects, and networking systems that can move huge amounts of data without choking the whole machine. That is where Marvell enters the frame.

The company reported record Q1 fiscal 2027 revenue of $2.418 billion, up 28% year on year, and guided Q2 revenue to $2.7 billion at the midpoint, representing expected growth of 35% year on year.

CEO Matt Murphy said revenue growth should continue accelerating through fiscal 2027, driven by strength in the data centre business.

That gives Marvell Technology a cleaner AI angle than a simple ‘NVIDIA alternative’ label. It’s not trying to replace the lead character. It’s building parts of the world that the lead character needs to function.

NVIDIA Still Top Billing, but the Cast is Expanding

The AI market still begins with NVIDIA, but it no longer ends there.

NVIDIA CEO Jensen Huang recently gave Marvell a market-moving endorsement, calling it a possible “next trillion-dollar company” during Computex. Axios reported that Marvell’s stock surged nearly 33% in one day after the remark, lifting its market value close to $250 billion.

That kind of move says a lot about market psychology. Investors are no longer only buying the most obvious AI winners. They are hunting for the next layer of beneficiaries, especially companies that supply the infrastructure needed to scale AI across cloud platforms, enterprise systems, and hyperscaler data centres.

Marvell fits that search because its business sits where AI strain is becoming more visible. More models mean more computing. More computing means more data movement. More data movement means more demand for faster, more efficient connectivity.

That is the less glamorous side of the AI boom, but it may become one of the most important.

The Real AI Bottleneck May Be Data Movement

The first stage of the AI rally focused on computing power. The next stage may focus on the plumbing.

AI workloads need to move data between chips, racks, servers, storage systems, and data centres. As models get larger and inference demand grows, slow data movement can become a performance bottleneck. In that environment, networking and optical connectivity stop looking like background infrastructure and start looking like strategic assets.

Marvell describes its AI infrastructure portfolio as spanning custom silicon, interconnects, network switches, and connectivity from rack to multi-campus data centre scale. The company positions itself as a critical AI infrastructure partner for hyperscalers undertaking large data centre buildouts.

That is the key point for traders. Marvell Technology isn’t selling the flashiest AI story. It’s selling systems that help the flashiest AI story run at scale.

The market often prices glamour first. Over time, it tends to find the companies that solve bottlenecks.

Custom Silicon is the Next Plot Twist

Custom silicon is where Marvell’s AI story gets more interesting.

Hyperscalers want more specialised chips as AI workloads expand. Off-the-shelf processors remain important, but large cloud firms are increasingly looking for chips built around their own performance, efficiency, and cost needs. That creates a stronger role for companies that can design custom silicon and connect it into larger data centre systems.

We’ve explored this innovation on the quantum level too. Read ‘AI Demand Is Reaching Quantum Scale‘.

Marvell has forecast that its custom chip business will exceed $10 billion in revenue by fiscal 2029, as cloud companies continue expanding AI data centres. The company has also raised its longer-term revenue outlook, helped by demand for AI infrastructure and specialised chips.

The AI multiverse theme is more than wordplay. NVIDIA may dominate the GPU layer, but hyperscalers want more control over cost, power efficiency, and supply. Custom silicon offers them another route.

For Marvell, that shift could support a longer growth runway. For investors, it also raises the execution bar.

Custom chip wins can be large, but they depend on design cycles, customer concentration, production timing, and hyperscaler spending plans.

The NVIDIA Halo Cuts Both Ways

The value behind NVIDIA’s business decision validates Marvell’s position in the AI infrastructure chain. When the market leader points to another company as important to the next phase of AI, investors pay attention.

Huang’s praise gave Marvell a powerful market boost. But it also raises expectations very quickly, making the stock move fast. A sharp rally can pull future growth into today’s price. That leaves less room for delays, softer guidance, or any sign that AI demand is becoming more selective.

The recent reaction around Broadcom shows how sensitive this part of the market has become. Broadcom shares fell sharply after its AI chip revenue forecast came in below the market’s highest expectations, even though the company still pointed to strong long-term demand.

Marvell Technology now sits inside that same market lens. The AI theme remains strong, but traders are looking for more than broad exposure. They want signs of continued data centre growth, custom silicon momentum, and margin strength as expectations move higher.

The Valuation Villain Still Lurks

Every multiverse needs a villain. For Marvell Technology, that villain is valuation.

The company has a credible AI infrastructure story. Its data centre business is growing. Its custom silicon targets are ambitious. Its networking and interconnect portfolio fits the market’s next bottleneck. Yet the share price reaction has already priced in a large amount of future success.

That means the stock may become more sensitive to guidance than to broad AI optimism. Strong revenue growth can support the rally, but investors will also want clearer proof that Marvell can convert AI demand into margin strength, cash flow, and durable customer wins.

Data centre revenue reached $1.83 billion in Q1 fiscal 2027, up 27% year on year, and accounted for the bulk of total revenue. Non-GAAP gross margin came in at 58.9%, while non-GAAP operating income reached $846.9 million.

Those are strong numbers, but the market is now looking forward. The next test is whether Marvell can keep accelerating while avoiding the kind of expectation reset that can hit high-multiple AI names.

Marvell Forecast and Signals to Watch

Marvell Technology’s near-term outlook remains cautiously constructive. The company has earned a more visible role in the AI infrastructure trade, supported by data centre demand, custom silicon growth, optical networking, and hyperscaler spending.

The base case is that Marvell stays in focus as traders look beyond the obvious GPU winners and into the companies powering the wider AI data centre stack. If revenue growth keeps accelerating and custom silicon demand holds, the stock could continue to command a premium.

The risk is that expectations have already risen quickly. After a sharp rally, Marvell may need more than broad AI optimism to keep climbing. Stronger guidance, cleaner margin expansion, and proof that custom silicon revenue can scale will matter more from here.

Traders should watch four key signals:

• Data Centre Revenue: This remains the clearest read on whether Marvell’s AI infrastructure story is still gaining strength.
• Custom Silicon Updates: New hyperscaler wins, changes to revenue timelines, or revisions to fiscal 2029 targets could move sentiment quickly.
• Networking and Optical Demand: Stronger demand for high-bandwidth interconnects would support the view that Marvell is solving one of AI’s biggest data centre bottlenecks.
• Semiconductor Valuation Pressure: If investors continue to reward AI infrastructure names, Marvell could benefit from sector momentum. If the market starts questioning AI capex or stretched forecasts, even strong companies may face sharper pullbacks.

The AI multiverse is getting bigger. Marvell Technology has earned a stronger role in it, but the market will now expect it to perform like more than a supporting character.

Interested in trading the next layer of the AI infrastructure theme? Open a demo account with VT Markets and explore global share CFDs, including semiconductor and technology stocks.

Start trading now – Click here to create your real VT Markets account