Marvell surged more than 32% this time. What really exploded wasn't a single stock, but Wall Street suddenly realizing that in the AI era, the most expensive thing isn't necessarily just the GPU, but the high-speed highway that connects GPUs into a superbrain. In today's episode, we'll explain why Jensen Huang's “Marvell could become the next trillion-dollar company” can instantly turn this company from an AI supporting player into an AI darling.We don't chase the loudest hype; we only break down the most valuable logic.
Let's first zoom in on the early June U.S. stock market, the surface looked calm, but underneath it was already as hot as a Texas parking lot in summer. The S&P 500 and the Dow Jones kept setting new highs, the Nasdaq rose slightly, and on the surface the indexes didn't move much, but the semiconductor sector had already started to smoke. The Philadelphia Semiconductor Index jumped nearly 6%, Marvell surged more than 32%, Broadcom also rose, HPE jumped on strong AI server demand, and memory stocks like Micron continued to sit on the crest of the AI infrastructure wave.
What does that tell us?It shows that the market is no longer simply speculating on one company; it is repricing the entire AI infrastructure supply chain.
Whenever people used to mention AI, the first reaction was Nvidia. But here's the problem: AI doesn't become a money-printing machine just because you buy a pile of GPUs and stack them in a server room. GPUs need to communicate with each other, servers need to communicate with each other, racks need to communicate with each other, and data centers need to communicate with each other as well. You can think of it this way: Nvidia sells the engine, while Marvell is selling the highways, traffic lights, switching hubs, and the entire transportation system that keeps the convoy from grinding to a halt.
That is the real reason Marvell was rediscovered by the market this time.
When many people saw Marvell surging, their first reaction was whether this was just another concept stock that took off after Jensen Huang singled it out. To be honest, if all it took was Jensen Huang standing on stage and praising a company once for the stock to rise 30%, then Wall Street would look way too much like a fan event.
But this time is different, because Jensen Huang's endorsement was only the match. The real fuel is Marvell's own earnings, AI interconnect demand, custom chips for major customers, silicon photonics, and the long cycle of data centers moving from standalone compute to clustered compute.
Let's start with the earnings. Marvell, whose Chinese name is Maiweier Technology, ticker MRVL, in its latest first fiscal quarter earnings report, revenue reached $24.18 billion, a company record, up 28% year over year and 9% quarter over quarter.
This isn't the pace of a traditional chip company slowly climbing uphill; it is AI infrastructure demand pushing it upward. More importantly, data center revenue reached about $18.33 billion, accounting for 76% of total revenue. In other words, Marvell is no longer the traditional semiconductor company people used to think of as making storage controllers, networking chips, and communication chips; it is increasingly looking like an AI data center infrastructure company.
This point is very important. Many people see Nvidia and think AI; see Marvell and think it's vaguely familiar, but can't quite say what it does. But Wall Street doesn't see it that way. It is now looking not at whether you have AI in your name, but whether AI is in your revenue. Marvell's data center revenue already makes up three-quarters of the total, which shows that its core story has changed.
Look at the company's guidance. Marvell expects next quarter's median revenue to reach $27 billion, up about 35% year over year. Management also said that revenue growth in each quarter of fiscal 2027 is expected to continue