The highways and bridges that connect all the chips together inside data centers
In the evolving architecture of artificial intelligence, the spotlight rarely falls on the connective tissue — the unseen infrastructure that allows vast networks of processors to think together. Yet when Nvidia's Jensen Huang named Marvell Technology as a future trillion-dollar company, markets responded as if a hidden truth had been spoken aloud, adding $47 billion in value within hours. It is a reminder that in every great technological era, the builders of roads matter as much as the builders of engines.
- A single public endorsement from the world's most powerful chipmaker sent Marvell's stock surging 25% in one trading session, erasing months of quiet drift in a matter of hours.
- Marvell's role as the maker of data center 'highways and bridges' — the connectivity chips linking thousands of AI processors — suddenly felt less obscure and far more indispensable.
- Nvidia's $2 billion investment in Marvell reveals this is not mere flattery but a strategic bet: the two companies are weaving their ecosystems together as AI infrastructure scales toward unprecedented complexity.
- Marvell's market cap now sits near $192 billion — up 158% year-to-date — but the trillion-dollar threshold Huang invoked remains roughly $800 billion away, keeping ambition and reality in productive tension.
On a recent Tuesday, Nvidia CEO Jensen Huang made a prediction that most of the financial world was not expecting: Marvell Technology, a chip company operating largely outside public awareness, could one day reach a trillion-dollar valuation. The market did not wait for proof. Within hours, Marvell's stock climbed 25 percent, adding roughly $47 billion in market value in a single session. Nvidia's own shares rose modestly — a quiet signal of alignment from the industry's dominant force.
Marvell's story in 2026 had already been one of quiet acceleration. After opening the year between $80 and $90 per share and drifting for months, the stock began climbing sharply in April. By the time Huang spoke, it had surpassed $200 — a gain of more than 158 percent year-to-date — lifting the company's market capitalization to around $192 billion.
What Huang was really endorsing was not a product but a function. Marvell does not build the processors that perform AI calculations; it builds the networking and connectivity chips that allow thousands of those processors to communicate inside data centers. Huang called these components the 'highways and bridges' of modern computing infrastructure — unglamorous, but essential. When AI workloads are distributed across entire data centers, the speed of the connections between processors becomes the limiting factor. That is Marvell's domain.
The endorsement carries strategic weight. Nvidia has already committed $2 billion to Marvell, part of a broader effort to integrate Marvell's custom AI chips with Nvidia's own systems. Marvell's leadership, for its part, believes its custom chip business could generate $10 billion in annual revenue by 2029. Nvidia is also placing longer-term bets on photonics — technology that uses light rather than electrical signals to move data — signaling that the race to build AI infrastructure is far from settled.
Whether Marvell reaches a trillion-dollar valuation remains an open question, with roughly $800 billion still to close. But the market has decided the possibility deserves to be taken seriously — and sometimes, that decision is itself the beginning of the story.
Jensen Huang, the chief executive of Nvidia, made a striking prediction on a recent Tuesday: Marvell Technology, a chip company most people have never heard of, could one day become a trillion-dollar enterprise. The market listened. Within hours of his remarks, Marvell's stock price climbed 25 percent, erasing and then adding roughly $47 billion to the company's valuation in a single trading session. Nvidia's own shares ticked upward by 1.8 percent in the same move—a smaller gesture, but a gesture nonetheless from the world's most valuable chipmaker.
Marvell's ascent this year has been steep. The stock opened 2026 somewhere between $80 and $90 per share. For months it drifted. Then, beginning in April, it began to accelerate. By the time Huang spoke, it was trading above $200, a gain of more than 158 percent year-to-date. The company's market capitalization now sits around $192 billion—still far from the trillion-dollar threshold Huang invoked, but no longer a company that can be dismissed as a minor player.
The reason for Huang's confidence lies not in what Marvell makes, but in what it enables. While Nvidia designs the processors that actually perform artificial intelligence calculations, Marvell manufactures something less glamorous but equally vital: the networking and connectivity chips that allow thousands of processors to communicate with one another inside data centers. Huang described these components as the "highways and bridges" of the modern computing infrastructure. Without them, he explained, even the most powerful AI systems cannot function at scale. When a computing problem is broken into pieces and distributed across an entire data center, the speed and reliability of the connections between those pieces becomes the limiting factor. That is where Marvell operates.
The strategic logic behind Huang's endorsement becomes clearer when you consider Nvidia's own investments. Earlier this year, Nvidia committed $2 billion to Marvell as part of a broader effort to make it simpler for customers to integrate Marvell's custom AI chips with Nvidia's own networking systems and processors. The two companies are, in effect, building a more tightly integrated ecosystem. Marvell has announced that its custom chip business alone could generate $10 billion in annual revenue by 2029—a figure that suggests the company's leadership believes in the same growth trajectory that Huang publicly endorsed.
Nvidia is hedging its bets elsewhere too. The company is also investing in firms developing photonics technology, which uses light instead of electrical signals to transmit data. These are long-term wagers on the infrastructure that will support artificial intelligence at whatever scale the technology eventually reaches. For now, though, the market has spoken most loudly about Marvell. A single sentence from Huang—a prediction about what Marvell could become—moved billions of dollars in a matter of hours. Whether the company actually reaches a trillion-dollar valuation remains an open question. But the market has decided that the possibility is worth taking seriously.
Citas Notables
Marvell is set to be the next trillion-dollar company as its networking and connectivity chips are essential to data centres where computing tasks are spread across thousands of connected chips that need to share data quickly.— Jensen Huang, Nvidia CEO
When you take a computing problem and distribute it across the entire data center, what's necessary is connectivity. That's the reason why Marvell is so essential.— Jensen Huang, Nvidia CEO
La Conversación del Hearth Otra perspectiva de la historia
Why does a connectivity chip company matter so much to Nvidia's future?
Because Nvidia's chips can't talk to each other without something in between. Marvell is that something. As AI systems get bigger and more distributed, the bottleneck shifts from raw processing power to the speed of communication between processors.
So Huang is essentially saying Marvell's problem is about to become everyone's problem?
Exactly. Right now, most companies are still figuring out how to use AI at all. But once they scale up, they'll hit a wall—not because Nvidia's chips aren't fast enough, but because the data can't move between them quickly enough. That's when Marvell becomes indispensable.
The stock jumped 25 percent on one comment. Doesn't that seem like a lot of faith in a prediction?
It does, but consider what Huang represents. He's not a pundit. He runs the company that's already won the AI chip war. When he says another company is essential, the market treats that as inside information about where the infrastructure is heading.
Is Nvidia's $2 billion investment in Marvell a vote of confidence or a way to control the supply chain?
Both. Nvidia wants Marvell to succeed because Nvidia needs Marvell to succeed. But by investing directly, Nvidia also ensures it has a seat at the table and some influence over how Marvell's technology develops.
What happens if Marvell doesn't hit that $10 billion custom chip revenue target by 2029?
Then the trillion-dollar prediction looks like wishful thinking, and the stock probably corrects downward. But right now, the market is pricing in the optimistic scenario—that AI infrastructure spending will be as large and as sustained as Huang believes it will be.