5 AI Stocks That Could Be the Next Nvidia - Before Wall Street Figures It Out

NVIDIA (NASDAQ:NVDA) has not solely become a trillion-dollar company because it made the best chips, but it became one because it was seemingly the only company that could deliver what the AI industry needed, exactly when it needed it. This kind of structural dependence is what turns a stock into a generational winner. As a result, the question every investor should be asking right now is straightforward: Who could be the next NVIDIA? 

Even as a GPU shortage helped define 2023 and 2024, as AI data centers scale in 2026 and beyond, the constraints are shifting, and the chips are shipping. Instead, what’s in short supply now is the physical infrastructure to make those chips useful. The cooling systems, high-speed networking, memory, and the optical interconnects, just to name a few. Hyperscalers will spend hundreds of dollars in capex this year, and a growing share of that money is flowing to companies most investors haven’t been paying enough attention to. 

The five stocks below each sit at a chokepoint in the AI supply chain that arguably didn’t exist at this scale three years ago. Best of all, these are not speculative plays, as they have accelerating revenue, record backlogs, and direct relationships with NVIDIA, but the market hasn’t fully priced in what these positions are worth. 

Why The Physical Layer Is Where the Next NVIDIA Lives

NVIDIA’s GPUs are useless without the infrastructure to power them, cool them, feed them data, and move signals between them at the speed AI demands. As clusters scale past 100,000 GPUs, every one of those functions becomes a potential bottleneck, and the companies that can help solve this will become increasingly indispensable. That’s the setup. 

Vertiv Holdings

If it’s NVIDIA shipping the chips, it’s Vertiv (NYSE:VRT) that is keeping them alive as the company that is manufacturing the power distribution and thermal management systems that every AI data center requires before a single GPU goes online. As these racks of GPUs go online and push past 100 kilowatts, their liquid cooling systems have shifted from optional to mandatory. 

The company’s Q4 2025 revenue indicated it hit $2.88 billion with organic orders surging 252%, and full-year revenue reaching $10.23 billion, all while its backlog stands at $15 billion, up 109% year-over-year, with a Q4 book-to-bill ratio of 2.9x. 

The NVIDIA parallel is even more direct in the sense that just as the GPU supply couldn’t keep pace with demand in 2023 and 2024, cooling and power infrastructure are now the binding constraint on how fast data centers can deploy AI capacity. Vertiv is guiding $13.25 billion to $13.75 billion in 2026 revenue, which would crush Wall Street’s prior $12.4 billion estimate. 

Marvell Technology

NVIDIA owns computing, but its Marvell Technology (NASDAQ:MRVL) owns connectivity as its AI clusters scale to hundreds of thousands of GPUs, the speed at which data moves between chips becomes just as critical as the chips themselves. Marvell’s customer AI ASICs, optical DSPs, and 1.6T interconnect solutions are the networking fabric that have helped hold some of these clusters together. 

As far as the numbers go, fiscal 2026 revenue hit a record $8.2 billion, up 42% year-over-year, with data center products accounting for 74% of total sales and non-GAAP EPS growing 81%. 

The next NVIDIA case with Marvell examines the company’s custom silicon business, which grew from near zero to $1.5 billion in annual revenue in a single year. The company has 18 design wins with hyperscalers, including Microsoft and Amazon, and it recently acquired Celestial AI for $3.25 billion to bring photonic interconnect technology in-house. In Marvell’s case, every major cloud company is likely to need its chips to connect to its AI infrastructure, it’s no longer just the supplier, but also a platform. 

Astera Labs

Every rack of NVIDIA GPUs needs Astera Labs (NASDAQ:ALAB) retimers to function, and the company’s PCI/CXL smart DSPs and cable modules ensure data flows between components with any kind of signal degradation. It’s this performance that has helped make Astera the connective tissue inside the GPU rack itself. 

Given its critical nature, it’s no surprise that Astera’s Q4 revenue hit $270.6 million, up 92% year-over-year, with 75.7% gross margins heading toward 74% in Q1 due to a $6.5 billion Amazon warrant agreement and higher-margin product mix shift tied to Taurus model expansion. 

The stock itself is down nearly 60% from its 2025 peak as the Amazon partnership introduced an estimated 200 basis point quarterly margin drag starting in Q2, but the structural position remains well intact despite any near-term profitability concerns. Astera is also expanding shipments to additional hyperscalers, and its Scorpio fabric targets a connectivity market estimated to reach $25 billion over the next five years. 

A $20 billion market cap and with revenue that is nearly doubling annually, Astera has the profile of an essential AI infrastructure name. 

Micron Technology

AI servers require approximately three times the memory of standard servers, and the specialized high-bandwidth memory that feeds NVIDIA’s GPUs is in a structural shortage that Micron (NASDAQ:MU) is uniquely positioned to take advantage of. Micron is the only American HBM manufacturer, and its Q1 fiscal 2026 revenue came in at $13.6 billion, up 57% year-over-year, with non-GAAP EPS of $4.78, beating estimates by over 20%. The kicker is that the company’s entire 2026 HBM supply is already sold out, including its next-generation HBM4. 

The valuation is where the NVIDIA comparison becomes compelling as the company is guiding Q2 revenue of $18.7 billion with an EPS of $8.42, a number that represents 440% earnings growth, yet it trades at roughly 9x forward earnings. 

Comparable AI infrastructure names trade at 25 to 30 times, all while the HBM market is projected to grow from $35 billion in 2025 to $100 billion by 2028, and Micron can currently meet only 50% to two-thirds of the demand from its key customers. NVIDIA is going to need Micron’s HBM3E for its Blackwell platform and will need HBM4 for everything that comes next. 

Lumentum Holdings

Without Lumentum’s (NASDAQ:LITE) lasers, NVIDIA’s scale-out AI architectures will not work, and the company makes the optical components and co-packaged optics that carry data across the massive clusters NVIDIA’s platforms require.

On March 2, 2026, NVIDIA validated this dependency directly by investing $2 billion in Lumentum, along with multibillion-dollar purchase commitments for laser components. This isn’t a partnership announcement, instead, this is NVIDIA locking down a supply chain it can’t build without. 

Looking at the numbers, Lumentum’s fiscal Q2 2026 revenue reached $665.5 million, up 65.5% year-over-year, and the company swung from a net loss of $78.2 million to a net income of $1.1 million. Better yet, Q3 guidance calls for $780 million to $830 million in revenue, while the optical backlog exceeds $400 million, and analysts project fiscal 2026 revenue of approximately $2.6 billion. The stock has surged over 900% in the past year, so the valuation is the clear risk, but with NVIDIA committing to a multiyear purchase, it’s telling you whether Lumentum is in a strong position.