Qualcomm is chasing a massive $40 billion transformation. Meta is already buying in.

Qualcomm is attempting to transform into a data-center giant.

Qualcomm is telling Wall Street that its nonhandset business alone will approach the size of the entire company today. Investors are buying the promise.

Qualcomm is still priced as a smartphone and wireless-technology company with budding aspirations in other segments. That perception has been fair, since most revenue still comes from handsets and licensing, and new markets have stayed mostly off the income statement.

At its investor day in New York on June 24, Qualcomm $(QCOM)$ set out to change its Wall Street image. The company brought numbers, products and a marquee customer - Meta Platforms (META) - to help complete the makeover.

Qualcomm raised its fiscal 2029 target for nonhandset revenue to $40 billion, up from a prior target of $22 billion, and said handsets would account for only about a third of chip revenue by then. It set a data-center revenue target of more than $15 billion by fiscal 2029, for a business that is practically nonexistent currently. CFO Akash Palkhiwala told investors the data-center unit would generate roughly $5 billion as soon as fiscal 2027, and put the company on a path to more than $18 in adjusted EPS.

Put plainly, Qualcomm is telling Wall Street that its nonhandset business alone will approach the size of the entire company today, which brought in about $44 billion in fiscal 2025. The company's shares jumped sharply after the targets were released, and the stock posted further gains on Thursday.

None of this is a sudden pivot. It is the diversification strategy that CEO Cristiano Amon has been building toward for years, arriving at the moment the company needs it most. The handset business that still funds Qualcomm is squeezed on two sides: The revenue Qualcomm earns from Apple $(AAPL)$ is set to fall toward zero by 2027 as Apple moves to its own modems, and the premium Android market that drives the rest has flatlined as smartphone growth has stalled. The data center is where that long bet either pays off or not.

A market growing this fast leaves room for a new entrant, and Qualcomm's attempt has moved from vision to commitment. There is now a server processor with real specifications, a memory architecture with a brand, a named customer in Meta and a revenue ramp with dates attached. What Qualcomm has laid out is a full stack that spans CPU compute, AI accelerators, a new memory design and the high-speed connectivity that ties a rack together.

Tough competition

Does Qualcomm actually have an edge against Nvidia and other established rivals, or is it just a late entrant talking a good game?

The harder question, and the one investors are clearly divided on, is whether Qualcomm actually has an edge against Nvidia (NVDA) and other established rivals, or whether it's just a late entrant talking a good game.

I do not think Qualcomm overtakes NVIDIA on raw AI performance - but it does not have to. Two pieces of Qualcomm's technology are more interesting than the consensus suggests.

The first piece is memory. Qualcomm's high-bandwidth compute product is positioned against the high-bandwidth memory, or HBM, that nearly every leading accelerator depends on. Rather than placing memory beside the processor across an interposer, Qualcomm stacks it on top of the compute in a 3D design, which it says cuts the energy required to move data.

Memory has become the gating factor for both cost and power in AI inference, and HBM is expensive and supply constrained. Near-memory compute as a broad idea is not unique to Qualcomm, and others are pursuing versions of it. But Qualcomm's design has a substantial patent portfolio behind it, which makes it defensible and harder for rivals to copy. If Qualcomm gets this right, it is a real technology differentiator. Rival memory makers, from SK Hynix (KR:000660) to Micron Technology $(MU)$ to Samsung Electronics (KR:005930), will be watching closely - even as they supply the low-power memory and stacking that make it possible.

The second piece is the CPU. Qualcomm introduced Dragonfly C1000, a data-center processor that the company says delivers more than twice the performance-per-watt of competing server chips. More important than the specifications, Meta signed a multi-generation agreement to use the C1000 across its server fleet when it enters production in 2028.

A marquee hyperscaler choosing a brand-new server CPU is no small thing. It is a direct shot at data-center processor giants Intel $(INTC)$, AMD $(AMD)$, Arm Holdings (ARM )and Nvidia, and lands at the exact moment the largest buyers want custom, power-efficient CPUs they can tune to their own workloads. CPUs have been one of the hottest corners of the data center, and Qualcomm just planted a flag with one of the most demanding customers in the world.

None of this closes the one gap that has always defined this competition, which is software. The real Nvidia moat was never just the silicon; the software locks developers in. Qualcomm is attacking that the same way it assembled the rest of the stack, by acquisition. It recently announced it will buy Modular, an AI infrastructure software company, in a reported deal of close to $4 billion.

Why the targets are reachable

Qualcomm does not need to win the market. It just needs a modest slice of an enormous one. Research from The Futurum Group puts the AI platforms market near $110 billion in 2025 and on a base case path to roughly $497 billion by 2030, with spending shifting from training toward inference, which is exactly where Qualcomm is aiming. Against a market approaching $500 billion, a data center business of $15 billion is a low-single-digit share. That is a far easier bar than beating Nvidia, and why hitting the targets looks plausible even if Qualcomm never takes the performance crown.

That said, almost everything Qualcomm showed at its investor day ships in 2027 and 2028, which means the revenue is a forecast, not a result. The HBC vs HBM comparison rests on vendor claims that deserve independent testing. The next two years are all about execution. Watch for more top-tier customers beyond Meta, as well as hard data behind the high-bandwidth compute claims and whether the fiscal 2027 ramp actually arrives on schedule.

Qualcomm can compete - not as the company that dethrones NVIDIA, but as a credible, differentiated option in a fast-growing market. Customer commitments have started, and the financial targets are now on the record. The hard part - turning vision into shipped revenue - starts now.

Ryan Shrout is president of Signal65 and founder at Shrout Research. Follow him on X @ryanshrout. He has provided consulting services for AMD, Qualcomm, Intel, Arm Holdings, Micron Technology, Nvidia and others. Shrout holds shares of Intel.

-Ryan Shrout

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

June 25, 2026 16:39 ET (20:39 GMT)

Copyright (c) 2026 Dow Jones & Company, Inc.