Qualcomm’s AI gambit: Cristiano Amon forecasts $15B data-center AI sales by 2029

Qualcomm is no longer trying to be “a smartphone chip company with ambitions.” At its investor day in Manhattan, CEO Cristiano Amon pitched something far more aggressive: a five-year plan to take on Nvidia’s dominance in data center AI chips. And he backed the pitch with a hard number that the market could either believe or punish. Qualcomm forecast annual sales of more than $15 billion from AI components in data centers by fiscal 2029.

That $15 billion target matters because data center AI is not just another product line. It is the battlefield where the AI stack gets bought and reused at massive scale, and where Nvidia has been the clear leader. Fortune reports that shares jumped as much as 15% on Wednesday after the forecast, before those gains mostly evaporated by Friday as Qualcomm got dragged down by a broader Nasdaq sell-off. The stock reaction still tells you the same thing the forecast did: investors are watching whether Qualcomm can earn a credible foothold beyond mobile.

To understand why this is a true “gamble,” you have to look at Qualcomm’s starting point and incentives. Qualcomm has long been best known for smartphone chips. Amon, who took over as CEO in 2021, has spent the last five years trying to change the center of gravity of the company. Fortune notes that since 2021, Qualcomm has repositioned itself into newer areas such as automotive tech, including driver-assistance and connected-vehicle systems, and into chips for smart home devices and wearables. The investor day frames the data center move as the next phase of that reinvention, not a sudden pivot.

The forecast is also paired with another line item that signals Qualcomm is trying to reduce reliance on its handset business. Qualcomm projected annual revenue of $40 billion for its businesses excluding handsets by 2029. Fortune calls out that this would be double a long-range forecast from two years ago, which is a big clue about internal confidence and how much the reinvention story has already changed the trajectory. In plain English: Qualcomm is telling investors that “phones” will stay important, but it wants a much larger non-handset engine.

But beating Nvidia in data center AI is not just about shipping chips that perform. Fortune highlights the practical reason that Nvidia is hard to dislodge: software and developer ecosystems. This week, Qualcomm announced the $3.9 billion acquisition of AI software company Modular. The deal gives Qualcomm a software platform positioned to compete with Nvidia’s CUDA. CUDA is where many AI developers anchor their workflows, and Fortune notes that Modular’s platform allows developers to develop AI programs and services and take full advantage of Nvidia’s GPUs. That detail is important because it means Qualcomm is not only trying to offer alternative hardware. It is trying to weaken Nvidia dominance by meeting developers where they already run.

Qualcomm’s investor day also leaned on credibility signals from customers and big tech buyers. Fortune reports that the presentations included video endorsements from the CEOs of Meta and Microsoft, Mark Zuckerberg and Satya Nadella, respectively, plus senior executives from Amazon and Google. This doesn’t prove Qualcomm will win market share from Nvidia. It does, however, suggest those companies are willing to publicly engage with Qualcomm’s platform and chip roadmap enough to lend their names at an investor event.

Amon leaned into another theme that’s especially relevant for boards and enterprise buyers: power efficiency. He recognized public concerns about AI and data centers, given the massive power needs, and pointed to Qualcomm’s opportunity to stand out with power-efficient CPUs. Fortune captures this as a strategic opening: he sees pushback as something that can drive the industry to look for alternatives. The underlying logic is that if power and cost become procurement priorities, the “fastest” chip is not always the only winner. Systems efficiency, total cost, and energy constraints can rewrite the ranking.

The plan also comes with scope risk. Fortune describes the investor day presentations as dizzying, with multiple initiatives underway at once. Alongside data center efforts, Qualcomm is pushing further into automotive, and taking on Intel and AMD in the PC chips market. That breadth could be read as dilution, except Amon argues it is consistent with the way Qualcomm reinvents itself. He dismissed the notion that it is too much to chase at once, telling Fortune that Qualcomm has a strong engineering culture and is “not afraid of taking on completely new challenges.” He also recalled answering questions in 2021 about chasing “so many rabbits at once” across automotive, PC chips, and industrial applications. His response then, “Because we can,” is now framed as “the exact same thing.”

For executives, the second-order question is not whether Qualcomm can design AI accelerators and CPUs. It can. The question is whether Qualcomm can translate engineering into adoption in a market where Nvidia is entrenched both in hardware and in software habits. If Qualcomm hits its fiscal 2029 projections, it would validate the reinvention narrative and likely change how boards think about platform investments in AI. If it misses, it could still move the company forward, but it would tighten the scrutiny on how quickly Qualcomm can convert customer interest into durable, repeatable revenue.

Either way, Nvidia has a competitor it can’t ignore anymore. For anyone running chip strategy, data center procurement, or platform software partnerships, Qualcomm’s investor day is a reminder that the AI cycle is not just about models. It is about the stack, the ecosystem, and who controls the “default” path from developer to production.