Qualcomm’s latest investor day in New York turned into a strategic relaunch of its data center ambitions, as the mobile chip giant moved to stake a claim in the AI infrastructure race long dominated by Nvidia and other incumbents.

Laying the groundwork

In the run-up to the event, coverage of Qualcomm Inc had framed the company as a key but still largely mobile-focused player, with investors watching for signs of diversification into higher‑growth AI and data center markets.¹

On June 24, Qualcomm used its investor day to unveil the Dragonfly product family and, crucially, to reveal Meta as the first named customer for its Dragonfly C1000 data center CPU.² The chip is designed as a general‑purpose server processor to sit alongside Qualcomm’s AI accelerators in data centers, with Meta committing to use the C1000 and its successors across its facilities once the part becomes available in 2028.³

Building a broader AI stack

At the same event, Qualcomm filled in details that had been teased at Computex, expanding Dragonfly into three lines: data center CPUs, AI inference accelerators, and custom silicon for hyperscalers.³ It introduced the AI300 accelerator, adding to an existing lineup that already includes the AI200 and AI250, all focused on running trained AI models at scale rather than training them from scratch.³

The company also confirmed an all‑stock acquisition of AI software startup Modular for about $3.9 billion, aiming to create a software ecosystem around tools like the Mojo language and MAX inference engine that can run models across chips from Nvidia, AMD, Intel, and Qualcomm.³

Market reaction and ecosystem response

Financially, the Meta win marked Qualcomm’s first major Big Tech customer for data center chips and helped lift its shares as much as 15% after the announcement, supported by a higher revenue outlook.²

On social media, AI community figures highlighted a “one more thing” reveal at the event: a partnership between Qualcomm and Hugging Face to bring open models to Qualcomm platforms and beyond, enabling users to “take any model, big or small, and deploy on Qualcomm platforms.”clem 🤗 (ClementDelangue)⁴ Hugging Face CEO Clément Delangue amplified the collaboration with a series of posts celebrating Qualcomm’s role and thanking CEO Cristiano Amon for the surprise announcement.clem 🤗 (ClementDelangue)⁵clem 🤗 (ClementDelangue)⁶

Despite the momentum, analysts note that Qualcomm’s mobile power‑efficiency strengths remain unproven at hyperscale in data centers, leaving open the question of how its new hardware and software stack will perform against entrenched rivals.³