Qualcomm has signed a deal with Meta as the first customer for its Dragonfly C1000, which is expected to launch in 2028, and confirmed its acquisition of Modular for $3.9 billion.
See also: Qualcomm launches Snapdragon Reality Elite and a toolkit for AI glasses
Qualcomm announced the deal at its investor day in New York, along with a new AI300 accelerator chip and a confirmed acquisition of AI software startup Modular for about $3.9 billion in stock. The Dragonfly C1000 is a general-purpose server processor designed to work inside data centers alongside Qualcomm’s AI accelerator chips. Meta has committed to using the C1000 and its successors in its facilities. The chip won’t be available until 2028, suggesting the partnership is a long-term commitment rather than an immediate implementation.
The Dragonfly brand, first revealed at Computex in early June, includes three product categories: data center CPUs, AI inference accelerators, and custom silicon built with hyperscalers in mind. The recent event provided additional product details that weren't included in the Computex announcement.
On the accelerator side, Qualcomm introduced the AI300 chip in a series that already includes the AI200 and AI250. The AI200, built with Hexagon neural processing unit with direct liquid cooling and up to 768GB of LPDDR memory, is on track for initial customer shipments later this year, while the AI250 is expected to follow in 2027.
These accelerators are designed for inference, which involves running trained AI models at scale instead of training them from scratch. Qualcomm claims that its decades of mobile chip design give it an advantage in energy efficiency, an important factor as data centers increase the strain on electrical grids worldwide. However, whether this mobile experience translates to data center performance remains to be seen at scale.
See also: Qualcomm Snapdragon C changes budget Windows PCs
The acquisition of Modular, confirmed at around $4 billion in an all-stock transaction, will see Qualcomm issue around 19 million shares to Modular owners. The deal is expected to close in the second half of the year.
Modular develops the Mojo programming language and MAX inference engine, software that allows AI models to run on chips from Nvidia, AMD, Intel, and Qualcomm without requiring developers to rewrite the code for each processor. This directly challenges Nvidia’s CUDA platform, which has locked AI developers to Nvidia hardware for two decades. Overcoming this lock is a central challenge for companies competing with Nvidia in AI infrastructure.
Qualcomm's strategy is clear: while it can design competitive chips, it needs a software ecosystem that encourages developers to use them. Modular's cross-platform tools could provide Qualcomm with the developer loyalty it currently lacks.
CEO Cristiano Amon presented the deal as part of a broader industry move toward open, multi-vendor architectures, positioning Qualcomm as an alternative to Nvidia by offering flexibility where Nvidia's CUDA requires commitment.
See also: Google: Vulnerability in Qualcomm Android Component used by hackers
Despite Qualcomm's ambitious goals, its track record in data centers is limited. The company primarily generates revenue from processors and modems for smartphones, and its previous attempt to enter the server market with its Centriq processor in 2017 ended in failure. The current initiative has more institutional support, a named hyperscaler customer in Meta, and a clearer market opportunity.
