Qualcomm (QCOM) is dramatically expanding its presence in the AI data center wars with a slew of announcements that include everything from new AI accelerators and central processing units (CPUs) to memory and software.
The company said it has already inked deals with high-profile customers, including Microsoft (MSFT) and Meta (META), which has signed a multigeneration agreement to use Qualcomm’s new Dragonfly C1000 CPU in its next-generation services.
“We’re launching four product lines over the next 24 months that all come together seamlessly to deliver unparalleled performance,” explained Tony Pialis, executive vice president and general manager of Qualcomm’s data center business.
“We are deploying a full turnkey agentic AI infrastructure that we’ve built from the ground up,” he said.
The products are part of Qualcomm’s broader transformation away from its reliance on smartphone sales. The company is also pushing deeper into the automotive space, as well as competing with Intel (INTC) and AMD (AMD) in the PC chips space.
Qualcomm stock is up 24% over the past 12 months and 13% year to date.
The news comes after Qualcomm announced it is acquiring AI software company Modular. The move gives Qualcomm access to a software platform to rival Nvidia’s CUDA, which enables users to develop AI programs and services that take full advantage of that company’s graphics processing units (GPUs).
CUDA serves as a kind of software moat for Nvidia, ensuring that developers who write programs using the service stick with the company for years.
But Qualcomm said its acquisition of Modular will give it the ability to create services that can run customers’ software built on CUDA on its own AI hardware.
“This is how we open up the industry and deliver greater value by building bridges rather than building moats around our solutions,” Pialis said.
In addition to its CPU and Modular acquisition, Qualcomm debuted its Dragonfly AI300 chip and rack server for AI inferencing. The company previously announced its AI200 and AI250 chips in October.
Qualcomm’s new rack server. · Qualcomm
Qualcomm said it expects the AI300 to begin commercial sampling in 2028.
The company also debuted a new memory technology it calls HBC, or high-bandwidth compute. The offering is designed to compete with existing high-bandwidth memory (HBM) and, according to Qualcomm, will provide a lower total cost of ownership and improved energy efficiency compared to HBM.
Tackling Nvidia, however, will be a difficult task for Qualcomm. Nvidia is the market leader in the space, thanks to early investments in AI, and sells some of the most sought-after chips in the industry.
Nvidia’s longtime rival AMD (AMD) is also entering the AI data center server market this year with its Helios rack server.
Then there are the hyperscalers, Amazon (AMZN), Google (GOOG, GOOGL), and Microsoft, which develop and use their own AI chips in their data centers. Meta (META) is also using its own chips. And on Wednesday, OpenAI announced its first custom chip with Broadcom.
While those companies still purchase third-party chips from the likes of Nvidia and AMD, among others, some have begun renting out their processors to customers.
Still, AI chips remain in high demand, with no signs of that slowing anytime soon, giving Qualcomm a shot at stealing away market share for itself.
Sign up for Yahoo Finance’s Week in Tech newsletter. · Yahoo Finance
Email Daniel Howley at dhowley@yahoofinance.com. Follow him on Twitter at @DanielHowley.
Click here for the latest technology news that will impact the stock market
Read the latest financial and business news from Yahoo Finance