Qualcomm Technologies announced on Monday (October 27) its entry into the artificial intelligence (AI) server market, challenging industry leader Nvidia. The company plans to launch the AI200 and AI250 inference chips and integrated rack-scale systems within the next two years.
Qualcomm stated that the AI200 chip will go on sale next year, while the AI250 chip is scheduled for release in 2027.
Both chips can be paired with a direct-to-chip rack cooling system which, compared to cooling methods used in other large-scale data centers, can significantly save on water usage.
Furthermore, Qualcomm plans to introduce more AI products to meet the rapidly growing demand in the U.S. AI data center hardware market.
The U.S. boasts over 5,400 data centers, leading the global count, and with the increasing prevalence of artificial intelligence, the industry is expected to grow at a rate of 12% annually by 2030.
Qualcomm mentioned that its AI chips can support high-load computing tasks such as machine learning frameworks, inference engines, and generative AI frameworks, with low power consumption and overall lower costs compared to competitors.
Durga Malladi, Qualcomm’s Senior Vice President and General Manager of Technology Planning, Edge Solutions, and Data Center Business, stated, “The Qualcomm AI200 and AI250 chips are fully compatible with major AI frameworks and offer one-click deployment of AI models, aiming to make usage simpler, facilitate quick integration, and drive AI innovation.”
Malladi highlighted that Qualcomm’s software tools and related system support make it easier for users to integrate pre-trained AI models into Qualcomm’s AI inference platform for model management, updates, and expanded applications.
Qualcomm is developing new technologies to alter expectations for rack-scale deployment of AI inference computing power, shattering previous constraints. “These solutions allow customers to deploy generative AI at a lower overall cost while maintaining the flexibility and security necessary in today’s data centers,” Malladi added.
McKinsey & Company reports that global enterprises are expected to invest over $7 trillion by 2030 in constructing new data centers. Qualcomm aims to secure a competitive position in the market by developing AI inference chips, as opposed to AI training chips.
Currently, the global market for AI chips and integrated rack systems is primarily dominated by Nvidia, with AMD also holding a portion of the market share.
Qualcomm has traditionally been involved in wireless communication and computing operations, renowned for its high-performance Snapdragon integrated processors used in laptops and tablets. During Trump’s visit to Saudi Arabia, Qualcomm announced a collaboration with Saudi AI company Humain to construct AI data centers nationwide and develop semiconductor infrastructure and manufacturing capabilities by 2030.
The Monday announcement marks another strategic adjustment for Qualcomm.
Following the news, Qualcomm’s stock price surged at the open and closed with a gain of over 11%. Year-to-date, Qualcomm’s stock price has risen by more than 22%.
Due to the ongoing robust global demand for AI computing capabilities, Nvidia’s stock price hit a new high on the same day (October 27). Year-to-date, the company’s stock price has accumulated nearly a 39% increase, reaching a market value of $4.65 trillion.