Qualcomm unveiled new AI chips designed to compete directly with NVIDIA and AMD in the data center accelerator market.

Long known as the dominant force in smartphone technology, chipmaker Qualcomm has announced an aggressive move into the burgeoning AI data center market. On Monday, the company unveiled a new line of AI inference-optimized chips—the Qualcomm AI200 and AI250—aimed directly at the high-growth segment currently led by rivals Nvidia and AMD. Investors responded enthusiastically, sending Qualcomm shares soaring by as much as 15%, marking the company’s largest single-day gain in over six months.

According to Qualcomm’s official press release, the AI200 and AI250 are part of a “multi-generation data center AI inference roadmap.” Designed to deliver “rack-scale performance and superior memory capacity for fast generative AI inference at industry-leading total cost of ownership (TCO),” the chips target enterprise workloads requiring both speed and efficiency.

The AI200, expected in 2026, supports up to 768 GB of LPDDR memory per card, enabling organizations to run large AI models at reduced costs. Its successor, the AI250, is slated for 2027 and introduces what Qualcomm describes as “an innovative memory architecture” that delivers “a generational leap in effective memory bandwidth and efficiency for AI workloads.”

Durga Malladi, Qualcomm’s senior vice president and general manager for Data Center and Edge Solutions, stated that the new chips “redefine what’s possible for rack-scale AI inference.” He added, “These innovative new AI infrastructure solutions empower customers to deploy generative AI at unprecedented TCO, while maintaining the flexibility and security modern data centers demand.”

The launch positions Qualcomm in direct competition with Nvidia, whose GPUs currently command over 90% of the AI chip market, and AMD, which has been steadily gaining traction as a secondary contender. Under CEO Cristiano Amon, Qualcomm has been systematically diversifying beyond its traditional smartphone business into automotive, PCs, and, most recently, AI infrastructure. This data center push represents a second attempt following the 2017 launch of its Centriq processor line, which ultimately failed to gain market traction.

Backed by significant industry investments, Qualcomm is now doubling down on the expanding AI sector. The company recently agreed to acquire Alphawave, a semiconductor firm specializing in data center technologies, for approximately $2.4 billion. It has also secured its first major customer: Saudi Arabia’s AI startup HUMAIN.

“This agreement marks a major milestone in our journey to build a global hub for AI infrastructure, which helps enable Qualcomm Technologies’ data center expansion, driving high-performance low-power computing and energy-efficient AI processing at scale,” said Tareq Amin, CEO of HUMAIN. HUMAIN plans to begin deploying 200 megawatts of computing power based on Qualcomm’s new chips starting in 2026.

While Qualcomm’s late entry into the AI data center race presents inherent challenges, its strategic emphasis on cost optimization, memory efficiency, and scalability positions it as a credible challenger to incumbent leaders. In a separate development earlier this month, Qualcomm announced plans to acquire Arduino, the Italian open-source hardware and software company renowned for its developer-friendly boards and global creator community.