Qualcomm Inc. plunged into the artificial intelligence (AI) accelerator chip market Monday with two new processors scheduled to launch over the next two years.
The chipmaker, best known for powering mobile devices, is set to challenge NVIDIA Corp.’s commanding presence in AI data center infrastructure, a market experiencing unprecedented growth as companies race to deploy generative AI applications.
The new processors announced, AI200 and AI250, will become commercially available in 2026 and 2027, respectively. Both chips are engineered specifically for AI inference, the process of running trained AI models to generate predictions and responses with enhanced memory capacity built to handle demanding workloads.
Qualcomm emphasized that its chips will support widely-used AI frameworks and development tools, ensuring compatibility with current enterprise software ecosystems. The company sees cost savings as a key differentiator, appealing to organizations seeking alternatives to premium-priced solutions that now dominate the market.
More important, Qualcomm said it isn’t selling individual chips. Following industry trends set by NVIDIA and AMD Inc., the company introduced complete rack-based systems built around the new processors. The shift from component supplier to full-system provider reflects the evolving needs of AI infrastructure buyers, who increasingly prefer integrated solutions that simplify deployment and optimize performance across interconnected processors.
Qualcomm plans flexible sales approaches, offering complete rack systems or individual components for clients like hyperscalers preferring custom designs.
The announcement represents a crucial diversification strategy for Qualcomm as smartphone demand remains stagnant. While the company maintains leadership in mobile chip technology, growth in that sector has plateaued, forcing executives to identify new revenue streams.
AI data center infrastructure represents one of technology’s fastest-growing segments, with demand accelerating as enterprises adopt large language models and other AI capabilities.
McKinsey estimates nearly $6.7 trillion in data center capital expenditures through 2030, predominantly for AI chip systems. The market has rewarded NVIDIA lavishly — its GPUs trained OpenAI’s ChatGPT models and propelled NVIDIA to a $4.5 trillion market value.
To that end, the landscape is shifting. OpenAI recently announced plans to purchase chips from AMD, the second-largest GPU manufacturer, and potentially acquire an equity stake. Tech giants including Alphabet Inc.’s Google, Amazon.com Inc., and Microsoft Corp. are developing proprietary AI accelerators for their cloud platforms, signaling growing resistance to NVIDIA’s near-monopoly.
Qualcomm’s entry into the AI infrastructure market represents far more than a product launch. It is no less than a strategic inflection point that could fundamentally reshape how investors value the company over the next several years, according to Daniel Newman, CEO of The Futurum Group.
