Investing.com -- Qualcomm is positioning itself as a broad-based artificial intelligence infrastructure play, moving beyond its smartphone roots to compete across data centers, automotive, and wearables, with agentic AI the unifying theme.

In an exclusive interview with Investing.com, Durga Malladi, Qualcomm's Executive Vice President and General Manager for Technology Planning, Edge Solutions and Data Center, said the company's expansion into custom silicon reflects a deliberate strategy built on decades of platform customization.

"We have a long, successful track record of taking a platform and customizing it for the specific performance, power, and connectivity needs of a given market," Malladi said. "We've been designing custom silicon at scale for different customers, and we see great opportunity to apply this discipline to the data center."

Qualcomm announced a custom silicon partnership with a leading hyperscaler during its most recent earnings call, with initial shipments expected later this year.

The company declined to name the partner but confirmed it anticipates a multi-generation engagement spanning CPUs, AI accelerators, and custom ASICs, which is a business model CEO Cristiano Amon has described as "bespoke."

Further details are expected at its Investor Day in New York on June 24.

Malladi noted that hyperscaler demand is extending across Qualcomm's full data center portfolio, including memory architecture designed to address inference-related bandwidth bottlenecks.

"I’m not able to share specific names, but I will say that the demand signal from these companies is clear," he commented. "In this Agentic AI era, there’s a need for inference-optimized energy-efficient solutions, and that’s exactly what we’re focused on, as we deliver leading CPU, AI accelerators and custom silicon, building a broad, diverse customer base in this space."

Beyond the data center, Malladi pointed to the company's automotive segment, which posted a record $1.3 billion in Q2 revenue, up 38% year-over-year, as evidence of its diversification strategy taking hold.

The fifth-generation Snapdragon Digital Chassis, shipping later this fiscal year, is expected to deliver 12 times the neural processing performance of its predecessor.

On agentic AI, Malladi argued the moment favors Qualcomm's architecture precisely because agent orchestration requires efficient, heterogeneous compute distributed across devices and the cloud, which is a design challenge the company has been working on for years.

"Agentic AI requires heterogeneous compute across CPU and AI accelerators, with intelligence flowing seamlessly from device to edge and data centers," he explained. "Our technology is purposely built for distributed, power-efficient, end-to-end systems that can scale across environments. Our industry partners see that the future of AI is distributed, and given the breadth of our portfolio, Qualcomm is uniquely positioned to deliver that architecture at scale."

Malladi believes "AI is no longer a feature; rather, it is the foundation of every digital experience moving forwards," adding that Qualcomm is looking to "re-architect and re-define AI working across the entire computing continuum — from devices to datacenters."