Editor’s note: Qualcomm provided travel, lodging and other accommodations associated with 6G Leadership Day.
Qualcomm’s diversification, rooted in its pedigree of low-power, connected computing, is expanding from a wide range of devices into data centers, software and physical AI. The company sees 6G as the horizontal platform connecting those businesses into a distributed but unified system built for the AI era.
Qualcomm has spent much of the past five years explaining how technology developed for smartphones could be extended into PCs, automotive, industrial systems and other edge devices. Its next phase of diversification is considerably broader. The company is moving into data center infrastructure, pushing further up the software stack, expanding its physical AI portfolio and developing new classes of agentic devices.
At Qualcomm’s June 2026 Investor Day, CEO Cristiano Amon sketched the throughline, and it’s worth quoting at length: “We have been designing 6G for this AI era… There’s a lot of improvement on the connectivity side…The other part of the story is the computing part. And the computing is going to be required because the 6G infrastructure is no longer dedicated equipment for communications…It’s now transporting bits, and it’s going to also generate tokens.”
As compute spreads across devices, networks, edge infrastructure and centralized data centers, Qualcomm is characterizing 6G as the horizontal connectivity layer binding those increasingly diverse computing assets together.
Qualcomm executives elaborated on these ideas this week at the 6G Leadership Day event hosted at the company’s San Diego headquarters. Nitin Dhiman, vice president of product management, described AI, satellites, data centers, quantum technologies and other technology vectors developing simultaneously while, “in parallel we have 6G developing as a standard…that is meant to support [them] as a platform and fabric.”
The scope of Qualcomm’s diversification provides important context for that vision. Amon told investors the company is now present across “the entire compute continuum,” spanning power envelopes from “sub-2 milliwatts to about 200 kilowatts.” Its portfolio now reaches from extremely power-constrained wearable and embedded devices through smartphones, PCs, vehicles and industrial systems to the Dragonfly data center infrastructure.
The strategy is twofold: to compete in each of those markets independently, and to deliver compounding value by reusing common building blocks — CPU technology, AI acceleration, connectivity, sensing, RF expertise and software — across the continuum.
This is particularly visible in Qualcomm’s emerging infrastructure strategy. Its data center push brings server-class CPUs, AI accelerators, high-speed networking and heterogeneous computing into the portfolio. During the San Diego event, Qualcomm executives described adapting those same technologies to the telco environment rather than building an entirely separate infrastructure stack.
The objective, according to the company’s presentation, is an architecture extending from intelligent devices through 6G RAN and RAN automation, into the telco edge cloud and ultimately the central cloud. Running horizontally across that architecture are three capabilities Qualcomm repeatedly identified as foundational to 6G: connectivity, compute and sensing.
Qualcomm’s Durga Malladi, senior vice president and general manager of technology planning and edge solutions, framed the technological convergence behind this architecture. AI and wireless, he said, have been evolving alongside one another for generations, but “the technologies are converging in their own way — they feed off each other.” In the run-up to AI-native 6G, he said, it is important to explain “how these things come together.”
The relationship runs in both directions. AI is already making communications systems more adaptive and efficient, and that role will expand, while advanced wireless connectivity enables AI systems to reach and understand the physical world.
Historically, cellular infrastructure has primarily transported information between endpoints and centralized computing resources. In Qualcomm’s 6G vision, the network becomes one layer of a distributed computing system capable of deciding where AI workloads should be run and providing context to those workloads.
Malladi described the emerging architecture as “a continuum of compute” extending from the network edge to the data center. That requires increasingly high-performance CPUs for traditional RAN functions, accelerators for AI workloads and significantly greater energy efficiency as radio configurations scale. “It’s not just an efficient bit pipe,” he said.
All of this does not necessarily mean every service provider becomes an AI cloud provider. Qualcomm executives acknowledged different operators will pursue different models. Some could expose distributed compute capacity or host sovereign and enterprise AI workloads; others may focus primarily on providing the connectivity underpinning those services. In either case, operators will participate in a distribution of AI inference capabilities.
The idea is that agentic devices will combine local processing with edge and cloud resources according to latency, model size, battery life, privacy and available connectivity. Cars and robots introduce another layer of real-time physical AI. Meanwhile, network infrastructure itself can become a source of perception through integrated sensing and communications.
This helps explain why Qualcomm’s disparate investments increasingly fit together. Snapdragon addresses personal and mobile compute, Dragonwing extends into industrial and embedded systems, and Dragonfly pushes into the data center. Modem-RF and RAN technologies provide the connectivity between them, while Qualcomm’s expanding AI software strategy is intended to make workloads more portable across heterogeneous compute environments.
The big idea is that connectivity becomes more valuable as intelligence distributes. Amon summarized the broader thesis at Investor Day. “The token economics will make distributed inference inevitable.” As intelligence spreads across the compute continuum, he said, “6G will be foundational infrastructure for the age of AI.”
Qualcomm built its original business by connecting mobile devices, its next wager is that AI turns nearly everything into a computing endpoint and makes those endpoints increasingly dependent on one another. The more distributed computing becomes, the more important the fabric connecting that compute becomes. And Qualcomm is positioning 6G as that fabric.
