Table of Contents
The futuristic vision of 6G — whether its agentic devices, mobile robots or RF sensing — starts with much more prosaic requirements
The mobile industry’s experience with 5G offers a useful lens for thinking about 6G. The original 5G proposition leaned heavily on Industry 4.0, ultra-reliable low-latency communications and massive machine-type communications. Some of that vision materialized; some hasn’t — yet at least. Meanwhile, one of 5G’s most commercially consequential U.S. use cases turned out to be fixed wireless access, which introduced a new source of competition into the home broadband market.
During Qualcomm’s 6G Leadership Day, Durga Malladi, senior vice president and general manager of technology planning and edge solutions, cited CTIA data showing the cost per gigabyte to consumers falling 97% between 4G and 5G. In a fireside discussion, CTIA President and CEO Ajit Pai called fixed wireless “one of the great applications for [5G],” adding that, “With 6G it’s going to be another step function above that.”
The lesson for 6G is that the commercial foundation of a cellular-backed service depends on improving the basic economics and performance of mobile connectivity. Malladi characterized better coverage, capacity, user experience, network efficiency and total cost of ownership as “tablestakes.” And regardless of what AI ultimately does to mobile applications, he stressed one hard requirement: “I cannot emphasize enough…At the end of the day, we’re not doing this with wires. We need a lot of spectrum.”
One of Qualcomm’s central assumptions is that AI changes not only how much traffic wireless networks carry but also its direction and characteristics. The mobile internet evolved primarily around consumption as smartphones became increasingly capable endpoints for streaming video and other downlink-heavy services. Agentic and physical AI introduce more data generated at the edge as images and video from intelligent glasses, sensor information from robots and vehicles, and machine-to-machine or agent-to-agent interactions are sent to various AI compute nodes. To put that another way, people click, machines swarm.
At Qualcomm’s June Investor Day, CEO Cristiano Amon described the connectivity requirement in vivid terms. Whereas 5G enabled widespread high-definition video streaming to users, he argued that 6G needs to support the reverse direction and “transform all of us into walking cameras in this world,” enabling high-definition uplink for things liek “see what I see” agentic experiences.
Qualcomm Vice President of Engineering Hemanth Sampath quantified the potential impact during 6G Leadership Day. “Even a modest use of these applications…can lead to 50-plus GB per month,” he said of a “see what I see” use case, adding that, “Just a modest adoption of these types of use cases is going to explode the traffic, particularly the uplink.” More advanced 3D calling and interactive AR would increase the requirement further.
Despite years of discussion about terahertz communications, Qualcomm’s practical 6G spectrum emphasis in San Diego centered primarily on familiar wide-area bands. Its framework spans low bands below 2 GHz, existing mid-band spectrum around 2 GHz to 5 GHz, upper-mid bands around 6 GHz to 8.4 GHz, and millimeter wave for localized high-capacity deployments. The critical addition is access to large, contiguous mid- and upper-mid-band channels, with Qualcomm arguing that IMT-2030 usage scenarios will require channel bandwidths in the 100 megahertz to 400 megahertz range.
The significance of a 400 megahertz channel goes beyond just delivering larger peak-throughput numbers. Qualcomm argued that a single wideband 6G carrier could be inherently more efficient than assembling equivalent bandwidth through carrier aggregation. Its Leadership Day comparison showed a 400 megahertz single component carrier against a multi-carrier configuration such as two aggregated 200 megahertz carriers. The single-carrier approach could simplify scheduling and RF hardware, allow faster bandwidth-part adaptation, improve uplink power efficiency and pool resources under one scheduler. Wider contiguous bandwidth also improves the precision available for positioning and RF sensing.
That links spectrum policy directly to network and device economics. Contiguous spectrum can simplify how capacity is created, not merely increase how much spectrum the operator possesses.
Qualcomm is also proposing a collection of physical-layer changes intended to drive more capacity from both existing and new bands. During the recent event, the company showed a target of 50% or greater spectral-efficiency improvement relative to 5G-Advanced from network baseband upgrades, using changes spanning coding and modulation, data-channel design, MIMO and uplink transmission.
Qualcomm Vice President of Product Management Sunil Patil appropriately qualified that engineering roadmap: not every proposed feature will necessarily be accepted by 3GPP, particularly in the first 6G release. But he said the company is pushing the package as part of a broader effort to improve spectral efficiency while correcting weaknesses identified during the 5G cycle.
Power consumption is equally fundamental. Wider channels, more compute and increasingly complex antenna configurations can’t scale energy use in parallel with network performance.
Massive MIMO, for instance, could evolve toward what Qualcomm calls Giga-MIMO, potentially scaling to 1,024 antenna elements. Malladi noted that a current 32-transmit/receive radio section consumes roughly 1 kW and argued that moving toward 128- or 256-transceiver configurations cannot be allowed to proportionally expand that power envelope. “It’s always about energy efficient, high performance compute processors,” he said.
This is where some of Qualcomm’s more futuristic 6G propositions meet the very traditional operator requirement of delivering materially more network capability without creating an unsustainable site-power or TCO equation.
Patil said Qualcomm’s goal is to bring the pieces together for initial 6G commercialization in 2029 on both device and infrastructure sides. Between now and then, standards work will determine which technologies make the first specification, regulators will determine what spectrum operators can actually use, and the infrastructure ecosystem will have to turn increasingly complex radio concepts into commercially deployable systems.
Agentic AI, physical AI, sensing and distributed computing will likely define what makes 6G different. But those capabilities are not likely to determine whether operators can afford to deploy it at scale. The more immediate test is whether 6G can provide more coverage and capacity, fix the uplink imbalance, use spectrum more efficiently and lower the cost of delivering increasingly valuable connectivity? As we saw with 5G and fixed wireless access, predicting breakout applications can be tough. The safer play is to focus on improving network economics.
