Tantra Analyst

In cellular technology, when Qualcomm talks, the industry pays attention and listens. That is because the company not only talks the talk but also walks the walk by making products that turn the vision into reality. That’s exactly what was on display at Qualcomm’s 6G Leadership Day, on Aug 25, 2026, at its global headquarters in San Diego. The company previewed where 6G is headed, committed to a 6G modem in the second half of 2029, and signaled that its IP licensing approach will likely follow the 5G model.

I got the opportunity to attend the event, listen to the presentations and discussions, speak with Qualcomm’s executives and engineers, and check out interesting live demos. Qualcomm also invited CTIA President and CEO and former FCC chair, Ajit Pai, to deliver the keynote. Mr. Pai is working across the US cellular industry to identify and secure one of the most important ingredients for 6G: spectrum!

Key highlights and main messages:

• 6G will be a confluence of connectivity + compute + sensing

• Will be AI native and Agent-centric: AI will be infused into every part of the system—radio (AI RAN), device, network, apps, services—with distributed AI compute

• Will be device-driven, with heavy emphasis on uplink (2-5x edge data rate) – Different from previous generations, where the system was network-driven and downlink-focused

• Biggest differentiator is “Dynamic QoS”- real-time QoS management for best performance & efficiency – was a vision for previous generations, but is being fully realized in 6G through AI

• Wide-area sensing is a major “NEW” use case. Most other use cases are enhanced 5G. NTN will be integral to the 6G offering

• Need for up to 400 MHz wide contiguous mid-band spectrum: 6–8.4 GHz band is a primary target

• Most fundamental radio technologies, such as modulation, coding, etc., to remain the same as 5G, with performance improved by AI

• Because of this, Qualcomm believes the IP landscape, be it leadership, licensing framework, competitive positioning, etc., will remain the same as 5G. This means AI enhancements will be more proprietary and for differentiation. No word on licensing fee structure

• No clarity on licensing “Sensing’ IP as it is implemented on the network (not device)

• Timelines: First 3GPP spec with 6G will be Release 21, expected to be finalized by early CY2029; First 6G modem by the second half of CY2029

• Many interesting live demos:

  • Giga MIMO – Transmitter and Receiver system with 1024 antenna elements using 8 GHz

  • 3D Sensing (aerial and terrestrial) using terrestrial network (base stations)

  • Many use cases of Dynamic QoS working with Distributed compute – low-latency configuration, in near real-time, when and where needed

Tantra’s take:

Given Qualcomm’s unique position in the cellular ecosystem, the event was an important step toward aligning the industry around the company’s technology vision and the likely path to commercialization. In that sense, it helped start the broader industry dialogue and served as a practical stepping stone toward 6G.

6G is surely missing the hype of 5G, and that’s a good thing. Most aspects of 6G are realistic and implementation-oriented, addressing real problems. In line with history, it seems that while 5G was a vision-setting generation, 6G might turn out to be the implementing generation of that vision, bringing up to a 1.5x increase in spectral efficiency and a magnitude increase in capacity. That’s where things like Giga MIMO play a crucial role.

The influence of AI on 6G is unmissable. And that makes total sense, not because of AI hype. But because the wireless link is one of the most complex technical and mathematical challenges to solve. AI is the perfect tool for solving and enhancing performance, be it in channel estimation, coding/modulation selection, or something else.

Multimodal and Agentic AI are already increasing uplink traffic. If AI usage increases as estimated, uplink capacity and performance will become critical. Dynamic QoS, which has been the vision since 4G, is now being implemented in the best possible way using AI. This will improve many use cases, such as connected glasses, and enable connectivity for Physical AI.  

Sensing seems like an interesting new application/use case that will allow operators to generate new revenue from their cellular networks with minimal incremental overhead. Its potential, however, depends heavily on broad adoption of drones and clear commercialization models for network-based sensing.

To sum it up, 6G will be 5G enhanced by AI, with some new things like sensing, requiring new spectrum and a new stand-alone network.

Closing thoughts:

6G’s biggest challenge will be finding capital to buy spectrum and deploy infrastructure.  During 5G, there was enthusiasm and expectation of new revenue, which attracted investors. Although 5G has improved broadband across the board and enabled new services such as FWA, it has not created the major new revenue streams envisioned. That will dampen investors’ enthusiasm.

Ultimately, the solution might be to bring deep-pocketed hyperscalers and cloud players into the equation. That means developing 6G-based services and solutions for those audiences, so they see the investments as worthwhile. The ongoing rumor that Starlink is interested in buying US cellular operators is also in a similar vein. I have some concrete ideas on this aspect, which I will present separately later.