Today, 6G work is dominated by research but over the course of the next two years we will see the balance shift from research to actual development.
The industry has aligned on the timing of the first implementable standard for 6G to be complete no earlier than March of 2029 - so we have some ways to go.
The list of enabling technologies that got lots of attention a few years ago has gone through some degree of cultivation. 2025’s list of “most popular” technologies will no doubt undergo change as further research, early development, and some rudimentary trials prove and, in some cases, disprove a technology’s viability.
First, are a few that have high probability of not being “culled” from the 6G enabling technologies list.
7-16GHz Mobile Terrestrial Radio Systems
Wireless technology relies first and foremost on spectrum availability. The growth in data consumption and wireless connectivity has and will continue to put ever-increasing demand on the spectrum.
For a mobile operator, the ideal (in some cases, the only acceptable) scenario is to have exclusive use of spectrum in its geographic areas of operation over which they can transmit radio power levels high enough to maintain a high-capacity and high-reliability network.
The increased capacity demand has led to the exploration of repurposing the radio spectrum between 7 and 24 GHz with special attention to frequencies between 7 and 16 GHz. This spectrum has significant use in radio navigation, radio location, and satellite applications.
This is complicated by heavy and exclusive use in this band by federal agencies around the world (especially Departments of Defence). In addition, these higher frequencies have increased radio propagation loss over that between 3 and 5 GHz. The latter is used in 5G but carries its own technology challenges due to higher loss than the lower frequencies that are so heavily used in 4G (most below 2.5 GHz).
For mobile wireless to work from 7-16 GHz, there is serious consideration for how some of the spectrum can be shared. Sharing mechanisms involve both complex policies and technologies, so both are getting attention. Even if some of this range is set aside for exclusive use in commercial wireless, the added propagation loss drives significant technology work. The most obvious solution to the issue of a lower signal-to-noise ratio at the receiver is to make the cell size smaller.
However, this is not financially feasible for the mobile operators due to site acquisition costs and the challenge of adding a very dense backhaul interconnection to more cells. Therefore, investigations on how to overcome these problems with advanced integrated radio and antenna systems is critical (see below for “next generation MIMO”).
Artificial Intelligence (AI)
The form of AI known as Machine Learning (ML) is very popular given the advent of multiple powerful large-language models (LLM’s) available for public use. But telecommunications engineers are exploring very different types of models.
Whereas LLM’s are trained in human language response given vast amounts of exchanges on the web, the mobile wireless industry is developing AI to optimise network performance, address the complexities of radio beam management, optimise circuit design, facilitate more efficient traffic flows, and reduce overall power consumption.
None of this uses LLM’s but rather ML models trained on technical data from networks, circuits, and even synthesized data from simulation and emulation tools.
The key technical challenges are driven by the need to ensure a reliable model that is consistent in out-performing conventional means - these can be summed up in 1) how to develop, refine, and train the model (this requires lots of data that developers can trust); and 2) how to validate that the model works under the vast majority of circumstances.
Next Generation MIMO
Multiple in/multiple out (MIMO) was developed to take advantage of the fact that radio waves can follow multiple paths between the transmitter and the receiver (e.g., a direct path, one or more reflected paths).
Before MIMO, multiple paths were a problem for radio communications and caused “multi-path interference” (some of us can remember an image “ghost” on our televisions when the only access was through an antenna-based broadcast system).
MIMO in cellular is now in its 4th generation. The latest manifestations were necessary to overcome the increased loss in the 3.5GHz spectrum allocated for 5G. The fundamental approach is to 1) use many antenna elements and complex digital signal processing (DSP) so that the antenna elements work together to improve the effective signal-to-noise ratio at the receiver; and 2) to constantly measure the state of the radio channel between the transmitter and receiver (mobile wireless channels are under a constant state of change) so that the DSP does continuous manipulation as to how the multiple antenna elements are used to overcome the constant change in the channel.
The move to 7-16 GHz while keeping the cell-size the same (e.g. keeping the max distance between transmit and receive the same as with 3.5GHz) means even more technical complexity in the MIMO system: more, and even distributed, antenna elements, and stronger DSP. This is an excellent place to leverage ML given the complexity of what is needed.
Open RAN
Radio Access Network (RAN) is the term used for the network of radio base stations required to interface with the mobile user equipment (e.g. smart-phones).
Before 5G, the RAN was a closed architecture, with each of a few large network equipment manufacturers having their own proprietary approaches. However, the idea of virtualizing the digital parts of the RAN (software entities running on high-performance general-purpose servers) has driven the industry to work to standardise the resulting disaggregation (radio unit, digital unit, centralized unit) and to standardize the interfaces between these architectural components.
This open RAN approach has led to new concepts including intelligent controllers of the RAN functionality (RAN Intelligent Controller or RIC) in which ML is already getting some degree of use. Open RAN (and other open standards) is seen by many as a necessary step for 6G and thus, further work in the space is happening to move the concepts to the next generation
Second, the following are getting plenty of attention this year, but have a higher risk for commercialisation in 6G.
Millimeter-wave technology (from 5G days—24-71 GHz)
Frequency Range 2 (FR2), as 3GPP refers to this band, is already in use in 5G, but the industry has struggled to make the services profitable. The technology remains expensive with no clear “killer app” to drive use and volume (and hence reduce the cost through economy of scale). There is also work necessary in standards and in implementation to improve the reliability of the radio links (especially smart beam management, which is something similar to MIMO in that it relies on accurate real-time channel state information and can also benefit from ML).
However, the demand for capacity and spectrum is profound and the amount that will be freed up in the 7-17GHz range will not be enough. Hence FR2, much of which is allocated but still underutilized, can be a necessary part of this.
Integrated terrestrial & non-terrestrial networks
There is much in the news of late in integrating terrestrial and non-terrestrial wireless networks (NTN) - leveraging satellite and high-altitude platforms (HAPS - balloons, sub-orbital stratospheric aircraft, etc.). This is about better coverage and increased reliability - especially in the case of natural disasters or maritime distress. The technologies are challenging:
1. Transmit-to-receiver distances of hundreds of kilometers (not hundreds of meters),
2. Managing data traffic between multiple disparate networks
3. Managing interference given an added dimension to the transmission direction (almost no cell-phone towers point the signals straight up or straight down and all standardized radio channel models are two-dimensional only)
This is exciting territory, and while the business model for satellite companies may seem obvious (same infrastructure, more users) for the terrestrial mobile operator, it is less clear.
Integrated Sensing and Communications (ISAC)
The opportunity to use communications signals to sense the environment is another area getting significant attention. Traffic management, drone management, crowd management, and myriad other applications are under consideration. The challenge is related to 1) the radio frequency, wavelength, and signal bandwidth and 2) capacity management.
The frequency, wavelength, and bandwidth of the signals have a direct relationship with the physical and time precision of the sensing. Capacity is also critical; setting aside radio resources for sensing-only means that they cannot be used for communications and capacity demand has been discussed above. However, signals that are ideal for communications are not necessarily ideal for sensing.
Also, if sensing and communications can be done with the exact same signal, there is no guarantee that the desired direction of the sensing requirement will be the same as the direction in which the system must transmit the necessary radio signal. So, the technical work means juggling these multiple challenges in addition to addressing the complexities of interference in sensing from multiple base stations and mobile devices. The business models here are not obvious so the ultimate utility of this technology remains to be seen.
Third, these topics are still getting research attention, but commercialisation possibilities are even less clear.
Reflective Intelligent Surfaces (RIS)
Indoor propagation and outdoor-to-indoor propagation are problematic in many radio systems. For example, parking garages, large commercial buildings, shopping malls, and indoor stadiums are addressed by distributed antenna systems and radio repeaters—sometimes even by additional independent base stations. The theory is that less-expensive approaches using large wall-mounted “surfaces” that use intelligent reflection can make a large difference in indoor reception. They would be smart enough to adapt to changing conditions (people, furniture changes, relocation of indoor machinery, etc.). The challenges are how to make them inexpensive, reliable, and flexible as well as with measuring performance.
Plenty of work remains and the challenges, especially to make them inexpensive, are significant.
SubTHz technology (>100GHz)
The attraction of very wide bandwidths available at frequencies above 100GHz has been dampened by the lack of commercial success at the more modest FR2 bands described above. This is exacerbated by the fact that SubTHz is even more expensive and difficult to manage than 24-71GHz.
Research remains significant in industry and in academia, but SubTHz is no longer under consideration for mainstream use as a 6G radio access technology. That said, there are significant and successful demonstrations of point-to-point “microwave” links using D-band technology (110-170GHz).
The significant demand on backhaul data capacity may drive further investment in these ever-higher frequencies in this and in other niche applications.
As expected, the technologies under investigation include semiconductors, antennas, beam management, high-speed DSP, and even in-band full duplex (double the data rate by Tx and Rx at the same time), and like all things, remains in the context of economic constraints.
Author details: Roger Nichols, 6G Program Manager, Keysight
