What you’ll learn:
- Today’s business trends in 6G cellular.
- What are the latest technology trends?
- The migration to 6G solutions.
The typical development cycle for a cellular generation is 10 years. With the first 6G technology deployments slated for 2030, that puts us roughly halfway through the development cycle as 2024 comes to an end and 2025 kicks off. Although it will be another two to three years before the first 3GPP standard for 6G is finished, this is a good time to reflect on the current research and economic trends that need to be addressed to make 6G a success.
6G Business Trends
For 6G, considering the current economic state adds clarity to why some technologies and applications are being prioritized over others. In the past four years, there was a global pandemic, multiple wars, disrupted supply chains, and an ever-present threat of recession. Economic uncertainty paired with large investments made into 5G infrastructure have created a financially conservative atmosphere for 6G development.
The cellular industry is experiencing a shift from a growth market model to a profitability model. The number of consumer cellular subscriptions is beginning to plateau (Fig. 1). Consumers are also keeping their devices longer.
These two market forces combined are forcing the information and communications technology (ICT) industry to evaluate new business models (for example, the concept of creating application programing interfaces to run on top of the network). Simultaneously, there’s a push to make networks as efficient as possible to lower operational expenses.
5G uses less power per bit than 4G, but the number of bits being transmitted today is exponentially higher than it was 10 years ago. Techniques for power reduction are being used in 5G, and 6G has an opportunity to natively adopt them and implement other improvements. This will make 6G networks cheaper to operate long term while addressing global Net Zero and sustainability goals.
Even if the desire to invest in new infrastructure from telcos is tempered, there’s still a clear need for 6G. The demand for cellular services continues to grow and those demands can’t be met with existing technology alone. Lessons from 5G will help 6G, too, like not having multiple split options and skipping non-standalone and moving straight to standalone.
In parallel, AI and quantum computing are maturing. Both technologies represent a threat to the security of current cellular technology. AI also represents a once-in-a-generation opportunity for innovation.
With all of this in mind, it seems likely that 6G will be a technological evolution, leveraging the massive improvements to the infrastructure and protocol stack that were started in 5G. And as that core technology matures, it can serve as the foundation to revolutionize the applications and use cases for cellular technology.
Technology Trends: From Massive MIMO to O-RAN
At the heart of 6G is spectrum expansion (Fig. 2). New spectrum is required to increase capacity, and the most promising bands for 6G are between 7 and 24 GHz (or “FR3” or the upper-mid band). The 7- to 15-GHz band has preferable propagation characteristics—it’s the spectrum that the ICT industry is pushing to be earmarked for licensed cellular transmission. There are many incumbents across this spectrum and regulatory challenges must be addressed.
Sharing spectrum is a proposed way to help alleviate some of these challenges, but sharing spectrum historically has been less than ideal. With advances in AI, which is uniquely well-suited to optimizing complex systems like wireless networks, new solutions for spectral sharing in 6G hold promise.
MIMO improvements have played a significant role in every 3GPP release since release 15, and this trend will continue. Moving into slightly higher bands makes it possible for massive MIMO to evolve thanks to smaller antenna sizes. The ability to add more antenna elements to MIMO systems without dramatically increasing their size is another way that spectral efficiency is being addressed.
The trend to make networks more disaggregated, open, and software-defined is expected to continue into 6G. While O-RAN has had lackluster success so far, it’s far from being a failure. With current geopolitical tensions, O-RAN is a public policy priority for the United States. That pressure will allow for further investment and evolution for open RAN architectures.
Another reason why the move toward O-RAN is important is because it enables innovation. Being able to develop apps to run in the RAN intelligent controller (RIC), for example, gives engineers a way to contribute to the stack outside of 3GPP, which is valuable for small companies and academic researchers who don’t have the resources to participate in the standards.
Non-Terrestrial Networks
6G hopes to address numerous applications, but two have significant traction within the industry: non-terrestrial networks (NTNs) and joint communications and sensing (JCAS, or integrated sensing and communications, ISAC).
Using satellites to deliver connectivity is a promising way to deliver global coverage on a scale that’s long been desired but impractical. The price point to build and launch a satellite has decreased significantly in the last five years. But whether the business case will have long-term success is yet to be seen. In 3GPP, there’s an effort to look at standardizing waveforms and ways to handle mobility.
New channel models are also being built to consider a 3D channel for signals that may be coming from space. The new channel models are also important for sensing applications to account for backscatter and features of sensing waveforms, like radar.
The ability to gain information about the physical world through the communications channel can help in applications like autonomous driving, manufacturing, and gesture recognition. How JCAS will ultimately be used is still to be seen, but adding sensing capabilities to wireless networks is a way to increase networks’ value.
The Omnipresence of AI
Underlying all of the technical advances driving 6G is AI. As mentioned, it can optimize complex wireless networks in ways that aren’t possible with humans. Research is ongoing into how AI can be used for everything from improving energy efficiency to waveform recognition to channel-state-information (CSI) improvements. 3GPP is actively working to build a framework so that AI can be added into cellular standards, but many challenges remain.
AI is already being used in some areas, but many potential use cases for AI need further research before they will be ready. Before AI models can ever be implemented, they need to be trained. Access to training data for cellular applications is limited. Carriers do have data, but for many reasons, including customer privacy, the data will not be shared (or perhaps it will at a steep cost).
Questions remain in terms of how much training data is needed as well as how often models must be retrained. Researchers are actively working to understand these factors. After training, AI models must be hardened and proven to work reliably, predictively, and better before replacing traditional methods. Testing AI models inside of full stack real-time networks is a crucial yet difficult step that’s needed to evaluate this kind of performance. Limited testing is happening, but much more is required.
AI has been touted to improve the energy efficiency of wireless networks. However, AI today is power-hungry. Thus, adding any AI models into a wireless network comes with a risk of using more total energy than making no changes at all.
AI is rapidly maturing, and so is GPU technology. It’s likely that, in the future, the energy efficiency of AI usage will improve. Until that time, the wireless industry needs to better understand the energy tradeoff between using traditional methods and AI. As of now, there’s no industry standard way to measure this tradeoff. Some ideas and methodologies have been proposed, but they need more work.
The wireless industry needs a way to deterministically measure and compare the energy usage of AI-enhanced and traditional networks. Research on this topic has begun and will likely improve over the next year or two.
Overall, AI is fundamentally changing how compute power can be used, and the ICT industry is looking to harness this new power. The first 3GPP release to incorporate 6G study items will start in mid-2025, adding more clarity to what technology will be in or out for the initial 6G roll out. The timing for 6G will enable it to capitalize on the power of AI. It could very well be the most efficient and sustainable cellular generation yet. And with any luck, it will unlock new opportunities for humanity to connect and interact with each other.
