The results of a recent IEEE survey indicates that 6G remains more of a research topic than a near-term business priority
The Institute of Electrical and Electronics Engineers (IEEE) recently conducted a survey of tech leaders about the impact of various technologies in 2026, finding that only 7% of respondents view 6G as a top area for AI impact — a notable finding coming from a technology often described as AI-native. The association said the finding suggests 6G remains more of a research topic than a near-term business priority.
Speaking with RCR Wireless News about the result, IEEE Senior Member David Witkowski said that to truly understand how 6G might impact AI, the industry first needs to define what 6G will be.
“From that perspective, we also need to look at where we in 5G? Because each generation of cellular technology does two things: It resolves issues with the previous generation or limitations with the previous generation, and it adds features that were not included in the previous generation’s definition,” he said. We’re about halfway through the 5G era, and Witkowski, like most industry experts, believes its impact has been underwhelming thus far in terms of adoption, advanced features, and ROI.
He brings this up to point out that, as a result, 5G is still the focus, with 5G — and in some cases 4G — radios still going up, making it difficult to truly predict exactly what to expect from 6G. “That being said, I do think that there are some things,” Witkowski continued. “Very likely, the thing that 6G will do in the AI space is that it will have to address that question of upload capacity.”
He explained that earlier cellular networks were architected primarily for downloads, which is why 5G can deliver stunning speeds — often exceeding a gigabit per second during speed tests — while upload speeds are typically a fraction of that. This imbalance exists because radio resources and spectrum have historically been optimized to send data to users, not receive large volumes from them. That design creates a growing challenge for AI, which increasingly depends on uploading data such as photos, video, live camera feeds, and voice inputs to the cloud. As AI agents become more interactive and real-time, limited uplink capacity becomes a bottleneck to performance and user experience.
“I mean, when you think about it … The network was designed to deliver videos to us. It was designed for YouTube. It was designed for Instagram. It was designed for TikTok. It wasn’t designed for AI, so 6G has to be,” he said. “It’s an upload versus download question.”
Does it go both ways? Will AI support 6G?
RCR also discussed with Witkowski how, if at all, AI could help 6G address spectrum efficiency or coordination. “I think it’s a question of using AI to predict the behavior of users based upon their past usage patterns,” he surmised. Essentially, instead of relying on a fixed downlink–uplink split, AI could predict network load by learning how different users typically behave. By anticipating which users will need higher upload capacity, the network could proactively shift spectrum and assign them to cells better suited for uplink, improving efficiency without overprovisioning unused resources.
“In general, versus having people try to guess how to configure the network resources, it’s more and more likely that that’s going to be handed off to some sort of machine learning algorithm. And then ultimately the network will try to optimize itself,” he reasoned. “That seems like an obvious use of AI technology to me.”
