NTT DOCOMO and Samsung have advanced their joint work on intelligent mobile networks with a successful validation of user level AI RAN technology designed to manage wireless resources for individual subscribers. The approach uses artificial intelligence to assess changing radio conditions and adjust bandwidth and network behavior for each user, helping reduce connection drops and improve the consistency of mobile service.
The achievement points toward a major change in how future cellular networks could operate. Instead of treating every device connected to a cell site in roughly the same way, an intelligent network can respond to the needs and circumstances of each person. For someone walking through a crowded station, traveling at speed, moving between coverage areas, or using a demanding application, that distinction could determine whether a video continues smoothly or suddenly freezes.
Why User Level Network Control Matters
Mobile networks have traditionally optimized performance largely at the cell level. A base station observes conditions across its coverage area and manages radio resources according to broader network requirements. That approach has supported generations of mobile technology, but it becomes more difficult when users experience very different conditions within the same coverage area.
Two people standing only a short distance apart can have different signal strength, device behavior, traffic demands, and movement patterns. One may be watching high resolution video while another is sending a few messages. A third person could be approaching the edge of a cell where a handover to another base station is becoming necessary.
NTT DOCOMO and Samsung have been investigating how artificial intelligence can respond to these individual circumstances. Their research has focused on using historical and real time information about factors such as throughput, latency and signal strength to identify patterns that can precede connectivity problems. DOCOMO has described this direction as part of its broader work toward intelligent 6G network control. :contentReference[oaicite:0]{index=0}
The companies’ work builds on a collaboration announced in 2024 that specifically targeted user level optimization. At the time, the partners explained that artificial intelligence could help optimize communication for people entering weak signal areas or moving between cell boundaries, potentially allowing services such as video streaming to continue with fewer interruptions. :contentReference[oaicite:1]{index=1}
How AI Can Adapt the Network to Individual Users
The central idea is relatively simple even though the technology behind it is complex. The network continually gathers information about a user’s wireless experience and uses AI models to identify conditions associated with future connection problems.
When the system detects a likely problem, it can determine an appropriate response for that particular subscriber rather than waiting for the connection to deteriorate. Possible actions include changing a user’s radio band or adjusting handover behavior as conditions change.
DOCOMO’s published material describes user specific control as including functions such as band switching and handover. The broader research also involves collecting historical user perception data and applying machine learning to predict connection failures. :contentReference[oaicite:2]{index=2}
This matters because a dropped connection is rarely experienced by a customer as a technical event. For the person holding a phone, it may simply appear as a frozen video, a failed voice call, a delayed message, or an online meeting that suddenly becomes unusable.
Reducing those moments could therefore have a more noticeable effect on everyday communication than simply increasing a network’s theoretical peak speed.
The Significance for 6G Development
Sixth generation mobile technology is being designed around networks that are expected to become substantially more intelligent. Artificial intelligence is increasingly being considered not merely as an application carried over the network, but as a component that can help operate and control the network itself.
That distinction is important. A conventional network provides connectivity for an AI application running somewhere else. An AI driven network can also use intelligence to determine how connectivity itself should be managed.
NTT DOCOMO has been pursuing several related technologies. In March 2026, the company reported a demonstration with VIAVI that used AI driven RAN control and a digital twin to improve throughput by reducing control overhead. DOCOMO said the demonstration produced a throughput improvement of up to 20 percent in the tested scenario. :contentReference[oaicite:3]{index=3}
The company has also been investigating how virtualized RAN infrastructure can support AI processing. Another 2026 demonstration showed AI applications operating alongside communication processing on general purpose CPU resources within a commercially deployed virtualized RAN environment. :contentReference[oaicite:4]{index=4}
Together, these projects illustrate a broader direction. Future networks could increasingly combine communications, computing, automation and AI within the same infrastructure.
From Network Wide Optimization to Personal Connectivity
The most significant aspect of the DOCOMO and Samsung work is the shift from network wide decisions toward individual connectivity management.
Consider a commuter walking through a large railway station. Hundreds or thousands of devices may be competing for radio resources. Some users are stationary, others are moving rapidly, and many are transitioning between coverage zones. Their applications also place different demands on the network.
A network that understands those differences can potentially allocate resources more intelligently. Someone approaching a weak coverage area while participating in a video call may require a different response from someone downloading a small file.
The objective is not necessarily to give every user more bandwidth at all times. Instead, AI can help determine when additional resources are needed, when a different radio configuration is appropriate, and when a handover should occur.
That could make network capacity more useful without requiring operators to solve every performance problem simply by adding more infrastructure.
Samsung and DOCOMO Have Been Building Toward This Point
The latest validation fits into a longer relationship between the two companies. Samsung and NTT DOCOMO have collaborated on mobile communication technologies and standardization activities through the 3rd Generation Partnership Project, commonly known as 3GPP. Their AI research is intended to contribute to the development of future communication systems and the broader transition toward 6G. :contentReference[oaicite:5]{index=5}
Samsung has also pursued user specific AI RAN optimization with other operators. In December 2025, Samsung and KT announced validation of AI RAN optimization technology on a commercial network. Their field testing involved approximately 18,000 users in selected areas of Seongnam, South Korea, and examined how AI could automatically apply network configurations based on individual wireless conditions. :contentReference[oaicite:6]{index=6}
That work provides useful context for the significance of the DOCOMO collaboration. The industry is moving beyond laboratory concepts toward tests that examine how AI based network control behaves under practical conditions.
What This Could Mean for Mobile Users
For consumers, the value of this technology will ultimately be measured through experiences rather than technical terminology.
- Fewer unexpected connection drops during movement
- More stable video and voice communication in difficult coverage areas
- More responsive handovers between network cells
- More efficient allocation of radio resources during periods of congestion
- Potentially more consistent performance for demanding applications
These improvements could become increasingly relevant as mobile services carry richer video, cloud applications, immersive experiences, connected devices and AI services.
For operators, the potential benefit is equally significant. Better resource allocation could allow existing infrastructure to serve users more efficiently, while intelligent automation could reduce the burden of manually tuning increasingly complicated networks.
AI RAN Also Brings New Questions
The promise of user level AI RAN should not be confused with an immediate commercial arrival of 6G. Standards, interoperability, security, reliability and large scale deployment still require substantial work.
AI driven decisions also need to be predictable and accountable. A system that continuously evaluates user conditions must be designed to protect sensitive information and prevent unintended discrimination in how network resources are allocated. Operators will need strong safeguards around data collection, model behavior and operational control.
Energy consumption is another consideration. AI processing requires computing resources, and future networks could generate enormous volumes of data. The efficiency gained through intelligent resource management will need to outweigh the additional computational demands of operating AI models across large telecommunications systems.
DOCOMO and SK Telecom’s 2026 white paper on the evolution of virtualized RAN toward AI RAN highlights resource pooling, separation of hardware and software, and the use of network infrastructure for AI computing as important elements of the transition. :contentReference[oaicite:7]{index=7}
A More Personal Vision of 6G
We often describe the next generation of mobile technology through headline figures for speed, latency and capacity. Those measurements remain important, but user level AI RAN suggests that the real test of 6G may be more personal.
A better network should not simply deliver impressive numbers under ideal laboratory conditions. It should understand when a person is likely to lose connectivity and respond before that failure becomes visible.
That could be particularly valuable in places where reliable communication carries consequences beyond entertainment. Remote workers depend on stable video calls. Emergency responders need dependable communications. Connected vehicles require timely data. Industrial systems and future robots may need networks that can react to rapidly changing conditions without interruption.
The broader technical direction can be followed through the NTT DOCOMO technology and research resources and the global work of 3GPP, where the standards foundations for future mobile systems continue to develop.
What Comes Next
The successful validation by NTT DOCOMO and Samsung represents another step toward networks that can make decisions at the level of individual users rather than treating an entire cell as one uniform environment.
The immediate significance is practical: AI can help anticipate connectivity problems, adapt radio resources and reduce the likelihood of service interruptions. The longer term significance is architectural. If this approach scales successfully, intelligence could become deeply embedded in the operation of future mobile networks.
For users, the most meaningful sign of progress may be surprisingly quiet. A train moves into a weak coverage area, a video continues playing, a call remains clear, and nothing appears to have gone wrong. That invisible reliability is precisely what intelligent network technology is being developed to deliver.
