The Digital Twin (DT) Panel, hosted by the ATIS Next G Alliance’s Technology Roadmap Working Group on March 25, 2026, brought together experts from industry and academia to discuss the future of Digital Twin in the 6G era. The panel included representatives from Ericsson, Qualcomm, InterDigital, Nokia, and the University of Texas at Austin and explored the strategic role of RAN digital twins in the evolution toward 6G networks.

Strategic Role of Digital Twins in 6G

Panelists broadly agreed that digital twins will become a critical development and experimentation environment for 6G. Because future networks will rely heavily on AI/ML-driven automation, operators and vendors will require environments where AI models can be trained, validated, and stress-tested without affecting operational networks. A RAN digital twin can replicate the behavior of the real network and enable large-scale experimentation that would not be feasible in live deployments.

Another key role highlighted was the ability of digital twins to generate synthetic data. Since collecting large volumes of real network data is expensive and often limited by operational constraints, digital twins can provide data for AI model training and validation, accelerating the development of intelligent RAN capabilities. The panel also recognized that digital twins may adopt different strategies to generate synthetic data, including use of ray tracing, full stack emulation, and generative AI, each of which has a different computational cost and will be driven by the downstream task that ingests the data.

Key Use Cases for RAN Digital Twins

Several compelling use cases were identified:

1. Network Design, Planning, and Optimization
   Digital twins can help operators simulate network rollout scenarios, optimize cell placement, and improve coverage and performance before physical deployment.
2. Operational Optimization
   Once deployed, digital twins can support tasks such as interference management, handover optimization, beamforming adjustments, and energy efficiency improvements.
3. AI Training and Agent Development
   Digital twins provide an environment where AI agents can be trained and evaluated safely, enabling more autonomous network management.
4. Integrated Sensing and Communication (ISAC)
   ISAC, an emerging capability expected in 6G, may require new deployment strategies and network configurations. Digital twins can help evaluate technical feasibility and performance trade-offs before operators invest in infrastructure changes.
5. Spectrum Sharing and Coexistence
   Future spectrum frameworks incorporating dynamic sharing approaches, such as CBRS-like models, may require complex coexistence evaluations. Digital twins could simulate spectrum-sharing scenarios and interference environments before real-world deployment.
6. Network Slicing and Service Optimization
   Digital twins can assist in designing custom network slices and admission control strategies tailored to specific applications.
7. TN–NTN Integration
   The integration of terrestrial networks with non-terrestrial networks (satellites) presents significant modeling complexity, including satellite mobility and dynamic channel conditions. Digital twins were highlighted as a powerful tool for validating these hybrid architectures.
8. Controlled Policy Experimentation
   A facet of many of the use cases discussed is policy experimentation. As networks become more autonomous, operators will need to define guardrails and operational policies governing AI-driven decision-making. Digital twins allow these policies to be evaluated in a repeatable and controlled environment, ensuring that network behavior remains within acceptable limits before deployment.

From Simulation to “Living” Digital Twins

A major theme of the discussion was the distinction between traditional network simulations and true digital twins. Traditional simulations typically provide a static and one-directional evaluation of network performance. In contrast, a true digital twin could become a dynamic, real-time replica of the operational network, continuously ingesting real-world data and enabling closed-loop experimentation. Further, a true RAN digital twin requires both network and  user equipment (UE) digital representations, enabling realistic modeling of traffic, mobility, radio behavior, and closed-loop AI-driven network optimization. Designing realistic digital twins and keeping such twins calibrated may need closed-loop participation from components of the network, including the UEs, base stations, and other external sensors.

This capability would allow operators to perform “what-if” experiments, such as testing scheduler changes, new AI algorithms, or parameter adjustments in the twin before implementing them in the real network. While panelists agreed that digital twins will expand the capabilities of network testing and optimization, there was also consensus that traditional simulation tools will likely continue to coexist with digital twins.

Impact on Standardization

The panel also discussed whether digital twins could influence future 6G standardization efforts in 3GPP. Today, many standardization studies rely on stochastic channel models to evaluate network performance. Digital twins, however, can provide site-specific and environment-specific models, potentially offering more realistic evaluations.

However, panelists noted that incorporating digital twin–based modeling into standards would require industry consensus on representative environments and modeling methodologies. While digital twins may eventually impact standardization, particularly in later releases, their immediate impact will likely be more prominent in network deployment, optimization, and operational experimentation.

Broader Opportunities and Challenges

Panelists also suggested that digital twins could provide prior knowledge of the network environment, for example through maps, propagation models, or AI representations of the environment. This information could enable simpler AI models, faster inference, and improved network performance.

Another observation was that digital twins may unlock new use cases that are not yet fully envisioned, particularly as networks become more programmable and AI-driven.

Despite the strong potential, several challenges remain:

• Developing accurate and scalable digital twin models, especially those that fuse multimodal sources of data going beyond radio and network time series KPIs
• Achieving industry consensus on representative environments
• Integrating outcomes from different digital twins that operate at varying timescales, resolutions, and model the same physical environment but from the viewpoint of different layers of the network stack
• Dimensioning the digital twins that model different parts of the network with varying scope and timescales appropriately to manage complexity of the digital twin, especially for twins operating with short timescales.
• Integrating digital twins with standardization frameworks
• Evaluating economic viability and business models that can support the use of digital twins for new use cases such as ISAC

Final Thoughts

The panel concluded that RAN digital twins are likely to become a key enabler for AI-native 6G networks, supporting network design, AI training, operational optimization, and experimentation. While Digital Twin technology is still evolving, its ability to create a realistic, controllable replica of the network environment could significantly accelerate innovation and reduce deployment risk in future wireless systems. The panelists speculated that by 2035, close to 100 percent of the pre-deployment network will be designed using digital twins and a significant portion of operational decisions will be based on Digital Twins.