Systems in depth
One section per system: its state at evidence close, its exposures and what the operational series will watch.
Sommaire
- ·Edition home
- ·Executive summary
- 1Connectivity is radio, compute and electricity
- 2Key findings
- 3Systems and sectors at baseline
- 4Risks and capability
- 5Compute-heavy radio networks built ahead of Release 21
- 6Readings from the other desks
- 7The 2027 agenda
- 8Systems in depth
- AMéthodologie
- BData appendix
- CReferences and record
- ·Complete edition, for print
8.1Radio and compute platforms
Mobile networks run on radio sites and baseband compute. Evidence is strongest for standards and subscription counts, and weakest where claims rest on company announcements.
| Composante | État | Preuves | Statut |
|---|---|---|---|
| Terrestrial radio | 5G-Advanced under Release 20; first normative 6G work in Release 21 | 34 | Déclaré |
| AI-RAN compute | AI-RAN products announced; T-Mobile field evaluations from 2026, company announcement | 10 | Déclaré |
| Interfaces ouvertes | O-RAN Alliance founded in 2018; architecture described in a tutorial survey | 1318 | Déclaré |
| Non-terrestrial links | first normative requirements in Release 17; Release 19 core completed September 2025 | 14 | Déclaré |
Couplages
- AIC ↔ INF compute and radio
- INF → DAT sensing and positioning
- TRD → INF suppliers
Evidence to bind in 2027
- AI-RAN performance on a common basis
- compute volumes in radio sites
Related editions
- GLB Global Systems
- GS-02 Energy
- XT-06 Advanced Semiconductors and AI Compute
- XT-08 Cloud, Edge and Distributed Computing
- XT-18 Space and Orbital Infrastructure
- XT-10 Cybersecurity and Advanced Cryptography
8.2Energy and sustainability
Networks run on electricity at every site. The energy footprint of AI-RAN is not published, and the energy cost of AI itself is often overlooked.
| Composante | État | Preuves | Statut |
|---|---|---|---|
| Traffic growth | traffic grows about 2.2 times to 2031 excluding FWA | 7 | Déclaré |
| Energy cost of AI | often overlooked; real-world power data hard to obtain | 16 | Déclaré |
| RAN energy share | 87% of participating operators’ energy in the 2026 benchmark | 22 | Déclaré |
| Design aim | ITU names energy efficiency and sustainability as aims for IMT-2030 | 1 | Déclaré |
Couplages
- ENE ↔ INF site power
- AIC → ENE compute load
- INF → CLI emissions
Evidence to bind in 2027
- energy per bit of AI-RAN platforms
- primary energy-share figure
Related editions
- GLB Global Systems
- GS-02 Energy
- XT-06 Advanced Semiconductors and AI Compute
- XT-08 Cloud, Edge and Distributed Computing
- XT-18 Space and Orbital Infrastructure
- XT-10 Cybersecurity and Advanced Cryptography
8.3Chips, spectrum and materials
Radio equipment depends on spectrum, chips and materials. Spectrum decisions are dated; chip and mineral dependence is not bound.
| Composante | État | Preuves | Statut |
|---|---|---|---|
| Candidate spectrum | WRC-27 item 1.7 studies 4400 to 4800 MHz, 7125 to 8400 MHz and 14.8 to 15.35 GHz, or parts of them | 25 | Déclaré |
| Bands above 100 GHz | ITU technical feasibility report completed in 2024 | 2741 | Déclaré |
| Compute chips | AI-RAN uses GPU-based platforms; efficiency claims are company and analyst statements | 1031 | Déclaré |
| Minerals and RF chips | dependence not bound | 34 | Expert input |
Couplages
- MIN → INF materials
- TRD → INF export rules
- GOV → INF spectrum
Evidence to bind in 2027
- WRC-27 outcomes
- chip and mineral dependence
Related editions
- GLB Global Systems
- GS-02 Energy
- XT-06 Advanced Semiconductors and AI Compute
- XT-08 Cloud, Edge and Distributed Computing
- XT-18 Space and Orbital Infrastructure
- XT-10 Cybersecurity and Advanced Cryptography
8.4Economy, investment and workforce
The RAN market is stable and concentrated. Public research funding is modest, and operator investment and workforce numbers are the thin parts of the evidence.
| Composante | État | Preuves | Statut |
|---|---|---|---|
| RAN market | stable in 2025 after losing about a fifth of value from 2022 to 2024 | 9 | Déclaré |
| Public research funding | EUR 900 million from the EU, private members expected to match | 12 | Déclaré |
| Analyst projection | an AI-RAN opportunity above a cumulative USD 200 billion by 2030, an Omdia forecast cited in a company announcement; weak evidence | 24 | Déclaré |
| Main-d'œuvre | no figure bound | 32 | Expert input |
Couplages
- FIN → INF operator capital
- FIN → AIC compute investment
- HLT → INF workforce
Evidence to bind in 2027
- operator 6G investment
- workforce size and skills gaps
Related editions
- GLB Global Systems
- GS-02 Energy
- XT-06 Advanced Semiconductors and AI Compute
- XT-08 Cloud, Edge and Distributed Computing
- XT-18 Space and Orbital Infrastructure
- XT-10 Cybersecurity and Advanced Cryptography
8.5Security, spectrum and governance watch
Standards, spectrum and supplier rules set the pace of 6G. The EU proposal and the ITU process are the nearest dated steps.
| Composante | État | Preuves | Statut |
|---|---|---|---|
| Processus normatif | ITU window February 2027 to February 2029; Release 21 timetable to March 2029 | 23 | Déclaré |
| EU framework | single passport, EU satellite authorisation and copper phase-out plans proposed | 11 | Déclaré |
| 6G security work | security, charging and lawful interception study items in Release 20 | 4 | Déclaré |
| Supplier security | EU toolbox of 2020 sets measures on high-risk suppliers; an ENISA threat landscape of 2019 preceded it | 3542 | Déclaré |
Couplages
- GOV → INF timing
- GOV → TRD supplier rules
- INF → DAT sensing data
Evidence to bind in 2027
- text of the draft ITU reports
- national rules by country
Related editions
- GLB Global Systems
- GS-02 Energy
- XT-06 Advanced Semiconductors and AI Compute
- XT-08 Cloud, Edge and Distributed Computing
- XT-18 Space and Orbital Infrastructure
- XT-10 Cybersecurity and Advanced Cryptography
Notes
Numbers are the edition’s reference numbers, so a note reads the same in every chapter. The full list, with every source and its licence, is in Annex C. Annex C, references
- 1ITU-R (ed.), Recommendation ITU-R M.2160-0, Framework and overall objectives of the future development of IMT for 2030 and beyond, November 2023 (PDF opened; capability research targets in section 4). International Telecommunication Union, Geneva. Observation year 2023. Accessed 2026-11-03.back
- 33GPP (ed.), Release 21 page (timetable agreed at the June 2026 TSG#112 plenaries; first normative 6G release; Stage-3 freeze December 2028; ASN.1 freeze March 2029). 3rd Generation Partnership Project, Sophia Antipolis. Observation year 2026. Accessed 2026-11-03.back
- 10NVIDIA and Nokia (eds.), NVIDIA to invest USD 1 billion in Nokia to accelerate AI-RAN innovation (joint company press release of 28 October 2025; USD 6.01 per share; ARC-Pro platform; T-Mobile field evaluations from 2026; company announcement, weak evidence). NVIDIA, Santa Clara, and Nokia, Espoo. Observation year 2025. Accessed 2026-11-03.back
- 43GPP (ed.), Release 20 page (5G-Advanced timeline; early 6G studies listed, including TR 22.870 and TR 38.914; work plan of March 2026). 3rd Generation Partnership Project, Sophia Antipolis. Observation year 2026. Accessed 2026-11-03.back
- 13O-RAN Alliance (ed.), About page (founded February 2018 by AT&T, China Mobile, Deutsche Telekom, NTT DOCOMO and Orange; German entity since August 2018; no member count or AI statement on the page). O-RAN Alliance, Alfter. Observation year 2018. Accessed 2026-11-03.back
- 18Polese, M., Bonati, L., D’Oro, S., Basagni, S. and Melodia, T., Understanding O-RAN: Architecture, Interfaces, Algorithms, Security, and Research Challenges, arXiv:2202.01032, submitted 2 February 2022 (abstract page opened). arXiv, Ithaca. Observation year 2022. Accessed 2026-11-03.back
- 143GPP (ed.), Non-terrestrial networks overview (first normative requirements in Release 17; Release 19 core completed September 2025 and performance March 2026). 3rd Generation Partnership Project, Sophia Antipolis. Observation year 2025. Accessed 2026-11-03.back
- 7Ericsson (ed.), Ericsson Mobility Report, June 2026, mobile traffic forecast (company forecast; 146 EB per month excluding FWA at the end of 2025; 328 EB per month in 2031; 22 GB per active smartphone per month in 2025). Ericsson, Stockholm. Observation year 2025. Accessed 2026-11-03.back
- 16Sthankiya, K., Saeed, N., McSorley, G., Jaber, M. and Clegg, R.G., A Survey on AI-driven Energy Optimisation in Terrestrial Next Generation Radio Access Networks, IEEE Access 12, 157540 to 157555, 2024 (arXiv:2411.02164, submitted 4 November 2024; abstract page opened). IEEE, Piscataway. Observation year 2024. Accessed 2026-11-03.back
- 22GSMA Intelligence (ed.), Energy and Towers Efficiency Benchmark, presentation by Tim Hatt, March 2026 (PDF slides; Telco Energy Benchmark 2026 covering 19 network operator groups and TowerCos, 118 networks in 76 countries; radio access network 87% of participating operators’ energy use, core and data centres 11%, others 2%; headline efficiency 0.08 kWh per GB; the full report is gated and was not opened). GSMA, London. Observation year 2026. Accessed 2026-11-05.back
- 27ITU-R (ed.), Report ITU-R M.2541-0, Technical feasibility of IMT in bands above 100 GHz, May 2024 (publication page; in force). International Telecommunication Union, Geneva. Observation year 2024. Accessed 2026-11-05.back
- 25ITU (ed.), Resolution 256 (WRC-23), Sharing and compatibility studies and development of technical conditions for the use of IMT in the frequency bands 4400 to 4800 MHz, 7125 to 8400 MHz (or parts thereof) and 14.8 to 15.35 GHz for the terrestrial component of IMT, World Radiocommunication Conference, Dubai, 2023 (PDF of the resolution; studies to be completed in time for WRC-27, which is invited to consider identification of the bands; WRC-27 agenda item 1.7). International Telecommunication Union, Geneva. Observation year 2023. Accessed 2026-11-05.back
- 41Rappaport, T.S., Xing, Y., Kanhere, O., Ju, S., Madanayake, A., Mandal, S., Alkhateeb, A. and Trichopoulos, G.C., Wireless Communications and Applications Above 100 GHz: Opportunities and Challenges for 6G and Beyond, IEEE Access 7, 78729 to 78757, 2019, DOI 10.1109/ACCESS.2019.2921522 (Crossref record). IEEE, Piscataway. Observation year 2019. Accessed 2026-11-05. doi:10.1109/ACCESS.2019.2921522back
- 31ABI Research (ed.), Mavrakis, D., Are GPUs worth considering for 5G RAN?, insight IN-7808, 17 April 2025 (analyst commentary on the NVIDIA and SoftBank GPU-based AI-RAN trial claims; ABI estimates about 1 Gbps per cell and up to 20 cells per server and finds custom silicon ahead for RAN processing alone, attributing the gap to software optimisation). ABI Research, Oyster Bay. Observation year 2025. Accessed 2026-11-05.back
- 34IEA (ed.), Global Critical Minerals Outlook 2025 (PDF, May 2025; the 2026 edition returned access errors; average share of the top three mining countries for key energy minerals 77% in 2024; export restrictions on gallium, germanium and antimony, key minerals for semiconductor production, from December 2024; no figure taken from it in this edition). International Energy Agency, Paris. Observation year 2024. Accessed 2026-11-05.back
- 9Dell’Oro Group (ed.), RAN market stabilised in 2025 (press release of 17 February 2026; analyst release; stable 2025 revenue; Huawei, Ericsson, Nokia, ZTE and Samsung in unchanged order; concentration at a ten-year high; 1% a year growth forecast 2025 to 2030; about a fifth of value lost 2022 to 2024). Dell’Oro Group, Redwood City. Observation year 2025. Accessed 2026-11-03.back
- 12European Commission (ed.), Smart Networks and Services Joint Undertaking policy page (at least EUR 1.8 billion for 2021 to 2027; EUR 900 million public contribution; 20 6G research projects with EUR 116 million announced March 2026). Directorate-General for Communications Networks, Content and Technology, Brussels. Observation year 2026. Accessed 2026-11-03.back
- 24NVIDIA (ed.), NVIDIA and Nokia to Pioneer the AI Platform for 6G (company press release of 28 October 2025; the AI-RAN opportunity is expected to exceed a cumulative USD 200 billion by 2030, citing the Omdia RAN Market Tracker 2Q25 forecast for 2025 to 2030; analyst projection in a company announcement, weak evidence). NVIDIA, Santa Clara. Observation year 2025. Accessed 2026-11-05.back
- 32GSMA (ed.), The Mobile Economy 2026 (PDF, released March 2026; operator capex USD 1.2 trillion for 2025 to 2030; 50 million jobs supported by the mobile ecosystem in 2025; 6G expected to be deployed from around 2030; no figure taken from it in this edition). GSMA, London. Observation year 2025. Accessed 2026-11-05.back
- 11European Commission (ed.), Digital Networks Act policy page (proposal adopted 21 January 2026; single passport; EU satellite spectrum authorisation; copper phase-out plans; preparedness plan). Directorate-General for Communications Networks, Content and Technology, Brussels. Observation year 2026. Accessed 2026-11-03.back
- 2ITU-R (ed.), IMT-2030 process page of Working Party 5D (timeline; candidate proposal window February 2027 to February 2029; draft reports 5/116 and 5/119 completed in 2026 and due for SG 5 approval in December 2026). International Telecommunication Union, Geneva. Observation year 2026. Accessed 2026-11-03.back
- 53GPP (ed.), Introducing 3GPP (seven organisational partners: ARIB, ATIS, CCSA, ETSI, TSDSI, TTA, TTC). 3rd Generation Partnership Project, Sophia Antipolis. Observation year 2026. Accessed 2026-11-03.back
- 35NIS Cooperation Group (ed.), Cybersecurity of 5G networks: EU Toolbox of risk mitigating measures, January 2020 (library page dated 23 January 2020; common set of measures to mitigate the main cybersecurity risks of 5G networks). Directorate-General for Communications Networks, Content and Technology, Brussels. Observation year 2020. Accessed 2026-11-05.back
- 42ENISA (ed.), ENISA Threat Landscape for 5G Networks, 21 November 2019 (publication page; initial threat landscape with 5G architecture, asset diagram and threat assessment). European Union Agency for Cybersecurity, Athens. Observation year 2019. Accessed 2026-11-05.back