Requirements come before radios
It is tempting to start 6G with waveforms and antennas. Standards bodies do the opposite. First they agree what the network must achieve — the usage scenarios and the capability targets — and only then do engineering groups choose technologies that can hit those targets. Getting the requirements wrong is expensive, so this phase is deliberate and, for 6G, it is the phase we are in right now.
There are two requirement tracks, and they feed each other:
- ITU-R IMT-2030 — the global, vendor-neutral vision and the capability framework that any "6G" must satisfy to be recognised as an IMT-2030 radio interface.
- 3GPP — the body that writes the actual specifications, working to its own service requirements (SA1,
TR 22.870) and radio requirements (RAN,TR 38.914), which it also submits as input to ITU-R.
The six IMT-2030 usage scenarios
In November 2023, ITU-R approved Recommendation M.2160 — the "IMT-2030 Framework". It defines six usage scenarios. Three extend familiar 5G scenarios (the inner ring below); three are genuinely new to 6G (the outer wedges).
Immersive Communication, Hyper Reliable & Low-Latency Communication (HRLLC), Massive Communication, Ubiquitous Connectivity, Artificial Intelligence & Communication, and Integrated Sensing & Communication (ISAC). The first three generalise 5G's eMBB/URLLC/mMTC; the last three are new.
itu.int — IMT-2030 2023-11Capability targets — and why they are not "specs"
M.2160 also lists 15 capabilities. About nine carry example numeric targets; the rest are described qualitatively (security, sustainability, interoperability, sensing accuracy, AI). The crucial nuance, repeated by ITU-R itself: these numbers are example targets in the framework, not the minimum requirements 6G will be tested against. Those minimum technical performance requirements were still in draft in 2026.
ITU-R Working Party 5D adopted a draft of the IMT-2030 minimum technical performance requirements and evaluation methodology in February 2026; Study Group 5 approval was scheduled for around December 2026, and the draft values were restricted to ITU-R members until then. So the only public numbers are the M.2160 framework examples.
itu.int 2026-03With that caveat front of mind, here is how the framework's example targets compare with the 5G (IMT-2020) baseline they build on. Toggle between the conservative and ambitious ends of each 6G example range.
| Capability | 5G baseline (M.2410) | 6G example target (M.2160) | Multiplier | Status |
|---|---|---|---|---|
| Peak data rate (DL) | 20 Gbit/s | 50–200 Gbit/s | 2.5–10× | requirement |
| User-experienced data rate (DL) | 100 Mbit/s | 300–500 Mbit/s | 3–5× | requirement |
| Area traffic capacity | 10 Mbit/s/m² | 30–50 Mbit/s/m² | 3–5× | requirement |
| Connection density | 10⁶ /km² | up to 10⁸ /km² | up to 100× | requirement |
| Mobility | 500 km/h | 500–1000 km/h | 1–2× | requirement |
| U-plane latency | 1 ms | 0.1–1 ms | 1–10× lower | requirement |
| Reliability | 10⁻⁵ (99.999%) | up to 10⁻⁷ (99.99999%) | 100× lower FER | requirement |
| Spectral efficiency | IMT-2020 baseline | 1.5–3× IMT-2020 | 1.5–3× | requirement |
| Positioning accuracy (indoor) | meter-level | < 0.1 m | >10× better | requirement |
| Security & resilience | 5G level | Enhanced — qualitative | — | requirement |
| Sustainability / energy efficiency | 5G level | Improved — qualitative | — | requirement |
| AI & communication | Limited | Native — qualitative | — | requirement |
| Coverage | Terrestrial | Global incl. NTN — qualitative | — | requirement |
Worked example — the peak data rate target
A worked number, done honestly. 5G's IMT-2020 peak data rate requirement is a real, published value; 6G's is a framework example. We are allowed to compute the ratio, as long as we label both ends correctly.
// 5G baseline — ITU-R M.2410, a real published requirement R_5G = 20 Gbit/s // label: 5G baseline // 6G framework example — ITU-R M.2160, an EXAMPLE target, not a spec R_6G ∈ { 50, 100, 200 } Gbit/s // label: framework example target // headroom = R_6G / R_5G low = 50 / 20 = 2.5× mid = 100 / 20 = 5× high = 200 / 20 = 10×
The 3GPP requirement studies
3GPP turns the vision into its own engineering requirements. Two reports matter most right now, and their status is very different — one is finished, one is roughly half-written.
The Stage-1 study on 6G use cases and service requirements was approved at 3GPP TSG SA#111 in Fukuoka (March 2026). At roughly 590 pages it is deliberately broad: immersive communication, hyper-reliable low-latency, more-massive IoT, ubiquitous connectivity, AI/ML services and ISAC. It becomes the Stage-1 anchor for Release 21 normative work.
3gpp.org 2026-03-13The RAN study on 6G scenarios and requirements was reported about 60% complete in March 2026, with approval targeted around June 2026. It sets deployment scenarios (dense urban, industrial, high-mobility, NTN), KPIs, spectrum and device assumptions, and the high-level 5G-to-6G migration story, and is an initial input to ITU-R's IMT-2030 work.
3gpp.org 2026-03-01The machinery is the same one that produced 5G NR: SA1 service requirements, then RAN scenario/requirement studies, then normative Stage-2/Stage-3 specifications. If you understand how 5G's requirements drove its design, you already understand the shape of 6G's process — only the contents are new.
3gpp.org — Release 20 2026-06-15Where this connects
These requirements set the bar that every other 6G topic must clear. The capability targets drive the radio physical layer and the spectrum choices; the new usage scenarios are why AI-native, sensing and non-terrestrial work exists at all.