The short answer
If you are wondering whether 6G is just a faster 5G with a new label, it is not. 6G is a different stack — different spectrum, a different control plane, and a different design goal. The targets agreed at the ITU in February 2026 call for peak rates around 100 to 200 Gbps, latency under 1 millisecond, and an air interface that treats sensing and on-device AI as first-class features instead of bolt-ons. The first commercial deployments are expected around 2030, with 3GPP Release 21 due in 2028.
You do not need to do anything yet. You do need to understand that the gap between 5G and 6G is bigger than the gap between 4G and 5G, and it lands in time to reshape phones, factories, and cars around 2031 and 2032. Below is what changes, what stays the same, and the trade-offs nobody puts on a keynote slide.
What 6G actually is
6G is the sixth generation of cellular wireless, formalised under the ITU’s IMT-2030 framework. Recommendation ITU-R M.2160-0 set the vision in 2023, and the technical performance requirements were agreed at the ITU-R WP 5D meeting in Geneva in February 2026. Formal approval is expected when ITU-R Study Group 5 meets in December 2026. 3GPP picks up the work in Release 19 in 2025, then Release 20 in 2026 and 2027, with the first true 6G specification landing in Release 21 by the end of 2028. Commercial rollout follows in 2030.
Three things separate 6G from a 5G refresh. The first is spectrum: 6G adds the 7 to 24 GHz mid-band — sometimes called the “FR3” range — as the workhorse, with sub-terahertz bands above 100 GHz for short-range capacity. The second is architecture: AI is native to the radio access network, not an overlay. The third is sensing: the same radio that carries your data also maps the room it is in, a capability the standards call Integrated Sensing and Communication, or ISAC.
5G vs 6G — the specs that matter
| Spec | 5G (IMT-2020) | 6G (IMT-2030 target) |
|---|---|---|
| Peak downlink | 20 Gbps | 100–200 Gbps |
| User-experienced rate | 100 Mbps | 300–500 Mbps |
| Latency (air interface) | 1 ms | 0.1–1 ms |
| Connection density | 1 million devices / km² | 10 million devices / km² |
| Primary spectrum | sub-6 GHz + mmWave (24–47 GHz) | 7–24 GHz mid-band + sub-THz (100–300 GHz) |
| AI in the stack | Application-layer add-on | Native to the air interface |
| Sensing | Not in scope | Integrated sensing and communication (ISAC) |
| Non-terrestrial coverage | 5G NTN added later | Satellite integration designed in from day one |
| Energy per bit | Baseline | Target: significantly lower than 5G |
| Standards body | 3GPP Release 15 (2018) | 3GPP Release 21 (2028) |
| Commercial launch | 2019 | ~2030 |
The headline numbers — 1 Tbps peak in some industry decks, 0.1 ms latency — come from research papers, not the ITU targets. Peak-rate marketing claims have always overshot what users actually experience by an order of magnitude. The realistic user-rate jump is roughly 3 to 5x over a good 5G connection, not 50x.
Why terahertz changes the rules
5G’s mmWave bands (24 to 47 GHz) already taught the industry that high frequency means more capacity and less reach. 6G’s sub-terahertz bands push that further. At 140 to 350 GHz, free-space path loss exceeds 100 dB at just 10 metres, water vapour in the air absorbs signal, and a brick wall is effectively a brick wall. That is why most 6G deployments will rely on the 7 to 24 GHz mid-band for coverage and reserve sub-THz for short-range hotspots — a stadium, a factory floor, a campus.
The practical effect is a network of much smaller cells, denser in cities, and a lot more reliance on reconfigurable intelligent surfaces — flat, programmable reflectors that can bounce a signal around an obstacle. None of that is free. The energy cost of generating and modulating sub-THz signals is high, and thermal management in the radio head is one of the unsolved problems on the IMT-2030 roadmap.
The use cases that justify it
Three workloads sit behind almost every 6G keynote: holographic and volumetric communication, real-time digital twins of factories and cities, and ambient sensing that turns the network into a passive radar. A European 6G-XR consortium led by Capgemini, Ericsson, i2CAT and Vicomtech showed holographic calling running on standalone 5G at MWC 2026 — the point being that 5G can demo this, while 6G is designed to do it routinely.
The less glamorous use case is the one that may pay for the upgrade: industrial automation. Sub-millisecond latency and 10 million devices per square kilometre is what lets a car plant run thousands of motion-controlled robots, AGVs, and sensors on one wireless fabric instead of cabled fieldbuses. That is also where the AI-native stack matters: AI-RAN, the work Nokia is doing with Nvidia and the OCUDU foundation Ericsson joined in 2026, is what makes ultra-dense, ultra-reliable operation tractable. Holographic FaceTime is the marketing; the factory is the bill-payer.
What does not change
A few things that get oversold. 6G is not going to replace fibre to the home for most households — wired backhaul is still cheaper and lower-latency at the building level. Your phone in 2030 will not get 100 Gbps; you will get a better 5G-like experience with much fewer dead spots, because mid-band coverage and satellite integration close the gaps. And the long tail of 4G and even 3G coverage in developing markets will not vanish on the 6G launch date — 5G itself only crossed roughly half of global mobile connections in 2024.
Pricing is the other quiet question. Carriers spent a reported $1 trillion on 5G capex and are still arguing about the return. They will not repeat that without a clearer monetisation story, which is why the industry is pushing enterprise, automotive, and fixed wireless as the early 6G revenue lines, not consumer plans.
Should you care yet
If you buy a phone in 2026 or 2027, 6G is not a reason to wait. If you run a factory, a logistics yard, a hospital, or a venue that is planning a multi-year wireless refresh, it is the reason to make your 5G build forward-compatible — same vendor, software-upgrade path, mid-band-ready radio heads. Ask your integrator whether the 5G-Advanced gear they are selling you now has a published 6G upgrade path; the honest ones will tell you whether the radio is reusable or not.
For consumers, the change you will actually notice arrives quietly. Phones around 2031 will have 6G modems, coverage will feel better rather than faster, and a few new things — your room mapping itself, your watch sensing a fall via the network — will start to feel routine. The headline 1 Tbps demos will stay in trade-show booths, where they belong.
The verdict
6G is not a marketing rebrand. The spectrum, the architecture, and the design goals are all different from 5G, and the ITU’s February 2026 agreement on technical performance requirements means the standard is now real engineering work, not slideware. Expect the first networks in 2030, real coverage in 2032, and the enterprise use cases — factories, logistics, healthcare — to lead consumer rollout by two or three years.
The buyer who should pay attention now is the operations leader signing a multi-year industrial wireless contract. The buyer who can ignore it for another three years is everyone else.
Specifications and timelines cited from ITU-R M.2160-0, the ITU-R WP 5D February 2026 meeting outcomes, and 3GPP Release 19–21 schedules; verify against the latest ITU and 3GPP publications before quoting.
