5 Questions Operators Should Ask About Capacity, E-band, and RAN Evolution.
Here’s the good news: you’re not going to wake up one morning to find out 10 Gbps backhaul has become obsolete. For most mobile network operators, it still provides substantial headroom for 5G and 5G Advanced. But that won’t be true at every site forever. As operators add spectrum and eventually introduce 6G, some high-capacity sites will need 20 or 25 Gbps transport. Major hubs and aggregation sites may need 50 Gbps or more.
So the question isn’t whether every cell site needs a huge upgrade now. It’s whether the network being built today has somewhere to go next.
This means looking beyond capacity numbers alone. Future RAN architectures could also change requirements for synchronization, latency, resiliency, and even the direction traffic moves through the network.
New Omdia research commissioned by Ceragon explores what those changes could mean for mobile backhaul. And while nobody can definitively draw the final 6G transport architecture yet, we know enough to start asking better questions.
Here are five we’ve come up with.
1. Is 10 Gbps backhaul enough for 5G Advanced?
For most sites, yes.
The move from 4G to 5G already brought a significant jump in transport capacity demand. Omdia identifies 10 Gbps Ethernet as the typical cell-site access speed for 5G, compared with 1 Gbps for 4G.
But there’s an important distinction here: interface capacity is not the same thing as actual traffic demand.
Omdia models a three-sector 5G macro site with 200 MHz of deployed spectrum at 2.3–3.7 Gbps of backhaul demand. A 5G Advanced site with 300 MHz increases that modeled requirement to 3.3–5.3 Gbps. Both remain within the 10 Gbps capacity limitation.
In other words, they’ve still got room to run with existing 10 Gbps sites.
And another reason to not make any expensive, knee-jerk decisions: there is plenty of 5G evolution left before 6G arrives. GSMA reported in early 2026 that only about one-quarter of operator groups worldwide had deployed 5G Standalone at scale.
For many MNOs, the immediate job isn’t ripping out 10 Gbps infrastructure in anticipation of 6G. It’s getting more from 5G while making sure their choices today don’t hamstring the network they’ll need to reach tomorrow.
2. So, how much backhaul will 6G need?
Ask how much backhaul 6G will require and you’ll get the telecom industry’s (and any politician’s) favorite answer:
It depends.
Annoying? Maybe. But in this case, it really is the correct answer.
Added spectrum, site configuration, traffic demand, and RAN architecture can all move the number around quite a bit.
A high-capacity site combining 200 MHz of existing 5G spectrum with 200 MHz of new 6G spectrum produces 7.3–11.7 Gbps of modeled backhaul demand, according to Omdia. In this scenario, Omdia identifies the next step of needing 20-25 Gbps.
And that’s where the conversation gets even more interesting.
The question isn’t whether “6G needs 25 Gbps.” More accurately, it’s which sites will need that much capacity, and when?
A macro site adding a modest amount of spectrum presents a very different transport requirement from a high-traffic site deploying hundreds of megahertz of new upper-mid-band spectrum. Aggregation changes things again too. Omdia expects very large 6G hub sites aggregating traffic from multiple cell sites—or supporting active RAN sharing—to potentially require multiple 25 Gbps links.
But again, 6G won’t arrive everywhere at once. Neither will its backhaul requirements.
That’s why it’s worth checking out the full Omdia model. It breaks the numbers down by spectrum configuration and site scenario rather than assigning one capacity target to an entire network.
3. Can microwave and E-band really keep up?
Yes. And this is where the “more capacity means more fiber” argument begins to fall apart.
Wireless transport isn’t just some fringe technology waiting to be replaced by fiber. Omdia estimates that around 40–50% of mobile cell sites globally use microwave transport.
Fiber and wireless solve different problems. Fiber makes sense where the infrastructure exists or construction economics make sense. Licensed microwave can cover longer suburban and rural link distances. E-band can deliver enormous amounts of wireless capacity. And wireless, as a category, can quickly add big capacity where laying new fiber would be slow, disruptive, or economically unfeasible.
E-band—operating in the 71–76 GHz and 81–86 GHz bands—has become particularly important for high-capacity 5G backhaul. In fact, it now accounts for the majority of 5G wireless backhaul links.
But more importantly, 10 Gbps is not the ceiling when it comes to E-band.
Omdia models a 20 Gbps E-band link over distances up to 3 km at 99.95% availability using dual-radio XPIC and Layer-1 aggregation. Its roadmap extends to 25 Gbps per link and 50 Gbps through aggregation.
That’s not just a little bump above 10 Gbps. E-band vastly changes your network planning conversations. So, an operator looking at a 20 or 25 Gbps site? The answer doesn’t have to be, “welp, I guess it’s time for fiber.”
Of course, there are engineering caveats to E-band. Distance and availability depend on channel bandwidth, antenna gain, modulation, local rain conditions, fade margin, and other link-specific factors. But the technology itself continues to move forward. Our latest E-band solutions are already pushing capacity well beyond 20 Gbps, while the new Omdia paper gets into the engineering behind these higher-capacity configurations.
The point isn’t wireless instead of fiber. It’s that you still have plenty of choices.
4. What if capacity is the easy part?
Here’s the twist: hitting higher capacity may turn out to be the easy part.
We spend a lot of time talking about 10, 20, 25, and 50 Gbps because capacity is wonderfully measurable. But future transport isn’t just a bigger pipe. It may also have to move traffic differently…and do it within tighter time constraints.
We can already see that in 5G.
5G TDD radios require synchronization within ±1.5 μs of time error and ±50 parts per billion of frequency accuracy, according to Omdia. Transport therefore has to preserve timing as well as carry traffic.
Now consider distributed MIMO.
Instead of geographically separated radios operating independently, multiple sites could coordinate transmission. Omdia examines a D-RAN implementation requiring approximately 1 ms one-way site-to-site transport latency for signaling and user-plane traffic.
That may sound like a small architectural change. But it isn’t.
Traditional backhaul traffic predominantly moves from cell sites toward aggregation and the core. Distributed coordination creates more east-west traffic between RAN sites. More centralized RAN architectures introduce another problem, requiring high-capacity, ultra-low-latency fronthaul between centralized functions and radio sites.
And distributed AI compute could create yet another traffic path.
So, to sum up, backhaul, fronthaul, and sidehaul aren’t interchangeable words for “transport.” They solve different problems. As the RAN evolves, operators may have to design them much more deliberately as parts of the same system.
5. What should operators actually do now?
They should NOT build a 6G network before 6G exists.
The better goal is simpler than that: just don’t build your network today into a corner.
The transport investments operators make during the 5G Advanced cycle will still be in service when early 6G deployments begin. That makes a practical scaling path more valuable than putting maximum capacity at every site before there is traffic to justify it.
A few additional questions to ask yourself that can tell you where flexibility matters:
- Can today’s 10 Gbps cell-site backhaul scale to 20 or 25 Gbps when demand requires it?
- Are microwave, E-band, and fiber being matched to the needs and economics of each route?
- Can capacity grow where it’s needed without forcing upgrades everywhere?
- Can the network adapt as latency, synchronization, and traffic patterns change?
- And, are RAN and transport teams planning those changes together?
That last question may become increasingly important. As the RAN evolves, decisions about radio architecture will have a greater impact on what the transport network needs to deliver.
In other words, transport needs a seat at the RAN planning table, not an invitation after the radio design is finished.
Build for today. Leave room for tomorrow.
The transition from 5G Advanced to 6G won’t happen in one fell swoop of an enormous capacity upgrade.
A baseline 5G macro site may remain comfortably within 10 Gbps. A high-capacity site with substantial new 6G spectrum may move toward 25 Gbps. A major aggregation hub could eventually require 50 Gbps or more. And some new RAN architectures could make latency, synchronization, and traffic topology just as important as raw throughput.
Wireless transport provides a solid path through that transition. The challenge is choosing the right combination of microwave, E-band, and fiber for each part of the network, and making sure those choices can evolve.
So maybe the question isn’t, “Are we ready for 6G?”
That’s too big to be particularly useful.
A better one is:
“Are the transport decisions we’re making today giving us enough options for what comes next?”
The new Omdia white paper, Wireless Transport for the Beyond 5G Era: Backhaul Network Evolution for 5G Advanced and Readiness for Early 6G, commissioned by Ceragon, goes deeper into the capacity models, spectrum assumptions, E-band engineering, emerging RAN architectures, AI Fabric, and the transport decisions operators will face on the road to 6G.
