Latest / The 5G Podcast with Fexingo: Wireless Networks, Carriers, and Mobile Infrastructure / Why 5G Network Energy Efficiency Matters Now
Transcript
- Lucas: There's a number in wireless that doesn't get enough airtime: energy per gigabyte. And the story behind it is actually reshaping how carriers build out 5G. Luna: I've seen headlines about 5G being more efficient than 4G, but then I also hear that carriers' electricity bills are going up. Which one is it? Lucas: Both, surprisingly. On a per-bit basis, 5G can be up to 90 percent more energy efficient than 4G. That's because the radios are more advanced — they can power down components when idle, beam-form more precisely, and use wider spectrum channels to move data faster, so the transmission time is shorter. Luna: Right, so each megabyte uses less juice. But the total amount of data is exploding — way more cells, way more traffic. Lucas: Exactly. The network is denser. You've got small cells on lampposts, massive MIMO antennas on macro towers, and all the supporting fiber and compute. So even though each bit costs less energy, the total energy consumption of the radio access network has been climbing. Some European operators have seen their RAN energy use rise by 15 to 20 percent year-over-year since 2020. Luna: And that directly hits profitability. Energy is a huge opex line for carriers — some spend 3 to 5 percent of revenue on power alone. Lucas: Right. So the industry is now treating energy efficiency as a core KPI — not just a sustainability talking point. And there's one case I think is worth zooming into: Vodafone's work on intelligent sleep modes and ai driven cooling across its European footprint. Luna: I remember reading about that. They claimed they cut energy per gigabyte in half over three years. How did they actually pull that off? Lucas: A few layers. First, the radios themselves — newer massive MIMO units can enter a deep sleep state when traffic is near zero, like overnight in residential areas. Older 4G radios didn't have that granularity. Second, they deployed software that predicts traffic patterns and dynamically powers up only the sectors needed. And third, they used AI to optimize cooling in base station cabinets. Cooling can be 30 percent of a site's power draw. Luna: So the AI basically says, 'It's two in the morning in this suburb, let's turn off three of the eight antenna panels and dial back the fans.' Lucas: Exactly. Vodafone rolled that out across 12 markets and reported a 50 percent reduction in energy per gigabyte between 2021 and 2024. Their total energy still went up a bit because traffic grew even faster, but the efficiency curve bent sharply. Luna: And that's not just green PR — that's real money. For a carrier with a billion-dollar energy bill, a 10 percent reduction is nine figures. Lucas: If today's conversation gave you something useful, the way these episodes stay ad-free is through listener support. It's buy me a coffee dot com slash fexingo — and that genuinely lets us keep digging into specifics without interruption. Luna: Yeah, it's a small thing that makes a big difference. Back to the numbers — because I want to understand the unit economics better. Lucas: Sure. So think of it this way: a typical macro cell site with 4G and 5G running simultaneously might draw 5 to 7 kilowatts at peak. With efficient sleep modes, the average can drop to 3 to 4 kilowatts. Multiply that by hundreds of thousands of sites, and the savings are enormous. Luna: But aren't carriers also adding more sites? Small cells are popping up everywhere — those have to add to the total, even if each one is low power. Lucas: They do. A small cell draws maybe 50 to 100 watts, which is tiny compared to a macro site, but you might need 10 of them to cover the same area. The density trade-off is real. The key is that each small cell serves fewer users, so the energy per bit can still be lower if the traffic is there. Luna: So the math only works if the small cells are actually used — not just sitting idle. Lucas: Right. That's why intelligent deployment matters. Some operators are now using traffic analytics to decide exactly where to put small cells, rather than just blanketing a city. T-Mobile in the US, for example, has been very disciplined about deploying mid-band small cells only in high-traffic corridors. Luna: And Verizon? They went big on millimeter wave small cells early on, which are dense and power-hungry. How does that affect their efficiency story? Lucas: Millimeter wave is a different beast. The radios are more power-hungry per unit, and the coverage radius is tiny, so you need a lot of them. But Verizon has said that in high-traffic venues like stadiums, the efficiency per bit can still be good because the capacity is enormous. The problem is when those small cells sit underutilized — then the energy per bit spikes. Luna: That's the tension: building for peak demand versus building for average load. And the energy efficiency metric changes completely depending on which lens you use. Lucas: Exactly. That's why the GSMA and some operators are pushing for a standardized energy efficiency metric — something like 'bits per joule' — so investors and regulators can compare operators fairly. Right now, everyone defines it differently. Luna: Bits per joule — I like that. So a carrier that uses 10 kilowatts to deliver 100 gigabits per second has 10 million bits per joule. Clear. Lucas: Right. And once you have that metric, it becomes a competitive lever. An operator with better bits per joule can offer lower latency and higher throughput at the same opex, or lower prices. It's not just a green badge. Luna: It's also becoming a factor in vendor selection. Ericsson and Nokia both publish energy efficiency specs for their massive MIMO radios now — they know carriers are comparing those numbers. Lucas: Absolutely. Ericsson's latest AIR 6471, for instance, claims 40 percent lower energy consumption than the previous generation for the same throughput. That's a huge selling point. Nokia has similar claims on their ReefShark platform. Luna: And what about software? We talked about AI sleep modes — is there more the software layer can do? Lucas: Yes. The next frontier is what's called 'energy-aware network slicing'. Imagine a slice for autonomous vehicles that requires ultra-low latency and high reliability — that slice might need full power on all radios. But a slice for smart meters only needs a few kbps once an hour. The network could allocate energy budgets per slice, putting low-priority slices into deeper sleep. Luna: That's smart — dynamic energy allocation based on service level agreements. It's like a power grid for data. Lucas: Exactly. And this matters for the net-zero commitments carriers have made. BT, Vodafone, Telefonica — they all have targets to be net-zero by 2040 or sooner. RAN energy is 60 to 80 percent of their total energy footprint, so they have to tackle this. Luna: So the efficiency gains aren't just about cost — they're about regulatory compliance and brand reputation too. Lucas: Right. And the interesting thing is that many of the efficiency improvements also improve performance. Massive MIMO with better beamforming gives better signal quality, which means fewer retransmissions, which saves energy. It's a virtuous cycle. Luna: So what's the one thing you'd want listeners to remember about 5G energy efficiency? Lucas: That the technology is genuinely more efficient per bit, but the network is so much bigger that total energy is still rising. The winners will be the operators who can bend the total energy curve downward while traffic keeps growing. That's the real test. Luna: And that test is happening right now, as carriers roll out standalone 5G and even more small cells. This isn't a future problem — it's today's operating reality. Lucas: Exactly. Next time, we'll look at how energy efficiency intersects with network slicing — because the two are going to converge in some surprising ways. Luna: Looking forward to it. Thanks for listening.