Latest / The 5G Podcast with Fexingo: Wireless Networks, Carriers, and Mobile Infrastructure / Why Millimeter Wave 5G Is Making a Quiet Comeback
Transcript
- Lucas: If you followed the early hype around 5G, you probably remember the phrase 'millimeter wave' being thrown around like it was the future of everything. And then, pretty quickly, the narrative flipped — it became the poster child of overpromise. Luna: Right — it's the spectrum that can carry massive amounts of data but can't go through a wall, or a tree, or even heavy rain. For a while it felt like carriers were quietly walking away from it. Lucas: Exactly. But that's not actually what happened. Millimeter wave — the 24 gigahertz to 40 gigahertz range — is making a real comeback, not as a blanket coverage solution, but in targeted deployments where its weaknesses become strengths. And the numbers are starting to back that up. Luna: So what changed? Is it the technology itself, or the way operators are thinking about using it? Lucas: Both. Let's start with a concrete example. Take a major NFL stadium — we'll use one that's been documented pretty well by the operators involved. On game day, you've got 70,000 people in a concentrated space, all trying to stream video, post to social media, maybe check fantasy stats. A single mid-band tower covering the area would collapse under that load. Luna: That's exactly where millimeter wave shines — massive capacity in a small area. Lucas: Right. In this particular stadium, they deployed about 60 millimeter wave nodes — small radios, each roughly the size of a shoe box. Those nodes are connected to a fiber backbone, and during a game they've measured peak throughput exceeding 50 gigabits per second. That's roughly 15 times what a typical mid-band cell site can handle. Luna: And that's not theoretical — that's measured under real conditions with tens of thousands of devices? Lucas: That's from operator data during actual games. Now, the key insight is that 60 nodes in a stadium works because the geometry is known — you can place them along the rafters, point them down at specific sections, and the line of sight constraints that make millimeter wave a problem in a city become a feature in a controlled environment. Luna: So the comeback is really about finding the right use cases, not trying to blanket entire cities. Lucas: Exactly. And that's been validated beyond stadiums. Airports are a similar story — dense crowds, predictable paths, high data demand. But the really interesting growth is showing up in industrial settings. Factories and warehouses are deploying private millimeter wave networks for things like real-time video analytics on assembly lines, or controlling autonomous robots with latency under one millisecond. Luna: And those private networks don't have to worry about interference from other carriers, which makes the propagation issues more manageable. Lucas: Right. You control the environment. You can put repeaters where needed. And the cost has come down dramatically. In 2020, a single millimeter wave radio node cost around $1,500. Today, that same class of radio is under $400 — driven partly by scale, partly by chipset improvements from companies like Qualcomm. Luna: Four hundred dollars is still more than a Wi-Fi access point, but for the bandwidth you get, it starts to make sense in certain verticals. Lucas: Exactly. And that cost decline is opening up a second use case that a lot of people wrote off: fixed wireless access in dense urban corridors. Think apartment buildings in cities like New York or San Francisco, where fiber-to-the-home is expensive to install. A millimeter wave node on a rooftop can beam 2 to 4 gigabits per second to receivers on the side of buildings, without trenching fiber through city streets. Luna: Verizon was pushing that hard a few years ago with their 5G Home product. I remember they had some early struggles with line of sight and tree coverage. Lucas: They did. And they scaled back their marketing for a while. But the technology has improved — beamforming and beam tracking are much better now. The radios can lock onto a user's device even if it moves around a room, and self-install receivers with phased arrays are more reliable. Verizon actually added about 200,000 fixed wireless subscribers in Q1 of this year, and a meaningful portion of those are on millimeter wave. Luna: That's interesting because T-Mobile has been the loudest about fixed wireless, but they're mostly using mid-band. Verizon is leaning back into millimeter wave for the speed advantage. Lucas: Right. And that's a strategic choice. Mid-band gives you a solid 200 to 300 megabits per second, which is plenty for most households. Millimeter wave can give you over a gigabit. If you're trying to compete with cable companies who are offering one-gig plans, having a product that can actually deliver those speeds matters for certain customer segments. Luna: Let's talk about the spectrum itself. The FCC auctioned a lot of millimeter wave spectrum back in 2019 and 2020. A lot of that went unsold or at very low prices. Are carriers actually using what they bought? Lucas: That's a great question. In the 2019 auction, the 24 gigahertz and 28 gigahertz bands saw some bidding, but the 37, 39, and 47 gigahertz auctions had much less interest. Verizon and AT&T did buy significant blocks in the 39 gigahertz band. And according to deployment data from the carriers, Verizon has about 30,000 millimeter wave nodes installed as of late last year — concentrated in stadiums, airports, and dense downtown corridors in about 60 cities. Luna: That's not nothing. But it's a fraction of their total cell sites, right? They have hundreds of thousands of sites overall. Lucas: Absolutely. Millimeter wave will never be a coverage layer. It's a capacity layer. And the industry seems to have accepted that. The shift in narrative from '5G will blanket everything with millimeter wave' to 'millimeter wave is a tool for specific high-density scenarios' has been healthy. It lets the technology be judged on its actual merits. Luna: I want to push back a little on the cost narrative. You said radios dropped from $1,500 to $400. But the backhaul requirements for millimeter wave are intense — every node needs fiber, and that's expensive. Does the total cost of ownership still work? Lucas: It depends on the revenue opportunity. In a stadium, where you can sell premium connectivity or charge for network slicing — like giving a broadcaster a dedicated slice for live 4K video — the economics work. In a factory, where millimeter wave enables automation that saves millions in labor or defect costs, it's a no-brainer. For a residential fixed wireless deployment, the math is tighter, but if you avoid trenching fiber at $50,000 per mile, the node cost becomes secondary. Luna: So the future of millimeter wave isn't about coverage — it's about niches with high value per square foot? Lucas: That's exactly how I'd frame it. And I think we're going to see more of those niches emerge as the chipsets get cheaper and the beamforming algorithms get smarter. One area I'm watching is in-venue augmented reality — imagine pointing your phone at a player on the field and seeing their real-time stats overlay. That needs sub-10-millisecond latency and high bandwidth, which only millimeter wave can deliver at scale. Luna: If today's episode was useful to you and you want to keep it ad-free, buy me a coffee dot com slash fexingo helps. Lucas: Yeah, it literally does. A handful of listeners chipping in is what keeps us from having to run ads or put this behind a paywall. Luna: So if you're in a position to do that, it makes a real difference. Now — back to millimeter wave. Lucas, you mentioned network slicing a moment ago. How does that play into the comeback? Lucas: Network slicing is the software layer that makes millimeter wave even more valuable. With slicing, an operator can carve out a virtual network with guaranteed bandwidth and latency for a specific customer. So in that stadium example, you could sell a slice to the league for in-game official video distribution, another slice to a streaming partner for fan-facing AR, and a third slice for general public data. All on the same physical millimeter wave nodes. Luna: And the millimeter wave spectrum's short range means you can reuse those slices in different sections of the stadium without interference, effectively multiplying the capacity. Lucas: Exactly. The same 60-node deployment can serve multiple revenue streams. That's what changes the economic equation — it's not just about selling data plans to fans. It's about enterprise-grade services with service-level agreements. The potential revenue per square foot in that scenario is orders of magnitude higher than a typical cell site. Luna: Are there any major obstacles still? I mean, the rain fade issue — does that still plague millimeter wave? Lucas: Rain fade is real, but it's been well characterized. At 28 gigahertz, heavy rain can cause attenuation of about 10 to 20 decibels per kilometer. In practice, that means a link that normally works at 2 kilometers might drop to 1.5 kilometers in a downpour. For most deployments — stadiums, factories, even fixed wireless in urban areas — the link distances are short enough that you can budget for it. The bigger issue is foliage, like dense tree canopies. That's harder to design around. Luna: So you need clear line of sight, or at least a Fresnel zone free of obstructions. Lucas: Right. And that's why millimeter wave works best in environments where you control the deployment geometry. Indoors, obviously, that's easier. Outdoors, it means mounting radios on street furniture or building facades at strategic heights. The carriers have gotten much better at using propagation modeling software to optimize node placement before they ever install a radio. Luna: What about the device ecosystem? Early on, only flagship phones had millimeter wave antennas. Is that still the case? Lucas: It's improved. In the US, all three major carriers require millimeter wave support in their premium phones, and it's trickling down to mid-range devices. Qualcomm's Snapdragon 8-series chips have integrated millimeter wave since 2021, and now the 7-series — used in phones around $400 — includes it. The antenna modules are smaller and more power-efficient. So the device bottleneck is easing. Luna: That's a critical enabler. If only 10 percent of phones in a stadium support millimeter wave, the capacity advantages are limited. But if that number hits 50 or 60 percent, the network really starts to perform. Lucas: Exactly. And we're seeing that shift happen. At the most recent Super Bowl, carriers reported that millimeter wave carried over 65 percent of all mobile data traffic in the stadium, despite being available only on a subset of devices. That's because the users who had it got such a better experience that they consumed more data. Luna: So the usage is concentrated among the users who can actually connect — pulling them off the mid-band and freeing up capacity for everyone else. Lucas: That's the beautiful network effect. A few high-capacity nodes serving a fraction of users can dramatically improve the experience for the whole crowd. And that's really the millimeter wave story now — not a magic bullet, but a precision tool that, when deployed smartly, makes the entire network better. Luna: What's the single biggest thing to watch in the next 18 months? Lucas: I'd say the integration of millimeter wave with ai driven network optimization. Right now, operators still rely on manual planning for node placement and beam steering. We're starting to see machine learning models that can predict traffic patterns and adjust beam directions in real-time — not just based on device location, but on predicted demand. That could squeeze another 30 to 40 percent capacity out of the same hardware. Luna: So the comeback isn't just about cheaper radios — it's about smarter software making the physics less limiting. Lucas: Exactly. And that's why I think millimeter wave is finally hitting its stride. It took longer than the hype cycle promised. But the deployment numbers, the cost curve, and the user experience data all point in the same direction: millimeter wave is becoming a practical, profitable part of the 5G mix.