I’m one of those philosophical guys that think most insights about the future are actually embedded in the past. So, earlier this week, I was thinking back to one of my first Wi-Fi surveys, which took place in a sausage factory. Yes, an actual sausage factory. I remember looking in a ceiling cavity to find a Proxim 802.11 (not b, g, n, ac, or ax… just 802.11) access point (AP) that supported a whopping 2 Mbps. It was an ancient AP, even at that time. I have several memories of the project (particularly the awful smell). Still, the one memory that seems relevant right now was installing 30-something APs in this facility but only removing 9 APs.
Why is that important? Because it reminds me that the early days of enterprise Wi-Fi focused on hotspots—in specific offices, conference rooms, or spot locations in non-carpeted space. You put APs only where you needed them because they were expensive and unproven, driven by specific connectivity needs that were still emerging. And those connectivity needs for Wi-Fi went through an evolution, from hotspots for specialty non-carpeted services to ubiquitous enterprise coverage; then Wi-Fi was all about BYOD; then Ethernet replacement and high density for data capacity.
This lens of market evolution is how we should think about private 5G today—yes, of course, cellular is much more mature broadly speaking, but we are just at the beginning of an adoption curve for indoor services beyond voice. For many private cellular cases today, enterprises may only need 4G/5G hotspots of service. These hotspots will serve specialized devices in outdoor areas or subsets of non-carpeted indoor space. For other cases (like stadiums or venues), 5G is needed for extra data density. Other organizations are replacing a subset of deterministic wired links with 5G in industrial settings, simply to benefit mobility. So, the enterprise cases are varied, but they’re early and patchy.
Meanwhile, the telecom industry as a whole has evolved. Most of the long cellular history revolves around voice, with a huge focus on macro network coverage. Then the focus was on services like messaging and 3G data, followed by challenges with declining average revenue per user (ARPU), defense of voice/text services against so-called over the top (OTT) apps, and now densification with 5G for more data capacity—and hopefully new use cases and revenue streams. Around all this, the technology ecosystem (IoT, AI, cloud, mobile) has shifted while the network pipe is more commoditized, which creates the hunt (desperation?) for new use cases and customers. Currently, those hopes ride in the sidecar of 5G because it is new and flashy.
All this leads to a kind of industry noise that has stirred enterprises into asking, “what should we do with 5G?”
So, down to brass tacks. If you come away from this entire blog series with one takeaway, consider this: private LTE/5G is a wireless option for a specific set of connectivity problems. Put private LTE/5G in a group of tools alongside Wi-Fi, BLE, Zigbee, LoRa, and other LP-WAN options, public 4G/5G, satellite, NFC, and yes, Ethernet and fiber.
There are many choices and many dimensions of consideration for each option, as illustrated in Figure 1. Marketing messages are built around one or two dimensions, like range or bandwidth while neglecting other dimensions of importance—including complexity and cost to buy, deploy, and operate, which are implicit in Figure 1.
Figure 1: Dimensions of Consideration for Wireless Solutions (Figure courtesy of Disruptive Analysis)
If you’re still not quite seeing 5G as one tool among many, consider a couple of final points.
Cellular chips are an order of magnitude more expensive than other options, like Wi-Fi, BLE, and Zigbee. According to Gartner, the average selling price for Wi-Fi chips today is around $2.75 (USD), compared to $24 for a 5G baseband. As a consumer, $24 might not seem like a significant cost, but it’s enormous. Some other components/costs go along with that $24 too. Think about device manufacturers trying to build a $50 (or even $300) connected widget. I’ve seen heated cost control debates over $0.13 components, so a $24 cost adder is enormous—it completely changes the price category for a device.
Also, keep in mind that the global spectrum is not harmonized. So cellular also introduces a layer of complexity in deciding which radio bands to support (sub-6GHz, mmWave, etc.), whether to build region-specific product SKUs or global product SKUs. You pay a cost and complexity toll on either side of that decision, in the form of hard product costs or operational costs. Wi-Fi will have the same issue with 6 GHz support, but most vendors already have non-Wi-Fi 6E options in their portfolio to fill in.
Then, consider adoption. Given pricing and demand, I haven’t seen any strong evidence showing a shift in preference from Wi-Fi/BLE to cellular. If anything, it seems the opposite for many device types. In devices like IoT, wearables, laptops, and tablets (where there’s tremendous consumerization in enterprise IT), cellular adoption is very low. There is more traction in other areas, like equipment in non-carpeted space (robots, AGVs, healthcare equipment) and mobile handheld platforms (point-of-sale, scanners, voice). However, adoption by vendors and enterprises still seems nascent, often for cost and complexity reasons.
Chipmakers are just now starting to announce combo Wi-Fi/IoT chips designed to integrate 5G modems, which means the adoption trickles down to manufacturers over the next year or more after availability, and then to customers after that, where there’s appetite.
And then some of the slow adoption simply has to do with lifecycle, which is my final point. Most Wi-Fi devices have at least a 5-year lifecycle. Specialty equipment in manufacturing, operations, healthcare, retail, and industrial environments often has far longer lifecycles, often greater than 10 years—there’s a reason 802.11b and WEP/TKIP still exist. Everyone rags on Wi-Fi security compared to cellular, but enterprises are often reluctant to invest in even modern Wi-Fi security solutions when it impacts device refresh cycles. The problem is not technology availability. Instead, it’s a willingness to spend.