I'm an IoT product engineer. For the past six years, I've handled component selection, board bring-up, and supplier calls for small industrial wireless projects. I've made and documented eleven significant mistakes, totaling roughly $46,000 in wasted budget. Some of that waste came from bad antenna assumptions. Some came from ordering during the chip shortage. Most of it came from treating LoRaWAN as a simple component purchase instead of a system design decision.
My opinion is simple: small orders are not the enemy of Semtech LoRaWAN. The enemy is a procurement mindset that only respects volume. That mindset hurts real projects, and the chip shortage made it worse.
What “semtech lorawan” actually means
If you search for the phrase semtech lorawan, you'll see modules, development kits, and gateway products. It sounds like one thing. It isn't. LoRa is the radio technology in Semtech chips. LoRaWAN is the open network protocol managed by the LoRa Alliance. Semtech supplies transceivers like the SX1262 and SX1276, but the protocol layer is not locked inside one chip supplier. A module that contains a Semtech radio can be sold by multiple authorized module vendors. That detail saved one of my projects.
Why did I almost miss that? Because during the chip shortage, I treated a long lead time as a dead end. In early 2022, I was waiting on a small batch of SX1276 modules and the factory date kept slipping. The solution was not magic inventory. It was an approved module built around a Semtech radio and certified for the region. From the network side, it was still semtech lorawan. From the physical layout side, I still had to review the changes. But the project didn't stop.
What most people don't realize is that the LoRaWAN regional rules and certification stack are as important as the chip. The radio may be able to transmit, but if it doesn't comply with regional duty cycles and power limits, you have a product that only works in your lab. The LoRaWAN specification is public through the LoRa Alliance, and Semtech publishes detailed transceiver reference designs. I keep telling engineers to read those two sources before complaining about range.
The chip shortage exposed a blue chip bias
Large suppliers had hard allocation decisions during the shortage. If you have enough annual purchasing power to get true blue chip treatment, this article may not apply to you. Most engineers I know are not in that position. We buy in hundreds, not hundreds of thousands.
Still, I understand why high minimum order quantities exist. A distributor that moves a million parts per month cannot spend twenty minutes on every 100-piece order. But the customer with 100 pieces is often the same customer that will eventually move a million parts. The companies who made me feel like a real customer on small orders got the larger orders later.
I remember one quote from a distributor during the shortage: "We can take 80 modules but not 500. If that works for you, they're yours." I took them. That 80-piece order was enough to keep the validation program moving while bigger orders went to companies with more purchasing power. Small didn't mean unimportant. It meant I was still in the game.
The phrase “semtech antennas” hides a real mistake
Let's clear up a wording issue. When someone searches semtech antennas, they are usually looking for guidance about antennas for LoRa modules. Semtech doesn't necessarily sell its own antenna catalog. More importantly, the company's reference designs tell you how the antenna should be matched to the transceiver and what the ground plane needs to look like. That's not a minor detail.
I once approved a 200-piece order using a module with a connector for an external antenna. The antenna had the right connector and roughly the right frequency, but the matching network and ground plane didn't follow the radio reference design. The range test was embarrassingly short. We spent about $4,600 on rework and lost two weeks. The fix wasn't a stronger radio. It was a properly matched antenna and a layout that followed the published reference.
Now I search for semtech antennas when I'm in the design phase, and I also check whether the module is LoRaWAN certified or just a bare RF component. The difference changes what you need to test. Antenna selection is more than a catalog choice. It's a field reliability decision.
Ask what is on my WiFi—even when LoRaWAN is the star
LoRaWAN is not Wi-Fi. That's usually a benefit. But a gateway still needs network backhaul, and the equipment beside it is often Wi-Fi. The question what is on my WiFi is not just a home user's paranoia. It belongs in your deployment checklist.
Here's what happened to me in 2023. A site test was going well. The gateway was online, the server was receiving LoRaWAN packets, and the LTE router was stable. I still had an uncomfortable feeling, so I logged into the router and asked: what is on my WiFi? There was a diagnostic node I had connected earlier that week. It was on the same subnet as the gateway management interface, and it still had the factory default password. The risk was not someone jamming LoRa. The risk was someone entering the network through that forgotten device.
So when I train teammates now, I tell them to audit the network around the network. Identify every wired connection, every access point, every forgotten test node. The strongest LoRaWAN link in the world cannot protect you from a default password on the backhaul side.
The objection I keep hearing
An experienced procurement manager might say: "You can't compare a 200-piece order with a 200,000-piece order. Allocation has to favor the bigger customer." I agree. I am not asking suppliers to ignore volume. I am asking them to see what a small LoRaWAN order represents. It represents a design win that might grow. It represents an engineer who will remember how the vendor treated the first attempt.
I don't know when chip shortages will fully disappear. My best guess is that component allocation will never feel as easy as it did before 2020. The solution is not to wait for that old feeling. It is to become good at working within allocation, and to choose partners who can support a thoughtfully small order.
Let me add a limitation. My experience is based on about 200 industrial IoT projects and orders, mostly between 50 and 1,000 devices, in North America and Europe. I have not managed a mass-scale utility deployment. If you are building millions of smart meters, your priorities around blue chip suppliers and allocation will be different. My opinions should be treated accordingly.
What I believe now
I'll end where I started. The chip shortage did not make Semtech LoRaWAN fragile. It made the ecosystem around it more visible. Small orders are not the problem. Ignoring antennas, ignoring the network around the gateway, and treating every purchase like a blue chip allocation exercise is the problem. If you respect the journey from 100 nodes to 10,000 nodes, you will build designs—and supplier relationships—that survive the next shortage.
