If you're a small IoT manufacturer trying to choose between Semtech and Cypress, you've probably noticed that most comparisons are either written by marketers or written for engineers with a budget for expensive test equipment. I'm neither. I handle LoRaWAN gateway design and component sourcing for a small industrial IoT company. I've been doing this since 2017, and I've personally made and documented 14 significant mistakes, totaling roughly $27,000 in wasted budget. I now keep a checklist so other people on our team don't repeat them.
Cypress vs Semtech isn't a clean chip-to-chip matchup. Semtech's SX1301 is the LoRaWAN gateway concentrator baseband that powers a huge share of public LoRa gateways. Cypress, which is now part of Infineon, builds wireless MCUs and combo modules like PSoC 6 plus Wi-Fi/BT. They solve different problems. But if you're building an IoT product, you need to decide which side of your board you build around. That's the comparison this article is about.
What I'm Actually Comparing and Why
Here's the honest framework: I'm comparing Semtech and Cypress on five dimensions that have cost me money: silicon fit, design-in difficulty, enclosures, small-order buying, and long-term survivability. I'm not going to compare benchmark numbers from datasheets. Datasheets don't tell you how many mistakes you'll make on the way to a certified product.
Semtech SX1301 is not a complete radio. It's the digital baseband part of a LoRa gateway, and it depends on companion transceivers like SX1255 or SX1257 and a carefully tuned RF front-end. I mention this because one of my first large mistakes came from assuming the SX1301 reference design would work exactly as printed. More on that in a minute.
You'll also see 'Semtech Canada Inc' on invoices or support documents from time to time, because Semtech's LoRa solutions group includes what used to be Sierra Wireless. It confused our purchasing team once because the legal entity didn't match the high-level brand name. That's not a red flag; it's just how large semiconductor companies split legal entities.
Both Semtech and Cypress are blue-chip names in their respective worlds. But a blue-chip name is not a design strategy.
Dimension 1: Silicon fit
If you're building a LoRaWAN gateway, the Semtech SX1301 is almost the default choice. It is designed to demodulate LoRa signals across multiple spreading factors in parallel. In other words, it's made for network infrastructure. That same strength makes it a poor fit for a battery-powered sensor node—it's too hungry and too complex for that.
Cypress/Infineon PSoC 6 and CYW43xxx combo chips are designed for edge devices. They integrate an MCU and Wi-Fi/BLUETOOTH in a way that gets a compact product going quickly. But they don't talk LoRaWAN directly. You'd need an external LoRa radio if you choose Cypress as the main MCU.
The conclusion here: if your product's job is to collect radio packets, Semtech SX1301 is the clear fit. If the product's main job is application processing plus local wireless, Cypress is the more obvious anchor.
Dimension 2: Design-in difficulty
The biggest lesson I've learned in eight years is that 'hard to use' is not the same as 'bad.' Semtech's SX1301 reference designs assume you understand LoRaWAN frequency plans and the RF chain around the chip. In October 2020, I ordered 50 gateway concentrator boards without checking the SAW filter specification against the EU868 frequency plan. It looked fine on my screen. When we tested them, the boards couldn't receive the intended channels properly. 50 boards, $11,400, straight to the trash. That's when I started treating the reference design as law, not suggestion.
For Cypress, the design-in difficulty is different. The toolchain is extensive, but the options are overwhelming. Honestly, I'm not sure why one integrated module can have so many antenna matching variants. My best guess is that they support high-volume consumer products where every dBm is negotiated and tuned. For an industrial 500-piece run, you don't need that level of obsession.
The surprise for me: for a small team with mixed RF experience, Cypress/Infineon got us to a working prototype faster than the discrete LoRa path. The integrated module fit on a board with fewer RF headaches. That surprised me because I assumed the blue-chip name meant harder to control. In practice, it meant fewer unknowns.
Dimension 3: Enclosures and mechanical reality
Almost every comparison I've seen skips enclosures. That's a mistake. The enclosure is where RF design, thermal design, and installation requirements collide.
Semtech SX1301 gateway boards often have vertical RF connectors, metal shields, and sometimes a heatsink on the main processor. Your enclosure needs clearance above all of those. I still kick myself for ordering 100 polycarbonate enclosure halves before checking the height of SX1301 concentrator board components. If I'd done the mechanical pre-check, I'd have caught that the cover pressed against the antenna connector. The rework cost $890 and added a one-week delay to a customer pilot.
Cypress combo modules, in contrast, are often designed for compact, thin products. The radio and controller share one package or small module, which makes a smaller enclosure much easier. You pay for that convenience with lower range compared to LoRa and with a different kind of regulatory complexity.
So the enclosure conclusion is direct: Semtech wins if your product needs outdoor range and multiple nodes. Cypress wins if your enclosure is thin, indoor, and tightly constrained. Cypress vs Semtech on enclosures isn't about who makes a prettier case; it's about how much height and antenna you're willing to accept.
Dimension 4: Buying small and the blue-chip gatekeeping problem
Now let's talk about the part that hurts most small teams: order quantities. Both Semtech and Cypress are blue-chip semiconductor suppliers. Neither one has a direct sales desk for a 100-piece prototype run. But that doesn't mean small orders are impossible.
I've bought Semtech SX1301 parts through distribution with no serious issue. The meaningful gate is finding a distributor's FAE who replies to a small customer. Infineon (Cypress) also sells through distribution, and you can get parts in small quantities if you accept standard lead times and packaging limits.
My personal rule is simple: When I was starting out, the vendors who treated my $200 orders seriously are the ones I still use for $20,000 orders. Small doesn't mean unimportant—it means potential. If a supplier acts like your 50-piece order is a nuisance, walk away before you give them the 500-piece order.
One practical note: ask about forecast requirements early. As of early 2025, I still see distributors willing to support small orders when the buyer brings a six-month forecast, even if the forecast is modest. Without it, you're just a stranger asking for a favor.
Dimension 5: Roadmap and survivability
Semtech and Cypress have different ecosystem risks. Semtech's LoRa business is closely tied to LoRaWAN and the spread of public/private IoT networks. Cypress is part of Infineon, which gives it a large automotive and industrial footprint. But Infineon's focus is not LoRa, so don't expect deep LoRaWAN integration from that side.
I learned this in 2020: vendor portfolios can change, and a part that looks safe can get a not-for-new-design notice sooner than you expect. I've never fully understood the roadmapping logic for IoT communication chips. My best guess is that large suppliers follow big customers, not small ecosystems. That's another reason I advise small teams to standardize on an ecosystem you can survive, not just a chip you can buy.
This was accurate as of Q1 2025. Both companies change their product portfolios quickly, so verify current lifecycle status with your distributor before you commit.
Final Scenario Recommendations
If you've read this far, you know I'm not going to give you a clean 'Semtech wins' or 'Cypress wins.' If your product is a LoRaWAN gateway or network device, build around the Semtech SX1301 and do the RF and enclosure homework I skipped. If your product is an edge sensor with Wi-Fi/BLUETOOTH and no need for LoRa, the Cypress/Infineon PSoC 6 path will likely save you time and get you into a smaller enclosure.
If you're still torn, buy both evaluation boards. Put them next to each other on the bench, measure power, range, and integration difficulty. In my experience, the board that gets you to a reliable pilot first is the board that will get funded for production.
Cypress vs Semtech is not a marketing battle. It's a question of which radio path, enclosure, and supply chain best fits your product. The wrong answer is almost never about brand; it's about ignoring the first-order constraints of your design.
