Proprietary radios and closed ecosystems

Closed is not one thing#

"Proprietary" covers four quite different situations, and they carry very different risk:

The useful axis is not open versus closed. It is what survives the vendor. That is the same question behind when a smart home company shuts down, and it is worth pricing before you buy, not after.

Lutron: the case for a closed radio#

Lutron runs Clear Connect, its own RF protocol, on a sub-GHz frequency in North America rather than in the 2.4 GHz band. That alone puts it outside every problem described in 2.4 GHz interference and channel planning: no Wi-Fi contention, no microwave, no channel plan.

What you get for accepting a closed radio:

• Dimming that behaves. Lutron's core business is dimming, and its dimmers handle awkward LED loads with fewer complaints than most standards-based competitors.
• Pico remotes. A battery wall remote that mounts anywhere, needs no wiring, has a very long battery life, and is the single most practical answer for renters and for rooms with no switch where you want one. Smart home for renters leans on exactly this pattern.
• A network you do not have to design. No coordinator placement, no mesh routing, no channel choice, no re-pairing when you add a device.
• Integration through the hub. Lutron's bridge exposes the system to the major platforms, so the closed radio does not mean a closed system at the automation layer.

What you give up: device selection outside lighting and shading, price, and the ability to mix in a cheaper switch from another vendor. Everything must be Lutron, and Lutron is not cheap.

That is a coherent trade if you are wiring a whole house of lighting and shading and want it to be boring for fifteen years. It is a poor trade if you want one smart switch in a hallway. Before buying, check the current state of Lutron's Matter and platform support for the specific bridge model you are considering, since bridging capability differs across their product lines and changes with firmware.

Philips Hue: standard radio, vendor network#

Hue is Zigbee. That surprises people who assume it is proprietary, and it changes the risk calculation substantially: the bulbs are speaking a standard protocol you can theoretically point at another coordinator.

In practice Hue is a vendor-controlled Zigbee network with its own bridge, its own local REST API, and its own app. Signify has added Matter bridging so a Hue Bridge can expose its lights to a Matter fabric, which means the lights appear in Apple Home, Google Home, Alexa and SmartThings without a cloud account in the path.

Three practical things to know:

1. You can leave, at a cost. Hue bulbs can be reset and joined to a generic Zigbee coordinator such as one running Zigbee2MQTT or ZHA. You lose the Hue app, Hue Entertainment sync, and some of the effects, and you take on the maintenance yourself. But the hardware is not bricked, which is more than most closed systems offer.
2. Bridging costs features. Reaching Hue lights through the Matter bridge is not the same as talking to them natively. Effects, scenes stored on the bridge, dynamic scenes and gradient behavior commonly do not survive the translation. Matter bridges and what you lose through one covers what generally gets dropped, and smart lighting that stays compatible covers the color capability loss specifically.
3. The app has drifted toward accounts and cloud features. The local API still exists, which is what matters for reliability, but treat the direction of travel as a factor in a long-term buying decision.

Hue is the middle of the spectrum: closed enough to be pleasant, open enough to escape. For most people that is a reasonable place to be.

Tuya: the white label under a thousand brands#

Tuya is not a brand you buy. It is a platform that thousands of brands build on, so the same Wi-Fi plug appears under dozens of names with different logos and the same firmware, the same app architecture and the same cloud.

The upside is real: it is why smart devices are cheap, why a niche product exists at all, and why an obscure brand can ship a working app.

The downsides are structural:

• Cloud dependency by default. Many Tuya-based Wi-Fi devices route control through Tuya's cloud even when your phone is on the same network as the device, which costs latency and breaks in an outage. See local vs cloud control.
• Local control is possible but fragile. Local integrations exist, but they typically require extracting a per-device local key and they break when firmware changes.
• Reflashing has gotten harder. Older Tuya Wi-Fi modules could be reflashed with open firmware over the air. Newer firmware and hardware revisions closed that path.
• Tuya Zigbee beats Tuya Wi-Fi. Tuya-based Zigbee devices join a standard coordinator and work locally, though they are notorious for non-standard clusters that need per-device quirk handling in Zigbee2MQTT and ZHA.

The rule: if you are buying a Tuya-platform device, buy the Zigbee version, not the Wi-Fi version. And if the only version is Wi-Fi and cloud-bound, price it as disposable.

433 MHz: cheap, dumb, and one-way#

The 433 MHz band is full of legacy smart home gear: door sensors, PIR sensors, remote-controlled outlets, weather stations, some alarm panels. It is sub-GHz, so range is decent and walls are less of an issue.

Everything else about it is a warning:

• Usually one way. Nothing confirms anything. Your hub does not know the light turned on, only that a command was sent.
• Usually unencrypted. Fixed-code devices can be recorded and replayed with a cheap radio. Never use 433 MHz for anything that grants access.
• No network. No mesh, no routing, no diagnostics. If it does not reach, it does not reach.
• Bridged, not integrated. Support for any given device depends on someone having reverse engineered its encoding.

There is one legitimate use: reading existing sensors you already own, or bringing an old alarm system's sensors into a modern platform through a receiver. That is genuinely useful and covered in upgrading an old smart home. Building anything new on 433 MHz is a mistake when Zigbee, Thread and Z-Wave all offer acknowledged, encrypted, two-way communication for similar money.

Insteon: the cautionary tale#

Insteon was a genuinely clever system: dual-band, sending commands over both the house wiring and RF, so a message that could not get through one path could get through the other, with every device repeating for every other. For a decade it was a credible alternative to Z-Wave in the US.

In April 2022 the company behind it shut down abruptly. The cloud servers went offline, hubs stopped working for anything that touched the service, and customers found out from their devices rather than from an announcement. The brand was later acquired, service was restored under new ownership, and Insteon products are sold again today. The lesson does not go away:

The test to apply before buying#

The exit cost calculator puts numbers against that, and the pre-purchase checklist turns it into questions to answer before checkout.

Where closed still beats open in 2026#

Being fair to the walled gardens, there are real gaps:

• High-end shading and whole-house lighting control. Standards-based blinds have improved, but coordinated multi-shade control with reliable position feedback is still a place where a single-vendor system tends to work better. See blinds, shades and garage doors.
• No-neutral dimming at scale. US wiring often lacks a neutral in the switch box, and vendors with their own hardware and radio have solved this more consistently. Smart switches, dimmers and the neutral wire covers the options.
• Cameras. Matter cameras arrived in Matter 1.5, but the shipping support is thin, so camera choice is still largely a vendor choice.
• Anything needing tight, low-latency scene coordination across dozens of fixtures, where a system designed as one product beats a fabric assembled from six vendors.

The pattern in all four: the closed system wins where the open standards have not finished the job. As they finish, the case narrows. That is what twenty years of smart home standards shows happening, wave after wave.