We build on standard Pi hardware. You skip the industrial markup.
Most "industrial Pi" vendors re-badge a reference board and mark it up 3–4×. We build on standard Raspberry Pi and Compute Module 5 — the same parts you could buy yourself — so you skip that premium and pay for the engineered system, not the badge.
What the hardware costs — before engineering
Indicative street price for a single comparable control node. Lower is better — a standard CM5 build sits at the top. Our engineering fee is quoted separately.
| Vendor | Build / model | CPU | On-board I/O | Edge AI | Real-time | Temp range | Street price |
|---|---|---|---|---|---|---|---|
| ◢ Hyleon | Standard CM5 build | CM5 quad-core | per build | Hailo-8 optional | MCU co-proc | −25→+60 °C | $80–900 |
| Revolution Pi | Connect 5 | CM5 | expansion | add-on | PREEMPT_RT | −40→+55 °C | $489 |
| Seeed | EdgeBox RPi 200 | CM4 | 4 DI · 4 relay | add-on | none | −20→+60 °C | $419 |
| Siemens | S7-1200 + HMI | proprietary | 14 DI / 10 DO | none | native | 0→+55 °C | $920 |
| Allen-Bradley | CompactLogix 5380 | proprietary | modular | none | native | 0→+60 °C | $1,840 |
Indicative single-unit street pricing, Q2 2026 — hardware only, engineering quoted separately. PLC rows exclude programming-software licenses (typically $1k–6k more).
Cheaper because of where we start — not what we cut
No industrial markup
We start from standard Raspberry Pi and Compute Module 5 modules — not a re-badged 'industrial' board sold at a 3–4× premium.
Open by default
Mainline Linux, off-the-shelf parts. No proprietary I/O tax, no per-seat license to program it.
You pay for the system
The hardware is a commodity you could buy yourself. What you're paying us for is the engineering that turns it into a control system for your line.
The platform we build on
A representative build on the Compute Module 5 — the industrial-grade, AI-ready baseline we start from and tailor to each line. Specs vary by build.
We scale the build to the job
Same OS, same toolchain, same security model — we scale the I/O and compute to the job.
Control node
Control plus optional on-device AI for the typical line.