How to develop hardware if you don’t develop hardware?

This might sound contradictory, but plenty of software or AI companies need hardware. How does data get into the software? How does the data get out to users? There are multiple paths to get to hardware solutions each with their own development cost, hardware cost and schedule implications. (time to first lot in parenthesis)

Buy it off the shelf (immediate)

Using ruggedized Samsung or Apple tablets have replaced cash registers and even controls for medical devices. Having a solid, well understood hardware with a known SDK can make sense for low volume products (or even high volume products).

White label (1 to 3 mo)

Lots of wearable devices are commoditized and use standardized protocols over BLE to communicate. ODMs will sell you their device with your name on it.

Semi-custom (2 to 3 mo)

Yesterday’s DIY companies have grown up and are now offering compute modules that can be easily integrated into custom interface PCBs at reasonable prices for medium volume applications. There are also firms that have already integrated CMs into useful IO boards. This means you can mix and match electronics, put it in a custom enclosure (or semi-custom enclosure) and get to a polished hardware solution with a light lift. (Make sure any solution used for production has an EOL policy!)

Prototyping (0 to 1 mo)

We are in the golden era of low cost dev boards that can be easily put together to make compelling proof of concept prototypes (do not go to production with this solution!)

Fully custom hardware (2 to 4 mo)

Didn’t I just talk myself out of work? Yes and no. As volume grow or required feature sets don’t match up with off the shelf options custom hardware is very much advantageous or required. This is where a fractional engineering team to can help deliver on your hardware solution. Having robust hardware can also create a barrier to entry for competitors due to development cost and schedule.

Reach out if you need help parsing all these options!