Curious research: "Plastic" micros using TFT fab processes

[Whales]

Just a curiosity: research level, only runs at KHz and doesn't have all memory onboard.

Summary and images: https://www.cnx-software.com/2021/07/22/plasticarm-is-a-functional-non-silicon-flexible-cortex-m0-microcontroller/
Paper itself (seems open to download): https://www.nature.com/articles/s41586-021-03625-w

Curious takeaway: TFT fabrication processes (ie those used to make LCD screens) instead used to fab a microcontroller.  Completely different substrate, authors claim it's much cheaper. 

Normally I would assume that the packaging & distribution costs of small micros is their dominant cost (not wafer+fab), but given the current shortages of certain models of small micro it might be worthwhile having a future where micros are fabbed on a wider variety of crazy processes.

[ataradov]

I don't see why it would be cheaper. Especially when specs are a bit closer to something usable.

We already have 3 cent MCUs that are much faster than this.

I also don't buy eco angle. The piece of silicon is just sand in the end, plastic packaging is a bit worse, but if we assume non user friendly package, then we can go with WLCSP and avid the plastic in a traditional process.

It is an interesting proof of concept, for sure, but I would not extrapolate too much.

[coppice]

In an era when the die for a small MCU costs less than the packaging and testing to turn it into a usable component, its hard to see how reducing the cost of the actual circuitry can do much for overall costs.

[Whales]

In an era when the die for a small MCU costs less than the packaging and testing to turn it into a usable component, its hard to see how reducing the cost of the actual circuitry can do much for overall costs.

Not at the moment, not everyone can get fab time.

[ataradov]

Sure, but what makes you think that there will be fab time with a weird TFT on plastic process if it gets mass adoption? They will be in the same exact situation.