WHY THIS MATTERS IN BRIEF
In the future computer chips won’t be made of silicon …
Love the Exponential Future? Join our XPotential Community, future proof yourself with courses from XPotential University, connect, watch a keynote, or browse my blog.
If you read my blog regularly then you’ll have read about the future of computing which includes everything from biological, chemical, and DNA computers, all the way through to weird liquid computers and ultra powerful neuromorphic and quantum computers which are hundreds of millions of times more powerful than the computers we use today.
There’s also another type of future computing platform though called MIS, which stands for Molecular Information System. First developed by IARPA, the bleeding edge US research arm of the US intelligence agencies MIS computers use a mix of polymers to create computers that have the performance and storage capabilities of a Google sized, or NSA sized, hyperscale datacenter into a for factor that’s no larger than a standard dining room table.
The future of computing is increasingly broad and versatile, and now British chip designer Arm have unveiled their newest chip design that will help build computing capabilities into every object in the world, and every Internet of Things (IoT) device – and it’s made from plastic.
The falling cost and size of silicon chips has made it possible to integrate sophisticated electronics into an ever-greater array of objects over the years. Think of almost any consumer good and you can probably find a “smart” version of it.
Get to know the tech
But there are limits to where we can reasonably add electronics. A disposable milk bottle that can automatically sense when its contents is spoiling is clearly out of reach, as is a throwaway wound dressing that can monitor moisture levels to ensure proper healing.
Arm thinks those kinds of applications may not be far away, though. In a paper published last week in Nature, researchers from the company detailed a 32-bit micro-processor architecture built directly onto a plastic substrate that promises to be both flexible and dramatically cheaper than today’s chips.
“We envisage that PlasticARM will pioneer the development of low-cost, fully flexible smart integrated systems to enable an ‘internet of everything’ consisting of the integration of more than a trillion inanimate objects over the next decade into the digital world,” they wrote.
Flexible electronics aren’t exactly new, I’ve covered them for years now, and sensors, batteries, LEDs, antennae, and many other simpler components have all been demonstrated before. But a practical microprocessor that can carry out meaningful computations has been elusive thanks to the large number of transistors required, say the researchers.
The new chip designed by Arm squeezed 50,000 transistors into a little under one square inch. This allowed them to create 18,000 logic gates, roughly 12 times as many as any previous flexible integrated circuit.
To do so, the company turned to British firm PragmatIC Semiconductor, which has developed a flexible electronics fabrication process that prints thin-film transistors directly onto a plastic substrate called polyimide.
The resulting chip is essentially a copy of Arm’s ultra-low power Cortex-M0+ processor designed for IoT applications, but with some significant caveats. The chip can only manage speeds of 29 kilohertz, more than 30 times slower than the roughly 1 megahertz the M0+ can achieve. Even worse, it uses 2,000 times as much power, though it’s still not much at 20 milliwatts.
The current iteration also doesn’t have a reprogrammable memory, so it can only run the operations it was built with, though the researchers expect to change this in future versions. The team also didn’t test the chip’s flexibility, despite this being pitched as one of its most promising characteristics.
This device is just a proof of concept, though, and many of its shortcomings will be worked on as the chip moves towards commercialization, which the company seems intent on. But Arm’s James Myers told The Register that these improvements were likely to be modest, as their goal was not to create chips as complex as standard silicon ones.
Despite the prominent mentions of the chip’s flexibility, its low cost is probably the more important factor. PragmatICs’ fabrication process has removed many of the more costly steps associated with building chips, and the raw ingredients are also considerably cheaper than standard processors.
These chips might not be nearly as powerful as their silicon cousins, but they will be much, much cheaper, and much more versatile – which is the important thing here. And there are many applications where price is a far bigger factor than processing capabilities, from smart clothing to intelligent packaging, and beyond.
Whether we really need to install chips in every object under the sun is another matter, but if an Internet of Everything is really in the cards, then these plastic microprocessors might be one of the key ingredients.