An Internet of Everything has long been promised, but progress has been hindered by semiconductor availability, cost and concerns around sustainability. Could a new approach help to achieve a smarter, more connected future?

Enter Pragmatic Semiconductor’s FlexICs, a new kind of semiconductor that’s set to make the Internet of Everything a reality. Shane Geary, SVP Manufacturing & Operations, Pragmatic explains.

From smart manufacturing to predictive maintenance, digital transformation is already impacting every facet of business and industry. But the hyperconnected Internet of Everything (IoE) – and the promised efficiency and insight – has remained elusive, despite the market potential being huge. McKinsey estimates the global value of the IoT alone to be up to $12.5tn by 2030; amplifying the level of connectivity to encompass billions of smart items could see that figure soar.

Why hasn’t the IoE materialised?
In part, it’s a matter of supply: ubiquitous connectivity is reliant on abundant supply of semiconductors – yet existing methods of semiconductor fabrication would struggle to satisfy the demand for billions of smart items. IoE devices typically don’t use the latest, cutting-edge chips – the powerhouses you’ll find doing the heavy lifting in your smartphone or in self-driving cars. They tend to rely on so-called ‘legacy’ chips.

Despite the name, legacy chips are not old technology. They’re constantly being adapted for new requirements and applications and play a central role in a manufacturing economy.

Defined by the US’s CHIPS Act as semiconductors built on 28nm or larger process nodes, these chips are hugely pervasive. Working hand-in-hand with leading-edge processors, they’re found in virtually every electronic device on Earth.

Their importance was underlined during the pandemic, when demand outstripped supply. The world watched in disbelief as entire production lines ground to a halt and everything from cars to TVs went unshipped, hampering an already beleaguered economy.

If supply is what’s holding back the IoE, why don’t we just make more chips?
The fact is that expanding the production of legacy chips isn’t easy. They’re often produced in fabrication plants (fabs) using older equipment – 200mm wafers rather than current-generation 300mm wafers – and this equipment is less readily available.

However, building a new fab is costly and time-consuming, requiring tens of billions of dollars and maybe two years or more to get up and running. In any case, the required profit margins are easier to achieve on depreciated equipment and manufacturers are naturally hesitant to make such an investment in a mature technology, as there’s a greater risk of obsolescence.

Sustainability is also an issue. Many of these legacy fabs were built when carbon emissions were less of a consideration and guzzling energy and water was considered a necessary part of production. While newer fabs tend to be highly efficient, with carefully considered wastewater programmes and an emphasis on energy reduction, legacy fabs tend not to benefit from such endeavours – primarily because sustainable methods of production are far easier to ‘bake in’ than they are to retrofit.

Supply issues aside, the long lead times – and costs – associated with chip manufacture are also a factor; ubiquitous connectivity requires a low-cost solution that can be swiftly proliferated, at scale.

Essentially, existing modes of chip production are ill-suited to realising an IoE. To sustainably achieve pervasive connectivity, we need to look to new methods and materials.

Unlocking the Internet of Everything
Earlier this year, Pragmatic Semiconductor unveiled its flagship fab in Durham, UK. It was a momentous occasion in more ways than one: not only is this the UK’s first 300mm semiconductor fab, it’s the first to produce chips based on thin-film technology.

This focus on advanced materials represents a radical shift in the semiconductor industry, The company’s FlexICs – flexible integrated circuits – does away with the complex, high-temperature processes required for silicon chip fabrication, instead relying on simple spin-coating of polyimide onto a glass carrier.

Since this process takes place at low temperatures, it requires significantly less energy, less water and far fewer chemicals. This has a dramatic impact on the carbon footprint of the finished product, as well as on set-up costs and production timescales.

Standard chips take many months to go from final design stage – known as tape-out – to delivery, yet FlexICs can be delivered in as little as four weeks. The implications for innovation are multiple; rather than adhering to ‘right-first-time’ workflows, designers can take advantage of rapid cycle times to amend and refine designs on the fly, iterating to achieve optimal performance. Low non-recurring engineering costs also lower the barrier to entry, making it cheaper and easier than ever before to bring designs to life.

The simplified fabrication also impacts the size of the fab. A minimum viable silicon fab typically requires thousands, if not tens of thousands, of square metres of cleanroom space, which naturally requires significant economic investment. By contrast, Pragmatic’s ‘mini fabs’ are just a few hundred square metres, making them easy to drop into existing production sites and to maximise productivity per square metre. In fact, production volume is anticipated to be in the billions of chips per line, per year.

The ‘mini-fab’ model may also appeal to original equipment manufacturers (OEMs) or brands, since they could benefit from their own dedicated, on-site fab through the ‘Fab-as-a-Service’ model, providing security of supply and reducing supply chain emissions.