Few areas of manufacturing are undergoing as much change and disruption as the electronics and semiconductor sector. The Manufacturer Editor Joe Bush speaks to Scott White, Co-founder and Executive Director of Pragmatic, a company at the coalface of this revolution.
A serial entrepreneur, Scott has been integral to the creation of a number of tech businesses and with colleague Richard Price, co-founded Pragmatic in 2010 with the aim of revolutionising the world of semiconductors with the development of the equivalent of a silicon chip, but in a form factor and at a cost point that made it viable to extend electronics beyond traditional applications and into areas such as smart packaging for fast moving consumer goods.
Scott commented: “If the goods you purchase in a supermarket can have integrated intelligence, that can allow the product to be traceable through the entire supply chain to reduce waste and improve inventory control. It could even track the packaging right through to end of life to ensure it gets recycled and reused in the most appropriate way.”
While most electronics today are made on rigid silicon wafers, Pragmatic’s integrated circuit (FlexIC) technology is thin and flexible, and its range of sensors and controllers are beginning to be deployed in applications where customers are keen to take advantage of the flexibility and/or lower cost that the company can offer.
An alternative to silicon
Global disruption in recent years has led to upheaval and a stifling of semiconductor supply around the world. As such, Pragmatic is experiencing an increase in demand for its technology as customers look to substitute silicon in areas where supply chain concerns are key. “
There have been many industries that have been impacted by a dearth of silicon chips,” added Scott. And it’s not the high-end silicon chips that have been in short supply; often demand has been in relatively simple functionality that’s made on legacy node silicon. To take the automotive sector as an example, there have been ongoing supply issues around controller chips for windscreen wipers and electric windows. They might only cost 5-10¢ but without them, that $100,000 car can’t be sold.
“We’ve seen that our technology is good enough to substitute that functionality but with a supply chain model that is far more scalable and guaranteed than legacy silicon.”
The greater reliability and lower cost point that Pragmatic can achieve results from a combination of the materials and the process. Due to the absence of silicon, the high cost involved in traditional wafer manufacture is immediately eliminated; the creation of a silicon wafer requires very high temperature processing and refinement plus expensive, high energy, long duration processes such as ion implantation and diffusion.
“All this ultimately means our entire process flow is simpler, more compact and faster,” Scott added. “For silicon, a typical cycle of wafer production from start to finish takes a minimum of three months (often it’s more like six, after you add in scheduling and tradeoffs etc); and that’s starting from a silicon wafer to get to an integrated circuit (IC). We can go all the way from raw materials to fully fabricated ICs in around two days.
“There’s no single secret sauce. We’ve achieved this through a combination of around 250 patents, covering different aspects of materials, device architecture, circuit design architecture and manufacturing processes, as well as a huge amount of know-how from our team.”
Current trends in a changing world
As previously mentioned, the world of electronics and semiconductors has undergone a period of unprecedented and unexpected disruption in recent years; as anyone who has recently purchased a new car or tried to get hold of a PlayStation 5 will testify.
However, with semiconductors a key component of just about every industrial product or process today, Scott explained that the sector is continuing to grow at an extremely fast rate. Typically, a new car rolls off the production line with over 1,000 chips onboard, providing a range of different functions. And that’s going to continue to grow over time.
Not only are the number of areas that now require this technology increasing, but the content of semiconductor devices themselves is also on the rise as they become more sophisticated. Within that, there is an interesting dynamic between what is needed in different types of semiconductor chips.
Scott explained that a closer look across the semiconductor industry will reveal a range of different materials used, beyond traditional silicon. As well as technology developed by Pragmatic, the industry is also developing compound semiconductors, like silicon carbide and gallium arsenide, that are used for specific functions such as high-power electronics or high performance radio frequency (RF).
He added: “There’s a lot of heterogeneity around semiconductor materials and chips that are relevant for different areas. And even within logic and mixed signal devices, there is a lot of granularity between the small number of applications that need very high-end silicon (which the latest three nanometre processes are aiming at), versus a huge range of similar functions, which actually perform better with mature node technologies.”
This has resulted in vastly different supply and demand implications across the sector. Memory, for example, is now in over supply, and Scott explained that Samsung is having to scale back production while TSMC is trying to put fabs in multiple locations around the world to balance supply.
However, with legacy node processes, there is still a very significant supply shortage. “Most of our customers who are looking to break away from that are doing so because they’re still seeing lead times of over 12 months or more for ordering chips. So, there are multiple elements underneath the surface of what many consider a single industry.”
Staying local
One trend that is common to all parts of the sector however, and driven by the fact that the industry has been built around a very concentrated supply, is the move towards localisation. Economies of scale in conventional semiconductor manufacturing mean that the most efficient method of production is in very large fabs.
That leads to very high geographic concentration of supply, and due primarily to historic government incentives, most of that manufacturing takes place in the Far East, particularly in Taiwan and China. Therefore, a key driver for many governments around the world is how to redress the balance and have a supply chain that has less geopolitical exposure.
Scott continued: “Our technology plays into that trend very nicely. Because of our far simpler process flow and smaller set of tools required to manufacture devices, we can not only build our fabs very quickly, we can also be more flexible in where we set them up. “
In addition, the cost of set up is 100 times less than a typical silicon fab and the physical and environmental footprint is also around 100 times smaller. Therefore, we can put fabs in places where it wouldn’t normally be viable to put a semiconductor manufacturing facility, making that localisation of production much easier.”
He added that this trend of localisation across the industry has become a very significant driver for Pragmatic’s technology. The company can even put fabs in individual customer premises (which Pragmatic call ‘Fab-as-a-Service’) to provide a dedicated production line for security of supply.
As the company scales over the next decade, the intention is to deploy over 100 fabs globally, with a significant proportion to be driven by specific customers who have either a volume or a strategic security of supply requirement to have their own dedicated capacity.
Semiconductor strategy
Because of the disruptive nature of the sector, the UK’s National Semiconductor Strategy was established to focus on areas where the country has distinct competitive strengths which can be sustained over time; namely the type of technology developed by Pragmatic. “Within that, there is an understanding that there needs to be support for semiconductor manufacturing,” he added.
“We currently have a strategy that is broadly sensible in terms of its direction, and there is good engagement from the government with industry on how that gets translated into implementation.”
An example of this is the Semiconductor Advisory Panel, of which Scott is a member. Established by the Department for Science, Innovation and Technology (DSIT), the panel’s aim is to enable the government to work closely with industry to grow the UK sector, ensuring a stable supply of chips and protecting the UK from national security risks associated with semiconductor technology. “
The panel includes people who understand and have participated in the entire breadth of the ecosystem over time. The idea is that we can be much closer to government and will be able to guide that thinking as it evolves, to ensure we’re heading in the direction for the broader industry.
“The devil is going to be in the detail on how that translates into real policy and other initiatives. And there’s still a lot more that needs to be done over the next months and years to make sure that the strategy is translated into something that is indeed valuable. But it’s a step in the right direction,” added Scott.
You could say that establishing the advisory panel is long overdue, as Scott highlighted that there has been a distinct lack of government level support for the sector in recent years.
This is reflected in how much the industry has changed over the last several decades. Back in the 1980s and 1990s, the UK had a very strong semiconductor manufacturing base that has almost entirely disappeared as underlying economic drivers and economies of scale drove production to the Far East. However, this decline was also impacted by a lack of government incentives.
“A few years ago, analysis by Boston Consulting Group (BCG) looked at different government incentives for semiconductor manufacturing versus manufacturing market share. Unsurprisingly, that showed a direct correlation between semiconductor manufacturing support and change in market share over time,” Scott added. Therefore, the places that have been given support have seen an increased and concentrated amount of semiconductor manufacturing.
He explained that when considering the fundamental economics involved, there is no reason the UK cannot scale manufacturing in this country just as effectively as elsewhere. However, to make this happen the country will need a level playing field. This is the nature of the discussions the Semiconductor Advisory Panel is having with government. Currently the landscape is being skewed by government incentives elsewhere, and if the UK wants to retain and scale manufacturing here, then a proportional response will be required.
“That doesn’t mean investment in the tens of billions of pounds,” continued Scott. “The absolute magnitude of it can be much smaller, because the technologies that companies like Pragmatic have at their disposal are far more capital efficient. However, they do need a similar support incentive to make sure that UK manufacturing is not disadvantaged compared to elsewhere.”
US expansion, UK commitment
There have been recent misleading reports in the mainstream press that Pragmatic is plotting a shift away from UK manufacturing as the company seek to establish facilities in the US. To quell such reports, Scott was quick to highlight that the sheer nature of the modular manufacturing model operated by Pragmatic means the company is able to set up fabs anywhere, based on economics and customer demand.
He explained that the US is an obvious geography to target as a significant customer base already exists in that market. It is predicted that Pragmatic’s Fab-as-a-Service model will be sought after by many of its largest customers to provide them the security of supply they need. And, because of the CHIPS Act, a $280bn government fund to boost domestic research and manufacturing of semiconductors in the US, there are healthy incentives in place to support deployment of capacity. “
US expansion was always part of our plan. We have established a US subsidiary and we are in discussion, both with our customers in the US and the government, around the specific plans of how we might roll that out,” Scott added.
This does not mean, however, that Pragmatic are any less committed to manufacturing in the UK. The company has made significant investment in its Pragmatic Park site in Durham, established to be a centre of excellence for flexible electronics design and manufacturing. The park is intentionally a much larger facility than the company require today to allow for future growth.
“We have significant scope for expansion,” Scott added. “We already have spots allocated on our floor plan for at least eight additional fabs and we intend to invest significantly to scale that up. There is, of course, an implicit assumption that we get the right support from government in doing that. “
But if that plays out as the current semiconductor strategy is implying, then our intent is to maintain headquarters, the critical mass of the business and all our R&D here. We also intend to scale manufacturing capacity in the UK to support not only our domestic customers, but also the export market which we can supply as effectively from the UK as anywhere else.”
He explained that what we need in the UK does not have to directly reflect The CHIPS Act, or indeed any other support packages that exist elsewhere (as other regions and nations have incentives to support onshoring of semiconductor manufacturing capacity). But it does need to provide a proportionately similar impact on the underlying economics of semiconductor manufacturing.
There is ultimately a dependency around the government having the right policies and incentives to sustain and support the growth of manufacturing in the UK. And while this country has a history of great semiconductor innovation and early stage technology development, there has traditionally been a shift overseas when it comes to scale up and manufacture.
Scott added: “Therefore, there is some inertia to overcome around changing people’s thinking and regaining the recognition that advanced manufacturing is a very high value activity. There’s sometimes a perception in the UK that manufacturing is a low value, metal bashing type activity – which it might have been 50 years ago.
“However, the country has moved into more value-added portions of the sector. Advanced manufacturing is something which is extremely valuable and profitable; one need only look at TSMC’s financials to see that, with net profits of over 40%. That’s not a low margin activity by any stretch of the imagination.”
The semiconductor strategy explicitly acknowledges the benefit and the need to support manufacturing. However, Scott highlighted that the mainstream press can sometimes misconstrue the intention of the strategy and fail to draw the distinction between the UK trying to catch up to other countries when it comes to silicon, versus taking advantage of advanced materials semiconductors, where the UK has unique strengths.
A decision was recently taken not to build a three nanometre advanced node silicon fab in Wales, due to a lack of ecosystem, environment and skills base to be able to support it. That has been interpreted as a lack of investment and support for UK manufacturing.
“The reality is that these days, advanced node silicon is too expensive and requires a specialist skill set which we no longer possess,” Scott added. “So, we’re going to have to work with the US and EU in that regard, to ensure that we have the right diversity of supply from a broader perspective.
“However, the UK has other strengths which CAN be effectively rolled out in the UK, namely technology such as ours and compound semiconductors.” So, the strategy has clearly outlined where the UK’s strengths are and is supporting those accordingly.
“The nature of our manufacturing model lends itself very much towards the right economies of scale for the UK at a much lower capital spend, and at a much more modular level. This has highlighted the potential to have many fabs scattered around the world, localising manufacturing to where our customers are.
“That trend has only been accelerated in the last few years with the shifting of the geopolitical landscape and the broader recognition that security of supply is a critical factor. There are obviously challenges in implementing that in practice, and in building a multinational organisation supporting manufacturing in multiple geographies. However, that’s part of the growth journey we’re on; it’s not unexpected or indeed, particularly new as a challenge.”
For more stories on Innovation click here.