Onshoring in the post-coronavirus future: local goods for local people

In a sign of how much the lockdown has changed the way things are done, the TED talks that normally take place on carefully lit stages in front of audiences enthusiastic to hear about technology trends took on a far more downbeat format when they invited hedge-fund manager Ray Dalio to talk about the future and the effects of coronavirus on business.

“This is bigger than 2008,” said Dalio. “We’re going to change how we operate and the supply lines are going to change. We are not going to go back to the way it was.”

Agreement is easy to find as forecasts look gloomier the longer the lockdown lasts. Malik Saadi, managing director and vice president, strategic technologies at ABI Research notes: “This is not like any downturn of the past. Covid-19 has affected all elements of the value chain almost simultaneously and it has become clear that Covid-19 has taken the world by surprise. There is no clear visibility on what the future holds.”

It is easy to make predictions of a new normal settling after a crisis while the crisis is still raging. Companies may fail to go through with anticipated changes. But like many others looking at business operations from the outside, Dalio expects them to focus far less on the seemingly endless cheese‑​paring of the past few decades and instead look much harder at resilience and robustness. Rather than subprime assets sinking the ship, this time the cause is supply chains with underlying frailties: ‘just-in-time’ quickly turns to ‘just about working’ and then fails completely when you start knocking out the weaker links.

Rian Whitton, senior analyst at ABI, says: “The post-Covid supply chain won’t be like the supply chain we currently have. It will stress diversification of manufacturing and reshoring. That is only going to be possible with significant investment in productivity-improving technology.”

Among others, ABI’s analysts are recommending manufacturers should not just avoid single suppliers but find ways to use components and subsystems that can be made in different locations. To a large extent, that means a shift away from China, or at least the near total reliance many industries have on manufacturing in the country with the world’s largest population.

Chaney Ho, executive director and co-founder of embedded-computer maker Advantech, is among the CEOs of technology companies who see that trend as having begun already but achieving greater urgency now: “The globalisation rule will be reset. Many factories will be decentralised. Factories will move closer to the market. Provide a service where it’s needed. That will be the trend. And I think many companies will have regional hubs.”

Although decentralisation away from China is likely to be important, it is not the solution in itself. Omdia’s principal analyst for semiconductors Hui He wrote in early April on the issues caused by single points of failure being distributed around the world. She noted electronics manufacturing remained affected around the world into April despite China reopening its factories.

In an online seminar hosted by the electronics-manufacturing trade group Semi, McKinsey & Company principal analyst Sven Smit said: “There are some companies in China who can’t get supplies out of Italy. If the world doesn’t have a V-shaped recovery, China will have it tough too.”

Despite the troubles in the flow of goods internationally, high levels of automation made a difference. “The supply chain for semiconductors was not as heavily disrupted as other industries,” notes Whitton.

In the current environment, political and corporate ambitions are realigning by putting greater emphasis on local production, potentially rewinding the clock by close to 40 years. In the 1980s, inward-investment policies in more developed economies tended to favour maintaining assembly operations while the components wound up being pushed to cheaper regions. Anti-dumping sometimes slowed the migration of component production, but not for long.

However, it is easy to overstate the alignment between politicians and business. One outcome is a possible retreat from globalisation that delivered many of the cost savings manufacturers wanted for political reasons but with little regard for whether the approach is desirable by either faction. The situation with 5G communications is just one example of the way that geopolitics, technology and manufacturing have come back together. In that case, it seems likely to favour greater competition in hardware and software but that is more of an accident of circumstance than anything.

‘The globalisation rule will be reset. Many factories will be decentralised. Factories will move closer to the market. Provide a service where it’s needed.’

Where manufacturing does come home, one potential difference between tomorrow’s local factories and the screwdriver plants of the 1980s lies in the higher degree and style of automation they are most likely to deploy. The problem with the screwdriver plants, particularly those created as the result of inward investment from large multi‑​nationals, lies in their relative inflexibility. If, as happened in the late 1990s when the Asian Tiger economies stumbled, they were no longer useful for their original task they either went idle or had to be stripped down and repurposed entirely. Some became little more than big sheds hosting call centres, which have themselves come under threat because distributing those calls to workers operating from home using voice-over-IP connections is far less infectious.

One possible result of the reversal of cost-focused globalisation is a change in the way manufacturing sites themselves are organised. A much higher level of automation may be an essential component in decentralising production.

In the shift to the Far East, many countries have lost the social infrastructure that supports manufacturing, a factor that has hindered a desire to boost onshore production. At its analyst meeting last September, Siemens Digital Industries played host to speakers from companies that have traditionally operated far from manufacturing but which now see opportunities. An example is Amazon Web Services (AWS), which sees opportunities in handling the flow of data between and around these sites. Cherie Wong, general manager of AWS IoT analytics services, said at the event: “There is a shortage of expertise. Skilled people in developed countries don’t think about careers in manufacturing. But now there is a trend towards reshoring. For that we have to be thinking about microfactories.”

The microfactory is not a small factory but a different way of thinking about production, in that a production plant can be formed of many microfactory cells that come together in different ways depending on what needs to be made at any one time. It denotes a way of setting up production that runs counter to the emphasis on high throughput in many existing facilities.

Alastair Orchard, digital enterprise vice president at Siemens, predicts the creation of multipurpose plants that deal with much smaller batches, avoiding the need for OEMs to build their own lines. They can instead call on local contract facilities, sending them the software needed to manipulate the line for a specific product. “You can automate processes without building in rigidity. You can’t easily personalise products halfway around the world, you need to run those lines onshore. With local production you can scale down to a lot size of one.”

Maurizio Cremonini, head of marketing at automation specialist Comau, says: “Even before Covid, we saw evolution in this area towards smart factories. Demand can be very erratic,” which drives manufacturers to build-to-order set-ups. Local production helps in these situations because it cuts the time to delivery to a minimum. “They want factories that are very easy to reconfigure.”

Siemens is working on methods to manage these more flexible machine tools and change what has been a dominant trend in production for several decades. Orchard says there is a move away from dedicated tools on the shopfloor that the manufacturer has to treat as a black box. “We are trying to break those black boxes down,” he says.

Cremonini adds: “This is where software is taking over from hardware solutions. Rather than having a very complex hardware manufacturing, the trend is going to very lean factories where the robotised and automatised cells are managed by software, loading different recipes as needed so they can easily modify the production process.”

The reprogrammable manufacturing-cell concept forms part of the much larger Industry 4.0 initiative: an attempt to make production lines more responsive using robotics and automation. At its core is the idea that each product carrier moving through the line tells the machinery what it needs next and where it should go afterwards using wireless communications. An onboard microcontroller either looks up that information in a central database or stores the production data in its own memory. Conveyors and automated guided vehicles move the product carriers around, and as each one arrives at a robot or machine tool, that tool downloads programs and data that carry out the task the carrier requests. Once manufacturing completes, the carrier’s data is overwritten and it is sent on to its next job.

Suppliers such as Bright Machines expect to capitalise on the emergence of micro­factories by offering highly roboticised machinery, arguing that modular machines can help overcome staffing shortages. In a recent move, Bright launched a leasing service, before following up with a programme for manufacturers building ventilators and similar products for the Covid-19 healthcare response. The company said it would provide free access to any microfactory deployment for up to 12 months if used for the pandemic response.

In one project, begun before the lockdown started, Cambridge-based HIV-diagnostics specialist Diagnostics for the Real World (DRW) decided to employ microfactory units to take care of some of the labour-intensive assembly and inspection tasks needed for the production of its test cartridges. DRW expects to be able to increase manufacturing throughput ten-fold over the course of a year, to an annual volume of one million units.

Microfactories may complement rather than replace larger facilities, which are themselves beginning to move to Industry 4.0 set-ups. They potentially make it easier for smaller OEMs to access personalised manufacturing through outsourced providers instead of having everything passed to a single provider in the Far East.

For larger producers, local micro­factories  may act as a supplementary source or backup production if another supplier is unable to complete a job. Some may also not form part of the core production flow but are recruited as needed for the manufacturing of spare parts – using additive processes such as 3D printing as an alternative to conventional moulding and assembly processes.

The set-up may partially reverse the approach used in the 1980s. Cremonini sees the possibility in the vehicle-manufacturing sector for microfactories feeding components and subsystems into large-scale local plants. “For some of the ancillary production you may see a good application of these smart cells and microfactories, providing the main line with a lot of parts. They may use this for supply-chain resilience or for full production,” he says.

To cope with the need for more heavily customised built-to-order situations, some vehicle makers, particularly in the light-industrial sector, are beginning to look at more modular designs based on ‘slices’. Cremonini notes: “Smart factories or microfactories could produce the standard slices, with final assembly putting the slices together.”

Some things will be highly resistant to reshoring and subject to the same migration pressures as they have been for the past 40 years. A prime example is chipmaking. Although the European Commission considered trying to encourage the creation of a shared fab capable of making the most advanced chips, the project never got off the ground. Today, most advanced production is in South Korea and Taiwan, with a limited amount in the US.

On top of that, many OEMs rely not just on a single location but a single fab for each product they employ, with little prospect of building a more resilient supply. Only the largest, operating on the scale of Apple or Samsung, are able to guarantee multi­sourcing, or those in sectors such as military production, which use older technologies available from multiple sources. Local factories will be confined to taking these devices and putting them on PCBs.

Another factor is the availability of technologies that help increase resilience that do not necessarily call for localised production. Lux Research, for example, expects a digital-twinning approach similar to that now being applied to product design to improve OEMs’ visibility over what they buy. In this environment, an OEM would use digital models to help switch suppliers and subcontractors quickly – and there would be no inherent requirement for these suppliers to be local.

One of the problems faced by OEMs in the recent crisis was the inability to forecast where supplies would dry up because the sourcing decisions were down to the contract manufacturers. Delaware North America managing director Richard Seel expects a role for blockchain technologies in improving visibility. “Traceability down the entire supply-chain hierarchy and understanding which suppliers provide certain materials, their batch numbers and countries of origin has never been more critical,” he says. The blockchain is useful in this context because it does not rely on the creation of a single, central database to store the information.

Augmented reality (AR) and virtual reality (VR) represent other potential challenges to onshoring. The pandemic has focused attention on the use of AR to help operate and even modify production lines. Toyota uses PTC’s Vuforia Chalk to help commission production lines. VR provides the ability to work on digital prototypes of production lines. Eric Abbruzzese, research director at ABI Research, says it may be easier to collaborate on designs using accurately scaled CAD models in a virtual space and that these technologies may get a boost “because companies realise there is value in using them to work from home”.

Although latency is an impediment for transcontinental communications, there is little other tangible difference between organising a product set-up in a micro­factory in Swindon versus one in Binh Duong, Vietnam. If delivery lag is not a concern, manufacturers will likely favour the lower-cost option.

For the moment, the balance has tilted in favour of resilience over cost – though it is highly possible that as the world moves beyond the pandemic, assuming a vaccine is found and distributed reasonably quickly, complacency and a lack of investment in production equipment may lead to a return to old structures as happened in the wake of the 2008 crash.