How Different the Factory of the Future Will Be?
So what is the factory of the future? It is one where bespoke products can be made at the cost of mass production, and where manufacturing and supply chains become ‘on demand’. It is being enabled through the mass connectivity of devices, through cheaper data collection, storage, and transmission, combined with advances in automation technology like robotics; it is being driven by consumers wanting bespoke products on-demand at mass-produced prices, and by manufacturers wanting shorter development lead times, lower production costs, and lower inventories.
It is one where products are designed, simulated, and tested digitally, and where these digital blueprints are translated automatically into production instructions ready for industrial control devices, to be sent anywhere in the world for production, leading to the potential for ‘distributed manufacturing’. The ‘digital twin’ of the manufactured product (the car, the aircraft, the phone) will remain and constantly evolve over the lifetime of the product to mimic its real-world counterpart, improving the product's lifetime performance.
It is one where augmented reality can aid in the setup and maintenance of the factory, and one where advances in robotic tools mean that dexterity, flexibility, and adaptability replace repetition and uniformity in the industrial process. Devices inside the factory are enabled with sensors and connectors and communicate over a dedicated 5G network, where advanced manufacturing processes like robotics thrive.
It enables circular manufacturing and where smarter product design and smarter production process design can improve the resource footprint and one which enables the verification of the provenance of products and components.
Factory automation has seen a constant stream of improvements since the advent of the assembly line (~1913), industrial robots (the 1960s), Ethernet-enabled factories (1980’s), and the advent of just-in-time and ‘lean’ manufacturing (1970s-1990s); even the theme of Industry 4.0 was first coined in 2011. Nonetheless, output per manufacturing employee has not increased notably in the U.S. since 2011, after enormous improvements from the 1970s through the early 2000s. We are now on the cusp of another wave of productivity improvement, driven by:
- The declining cost of data and computing;
- The declining cost of components;
- A pending explosion in the capabilities of industrial wireless connectivity (5G);
- Significant advances in automation technology like robotics.
Why Does This Matter?
Manufacturing will be greatly impacted by several emerging technologies in robotics, artificial intelligence, and machine learning. Today, robots have enabled mass fabrication. Tomorrow, they will enable a world of customized production where computational design and fabrication processes will support rapid manufacturing.
Automation suppliers have gradually shifted their strategies towards software, but until recently this has been (largely) domain-specific software; many industrial automation companies are shifting towards digital and data-driven strategies:
- It provides new revenue streams: Predictive maintenance is valuable to the customer through increased uptime and reduced repair time. Supplying software-as-a-service (SaaS) for product and process design mutes the traditional cyclicality of industrial hardware sales. For software offerings, advanced analytics and machine learning can be put to work operating on data in managed services scenarios and significantly boosting the value of these already high-margin offerings. There are also whole new markets; industry research firm IoT Analytics puts the software industrial IoT market at $1.7 billion in 2018 growing to $10 billion by 2023, including both the contribution from the Industrial Internet of Things (IIoT) and Augmented Reality/Virtual Reality (AR/VR).
- Manufacturers are demanding it: Often through lower production costs and shorter development times. One major industrial player estimates development times for aircraft can be reduced from six years to 2.5 years using digital tools, and that machine commissioning times can be cut by 25%.
- It is defensive in a period of technology disruption: In the past, Germany has lost its leadership in sectors such as consumer & communication electronics or carbon materials. Against a backdrop of technology disruption, the German government has promoted ‘Industrie 4.0’ to retain leadership in industries like automotive, medical devices, and more. China is also promoting this shift. Geopolitical forces alongside technological advancements can accelerate the desire to localize or re-shore production closer to the end market, which is far easier when the process is digitized.
- It is happening during a period of heightened trade tensions: This may mean multinationals will need to reduce the concentration risk in their operations away from China. In the May 2019 joint survey of member companies, the American Chamber of Commerce in Shanghai and China found almost 41% of respondents were considering relocating — or have already relocated — manufacturing facilities outside of China, versus only 18% of respondents who have indicated the same in an earlier 2018 survey.
- There is an increased focus on data on waste and resource usage: Analysis by McKinsey, as part of a study with the Ellen MacArthur Foundation found material cost savings worth up to $630 billion per year by 2025 in EU manufacturing sectors by increasing resource productivity. In the U.S., a study found benefits of 250–350mn metric tons of CO2 equivalent and $2 trillion in annual U.S. revenues could be generated from circular manufacturing. These points suggest that a move towards more circular manufacturing is likely to reduce costs for manufacturers.
We Can See This as a Secular Growth Theme Against Recent Cyclical Weakness
Many automation tools (from the assembly line itself to the industrial robot) have their roots in automotive markets, and automotive customers are still the single largest market for many factory automation suppliers, with electronics/tech hardware assembly another key market. Recent weakness in these markets has kindled fears for the growth outlook in automation markets. But as the technology spreads to other industries we see this current weakness as a soft period against a longer-term growth trend.
According to the Robotic Industries Association (RIA), the North American association of robotics, North America saw a decline of 3.5% (in unit terms) YoY in robot orders in the first quarter of 2019, attributed to a decrease across most major industries including Auto components (-16% YoY), Plastics & Rubber (-16%), Electronics (-17%), and Metals (-17%).
Looking at robot shipments versus orders, the decline was much sharper with a unit decline of 29% YoY. The RIA stated the weakness wasn’t unexpected and is mainly due to tough comparisons, as 2018 was a record year, and long-term signs remain healthy. Data for Japan (from JARA, see chart below), show a similar decline in the first quarter of 2019 for Japan. As I explain throughout the article, I see substantial opportunity despite the current weakness.
What is the Factory of the Future?
New manufacturing technologies and the proliferation of data and computing power are combining with customer demands to shift to ‘on-demand’ manufacturing, driving significant changes in the way goods are produced. This has far-reaching implications for companies that supply and use these technologies, and also for employees and economies. So what is the Factory of the Future? Definitions vary but summarize as follows:
- Flexible: Mass personalization (the ‘batch of one’) requires fully automated production and supply chains. On-demand production and supply chains become key enablers. Rather than a world of fixed products created offshore, this is a world where everything exists in design or cyberspace, where the customer creates the product they really want, and that product ends up being produced locally due to new advances in robotics and fabrication processes.
- Digitalized: Design, production, and service have to be digitally driven. To this end, companies talk of three ‘digital twins’ for manufactured products a virtual version in design, in production, and in use. Key themes here include the data from connected devices the IoT and the convergence of software and hardware. Digitization becomes a key enabler of distributed manufacturing, with ‘digital twin’ designs sent to and produced in fully-automated factories around the globe. Worker productivity can also be boosted by virtual and augmented reality devices to bridge the digital world and the real world in production, training, and service.
- Zero-waste manufacturing: This not only refers to energy (and other resource) efficiency but also the concept of ‘circular manufacturing’. The United Nations Sustainable Development Goals highlight the need for responsible production.
Who Are Likely to Be the Disruptors?
The ultimate winner might be the consumer, but between corporate players we see two key battlegrounds — between the traditional industrial automation suppliers themselves, and over time between industrial automation companies and technology companies, i.e., ‘new tech’ vs. ‘old industrial’.
- Optimizing the process through ‘vertical software’: ‘Vertical software’ is designed and used for a very specific industry, and is already one of the largest and fastest-growing software end-markets. The widespread adoption of connected devices creates a virtuous circle as more data collection allows for more advanced analytics, and so further product and manufacturing process improvement through software. Vertical (i.e., industrial) software companies are increasingly competing with traditional industrial players, and while domain expertise (i.e., industry and application-specific knowledge) is key, we see this as an area of increasing competition.
- Owners and interpreters of data: It is too simplistic to say that data is the new oil; industrial data is valuable but not a commodity. Much industrial data is not readily available by default — the vibration of a motor is analog by default and needs to be interpreted and digitized. Domain expertise and an installed base are both keys to collecting and contextualizing this data. We see industrial companies with large installed bases as having an advantage here.
- Suppliers of advanced manufacturing systems: While the penetration of robots in the automotive industry is quite high (>90% for some tasks), the penetration in the general industry is far lower. The adoption of collaborative robots (cobots) to work alongside humans is still in its infancy. For years, robots have supported human activity in difficult, dangerous, and dirty tasks. Increasingly more capable robots that are able to adapt, learn, and interact with humans and other machines at semantic levels will create new jobs, improve the quality of existing jobs, and buy us time.
- Suppliers of connectivity: High reliability and low latency are key for connected industrial components, and I see enablers of connectivity (the suppliers of connectors and sensors, as well the enabling communication technologies) as benefiting, although the adoption of 5G presents its own risks of disruption for incumbents.
- Cloud infrastructure is the backbone of the IoT: IaaS (Infrastructure as a Service, the ‘bare bones’ of servers, storage, and networks) and PaaS (Platform as a Service, to additionally include operating systems and databases, etc.) are key building blocks for Industrial IoT platforms; while industrial companies had moved towards PaaS models, the requirement for scale and technical expertise (rather than industrial expertise) is arguably beneficial to the cloud players here. Collaboration rather than competition between tech and industries seems the most likely outcome here.
Who Is Likely to Be Disrupted?
If distributed manufacturing means factories can be anywhere, what does that mean for labor, economies, and societies, particularly for manufacturing powerhouses like China and Germany? Will technology augment or replace the workforce, and if robots replace labor what is the impact on the workforce?
Composition of U.S. Job Market over the Last 150+ Years
If earlier industrial revolutions were about capital versus labor, is Industry 4.0 all about data? How will factory owners, industrial automation companies, and software companies compete and collaborate to maximize the utility of this data? If the manufacturing industry produces more data than any other sector in the U.S. economy, what is the risk from cyber security?
Distributed Manufacturing vs. Contract Manufacturing
Asset-light models of outsourcing manufacturing are not new — ‘fabless’ semiconductor manufacturing, where design and sales are kept in-house but manufacturing completely outsourced, was first pioneered in the late 1970s and early 1980s.
These models were built on long-term agreements; distributed manufacturing allows a much more flexible approach where small batches (or even a single product) are digitally instructed in real-time — digital instructions to digital assembly massively opens up the number of factories that can produce a given component.
Contract manufacturing has been around for many years in industries ranging from consumer electronics to clothing. Foxconn, the world’s largest contract manufacturer for consumer electronics, now has revenue of over $170 billion. As the name suggests, the manufacturing is contracted out with agreements lasting many months or years.
In contrast, distributed manufacturing uses short-term outsourcing of manufacturing for a given product or component, potentially for only one unit to be made.
The concept of ‘distributed manufacturing’ was first popularized in 2015, when the World Economic Forum (WEF) listed distributed manufacturing as one of its ten ‘emerging technologies’ for 2015. WEF commented, “the idea of distributed manufacturing is to replace as much of the material supply chain as possible with digital information.”
In this example, if a company needs 20 subcomponents these can be precisely designed with plans digitally transmitted to potential suppliers, who can use these digital plans to feed a fully-automated manufacturing line to make the component. While simple in principle, there are limits; a 3D printer can be programmed to print any plastic object, or a lathe can be programmed to cut any metal shape, but a factory built to produce shoes cannot suddenly start assembling aircraft.
Distributed manufacturing is enabled by the digitization of design and production that we discuss later, and there are some benefits — and risks — to this development.
Key potential benefits include:
- Reduction of inventory and transport times;
- Diversification of the supplier base;
- Optimizing industry capacity / lowering fixed assets ;
- Differentiated experience for the end customer.
Risks and disruptions could be:
- The disruption of traditional labor markets;
- The disruption of supply chains;
- The challenges of regulatory enforcement in industries where facilities may need to be certified (for example by the FDA);
- The ‘land grab’ — both organically and through acquisition — to build IoT platforms and access data may not always create value for every player; a small number of big winners may emerge.
We discuss the digitization trend in detail — in product design, production, and in-service — exploring the concepts of the ‘digital twin’ and the ‘Industrial Internet of Things, or IoT.
This also begs questions on data ownership and security, which we cover in our section on Managing Cyber Risk in Manufacturing As the IoT becomes more prevalent and industrial production leans more toward connectivity, data security becomes a bigger focus. According to Intel, a connected factory is estimated to generate 1 petabyte per day, the equivalent of data of 1.05 billion minutes of MP3 songs, or data of 160 million books.
According to Cisco:
- Manufacturing produces more data than any other sector in the U.S. economy;
- 40% of manufacturing security professionals said they do not have a formal data security strategy;
- There is a five-year lag-time in IoT adoption in factories due to cybersecurity concerns;
While there is much focus on the impact of technology on jobs, the role of technology augmenting, rather than replacing, workers are also prevalent. Cobots (collaborative robots) are designed for collaboration with workers and augmented reality (AR) can be used for instruction and guidance on production lines or for maintenance and service.
The European Commission has targeted the ‘Circular Economy’ as part of its Horizon 2020 program, commenting that “…the circular economy starts at the very beginning of a product’s lifecycle — smart product design and production processes can help save resources, avoid inefficient waste management and create new business opportunities. By providing instruments and incentives to improve the production phase, the actions put forward will not only help save resources but also boost innovation and cross-border trade in the EU single market”.
- Better product design: More durable products with replaceable parts “the circular economy”.
- Fewer resources in production: German Auto manufacturer Daimler intends to cut CO2 emissions at its new ‘Factory 56’ by 75% as compared to its existing S-Class production facility, including by having a CO2 neutral energy supply.
- A focus on by-products and waste: One in which all waste products are productively reused and as such, there is no absolute waste burden.
New Technologies Are Defining Industrial Policy
The impact of automation has been seen for years, although arguably the impact is less recently compared to the late 1990s and early 2000s. In the U.S., manufacturing employment in absolute terms peaked in 1979, and while manufacturing employment has been growing since the financial crisis, employment is not back above the level seen in the pre-financial crisis and is still far below the levels seen in the 1970s, 1980,s and 1990s.
In output terms, however, U.S .manufacturing output is close to a new high, only 1% below the 2007 output peak in 2018 (see Figure 10 below). If we look at output per employee, however, U.S. manufacturing produces more than 3x more units per employee as compared to the employment peak in 1979. Output per employee has actually been quite stable since the financial crisis, with most of the improvement coming in the late 1990s and early 2000s.
It All Started in Germany…
The term ‘Industry 4.0’ — for the fourth industrial revolution — was first coined by the German government. According to a European Commission paper, for Germany, Industry 4.0 is a “national strategic initiative from the German government through the Ministry of Education and Research (BMBF) and the Ministry for Economic Affairs and Energy (BMWI).
It aims to drive digital manufacturing forward by increasing digitization and the interconnection of products, value chains, and business models. It also aims to support research, the networking of industry partners, and standardization.”
In Germany, the moniker ‘Industry 4.0’ has become the widely-adopted label for next-generation manufacturing, referring to the fourth industrial revolution (after the steam engine in the 18th century, the advent of the assembly line in the nineteenth century, and the introduction of automation technology in the 1970s).
Since the launch of Industry 4.0 at the Hannover Trade Fair in 2011, we’ve seen an increasing number of companies define their strategies around the concept. The German government supports the initiative to ensure the global tech industry — largely based outside of Europe with software titans typically U.S. based and hardware specialists now largely Asian domiciled — doesn’t benefit to the detriment of Germany’s manufacturing industry.
… with China as the Driver
China’s growth in higher value-added manufacturing has been a theme running for years, although the ‘Made in China’ 2025 plan, introduced by Premier Li Keqiang in 2015 formalized the initiative. While the ‘Made in 2025’ program may have been deemphasized publicly, the tenets of achieving higher domestic share targets in critical industries will still resonate and likely intensify due to U.S.-China tensions.
Expedited by the recent trade war with the U.S. and increasing wages, the shift is visible; China is moving away from low-end goods (textiles, food products, etc.) to medium high-tech goods (vehicles, electrical machinery, construction machinery, etc.), and is now the world’s largest producer in the latter (32% global share according to the National Science Board).
Moreover, this is helped by not just the large domestic market but also the country’s growth in exports (which has been growing at a resilient pace of an average of 5% per year since 2009) enabling export-focused Chinese capital goods companies and foreign companies investing in China to boost both their size and scale of investments within the country.