By Daniela Raffel

In Part 1 of our series on the future of manufacturing, we made the case that the intersection of a broad range of macro trends including generational shifts, rising energy prices, geo-political instability, near-shoring, and the rise of new technologies, means that the time really is now for Industry 4.0. In case you missed it — here’s Part I.

In Part II, we dive into the details of the innovations we are most excited about that are already driving the transition to the interconnected factories, plants, and warehouses of the future. We’ll also give some examples of companies pioneering these innovations, but it’s by no means an exhaustive list!

A factory without workers, even the most automated, is nothing — so we will start with the people that make it happen and the generational shifts that are changing everything.

As in any business, manufacturers face the reality of drastically changing workforce requirements. Not only are the stalwarts of the shop floor starting to retire, with one-third of the manufacturing workforce over 55 years old, but more automation is, in fact, increasing the number of vacancies. This is already creating headaches for the 77% of manufacturers that struggle to fill vacancies today. Accounting for additional vacancies due to automation, this could cost the industry up to $1 trillion by 2030 if they remain unfilled.¹⁴ ¹⁵

Against this backdrop, we see an opportunity for industry-specific networks that can support businesses in finding the right talent with the right experience to fill the ever-present gaps in their workforce whilst providing skilled workers with a better view of the opportunities available to them. Once a position is filled, the challenge doesn’t go away — it evolves into one of effective onboarding and training. To complicate matters, new employees often don’t stick around long enough to pick up the decades of experience their colleagues might have. In fact, 40% of frontline manufacturing workers changed jobs in the last year.¹⁶

The good news is that 46% of frontline workers think that tech could make their jobs easier, and so do we.¹⁷ Innovative companies, like Pixaera, SynergyXR, Seabery and LuminousXR, leverage VR and AR to create engaging, impactful, and cost-efficient training, without burdening the experienced workforce with training new hires alongside their daily responsibilities. Once training is complete, companies like Kit-AR and Sphere Tech support workers in their day-to-day work, reinforcing training and providing real-time instructions for assembling increasingly complex products. This helps ensure that new employees quickly become as efficient as their more experienced colleagues. In fact, VR/AR technology alone represents an $8b opportunity — and according to PwC over 50% of US manufacturers will employ it by 2025.¹⁸ ¹⁹

It’s not all VR and AR — companies like Fabriq, Azumuta and Lumiform are digitising checklists and tasks, taking them from clipboard and pen to mobile phone or tablet. This makes it easier to adhere to processes, whilst reducing the burden of manual data reconciliation and allowing for real-time insights and updates to be surfaced to management.

Helping workers be more efficient is one thing but making them safe can be an additional, non-negotiable, challenge. One executive summed it up: ‘If you think training and safety is expensive, try having an accident’. It’s not just about reporting: experts we have spoken to confirm that workers who feel safe feel valued which can help make them less likely to leave, helping to solve the first challenge of filling vacancies. Companies like Protex and Intenseye are helping Health and Safety leads create a safe environment by keeping an AI-eye on conscious and unconscious behaviours that might become a hazard and helping businesses promote best practices across all their sites.

These solutions, either independently or in combination, can help businesses onboard and train new workers more effectively to help them reach higher levels of efficiency faster. So now we turn our attention to the actual process of manufacturing a product.

An efficient production facility ultimately comes down to:

i) having a well-designed standard operating process

ii) making sure that the operating process is put into practice

iii) ensuring there are no unforeseen interruptions to the carefully planned process

iv) continuous improvement

There have been innovations across each element so let’s start at the beginning.

Shopfloor optimisation starts in the process design phase. Traditionally this has been done by siloed teams, across product specification, product design, and manufacturing process design, each using their own specific software design packages. However, without collaboration tools, naturally, there are numerous iteration cycles between the teams, and it is difficult to optimise all parameters in parallel — and all these changes must be carefully tracked by the quality and compliance team. Companies such as Aletiq, Trace.Space and Assemblio are changing the game here and supporting companies in closer collaboration between each of these teams, collating audit-ready trails of design decisions and revisions, the definition of standard operating procedures and production pathways through the shopfloor.

Having planned the theoretical process, it’s about putting it into practice. Operations planning is a critical part of a healthy production line. However, it is often still done across disparate spreadsheets and whiteboards, or it sits in inflexible systems such as SAP S4/HANA that are too costly and rigid for mid-market companies. Companies like Pelico, Oplit, and AMFG are helping factories and plants of all sizes connect their machines, Manufacturing Execution Systems (MES), and ERPs to optimise their scheduling operations. This connectivity in turn allows production to react to internal and external factors more easily to take planned maintenance, worker shortages, malfunctions, customer demand, and supply chain disruption in their stride.

Having designed and connected the ideal process, it would be a shame for production to suddenly and unpredictably cease. Indeed, keeping the assembly line running is top of mind for all plant managers.²⁰ The UK alone spent $7.3bn on new machinery last year and with unplanned downtime costing more than $60k an hour, the payback time on that investment gets longer and longer. Indeed, the risk of unplanned downtime can often make implementing new technologies challenging as the risk and cost of it is too high.²¹ ²² ²³

This is where maintenance programmes and strategies become important and predictive maintenance is a key lever. These can reduce the stores of spare parts, minimise the impact of supply chain disruptions, and reduce pressures to remediate breakdowns rapidly, lowering worker stress and risk of safety incidents.²⁴ To quantify this, effective predictive maintenance strategies are estimated to deliver up to 25% improvements in throughput and 70% reduction in breakdowns.²⁵ No wonder it’s one of the top priorities for CEOs.²⁶ ²⁷ ²⁸

Historically, achieving predictive maintenance programmes has been challenging: large volumes of labelled data to train models are not readily available, the cost of sensors can be prohibitive to SMEs, and experienced machine operators may be hesitant to trust the insights from algorithms over their planned schedules.²⁹ But businesses like Fracttal, Infraspeak, and Remberg, along with the tailwinds of declining sensor costs, are making it easier for mid-size factories to reap the rewards of predictive maintenance. This all contributes to a growing predictive maintenance market that is forecast to surpass $14b by 2028.³⁰

Even with the best-laid plans and strategies, production can still go sideways, literally. That is where companies like Ethon AI and Cerrion can help. These businesses help factories detect production line issues and quality defects immediately. They can also aid diagnosis of the root cause so that action can be taken immediately to stop the acute issue and avoid it happening again. Driving efficiencies across a complex and connected production line is a complicated task and of course, can be improved over time, be it through new insights or simply ensuring that standard procedures are continuously followed. This is why companies like Deltia.ai and Vision Intelligence harness computer vision to recommend improvements in set operating procedures, and Oden Technologies can help factories optimise to their specific goals. Within our own portfolio, Dataiku is developing a number of industrial solutions to ingest manufacturing data and deploy AI for, among other solutions, identifying the root cause of production defects and optimising batch production scheduling.

But what is efficient and predictable production without inputs?

Whilst globalisation has opened opportunities for competitive international sourcing, it has strained the complexity of supply chains, which are only stressed further by heightened geopolitical tensions, climate change, ever-changing consumer expectations, escalating transportation and warehousing costs and the spiralling numbers of relationships to manage. Automotive manufacturers, for example, have on average ~250 tier-one suppliers but this actually represents about 18,000 suppliers operating across numerous countries and continents.³¹ ³²

Increased supply chain intelligence is critical in light of unstable supply chains to ensure factories can continue operating at full capacity. Indeed, a Bain survey of nearly 300 executives highlighted that increasing resilience, flexibility and speed in the supply chain is top of mind for them over the next three years, with a higher priority than cost and customer service.³³ But, supply chain intelligence cannot occur without first deriving the baseline and for that, we need supply chain transparency. This need has permeated the industry and will become all the more important in light of regulatory changes. For instance, the EU Digital Product Passport (DPP), which it is anticipated will be approved this year, will mandate traceable information around a product’s composition, including raw materials and energy consumption, throughout its full lifecycle.³⁴ This acts as a step up from a number of national initiatives already in place, such as 2023’s German Supply Chain Act, currently applicable to ~3000 companies operating in Germany, by requiring that sustainability data is consistently publicly available to consumers, rather than ‘just’ an internal reporting obligation.³⁵ ³⁶

As producers increasingly need to adopt solutions to become compliant with a new wave of regulation, there will be plenty of opportunities for new companies to emerge to serve them. In the long term we believe successful builders in this space will look to proactively add value to manufacturers — such as leveraging AI to recommend supply chain changes, design or material changes, demand intelligence, or supporting new business models like products-as-a-service. We’re already seeing this happen with supply chain intelligence companies like Prewave leveraging AI to uncover the complexity and interrelation of supply chains and events that could cause disruption. From the DPP angle, solutions like Carbonfact and Tset which started as emissions management solutions have already evolved to move beyond purely regulatory compliance towards proactive supply chain management or design decisions to drive sustainability and cost reduction.

With just some of the solutions we’ve laid out, there’s a path to running a fully trained, stable and productive supply chain. If you think there is a missing link here, then you would be right. That missing link for us is data, and we see a huge opportunity here to help factories capture and use the data needed to adopt many of these solutions and more.

Laying the infrastructure for a data-driven factory

The reality is that for every connected factory, many more are still in the early stages of leveraging a portion of the data that some of their machines produce in real time. Taking advantage of all the benefits of optimisation and automation can only take place if real-time production data is being efficiently collected, stored and analysed.

Data collection needn’t mean a new machine: it can mean a new $0.50 sensor or a USB-connected camera and companies such as Ecoplanet are leveraging this to help factories get a better grasp on their energy consumption. Given energy prices have soared by up to 10 times in recent years, leading to 4.1% of revenues being used to pay for energy inputs, finding a baseline of energy consumption and understanding when and why there are deviations to this is important.³⁷ ³⁸

Sensors that better monitor production, convert analogue to digital data, and embed edge computing for real-time decision-making are broadly referred to as IoT devices. It’s the technology that is today perceived to have the highest potential impact on a company’s future success and profitability and, to date, over 16 billion IoT devices have been installed globally.³⁹ ⁴⁰

Once the data has been collected, finding ways to understand it is important. As businesses have taken different approaches, and used different sensors, solutions that normalise and standardise these signals will become key. It is no mean feat to set operational standards across industries, and Europe’s International Data Spaces Association is seeking to encourage data sharing to ‘lift all ships’ in areas such as supply chains and predictive maintenance.⁴¹ In the meantime, Industrial IoT (IIoT) companies like HiveMQ, and BarbaraIoT are harmonising data between different sources allowing for edge computing, ​​securely and reliably transferring the data at a large scale. Within our own portfolio, CluedIn’s data management platform is also being used to address this problem, allowing manufacturers to maintain a golden record and reconcile all the data streaming in from their factories. At the most sophisticated end of the spectrum solutions like Tomorrow Things allow businesses to analyse the data they collect as digital twins to make data-driven decisions and OctAiPipe and Namla can even do this at the edge.

These innovations are just the start, and as businesses transition to the smart factories of tomorrow, new needs will emerge, including cybersecurity and truly integrated supply chains and production. We can’t wait to see these enter the mainstream.

Despite the clear advantages of a more connected and enabled factory, growing and scaling an Industry 4.0 business hasn’t been straightforward to date. In the third and final part of our deep dive into the future of manufacturing, we will share lessons from experts in the field on the success factors they have deemed important in scaling a business in this industry.

If you’re a founder innovating in this space, please do get in touch with: mina@dawncapital.com, daniela@dawncapital.com and skye@dawncapital.com.

¹⁴Manufacturing Industry Outlook | Deloitte
¹⁵Creating pathways for tomorrow’s workforce today | Deloitte and The Manufacturing Institute
¹⁶Frontline Workforce Report | Beekeeper
¹⁷Work Trend Index Special Report, “Technology Can Help Unlock a New Future for Frontline Workers” | Microsoft
¹⁸IDC
¹⁹US Metaverse Study | PwC
²⁰2024 Predictions: Smart Manufacturing | Forrester
²¹Three tailwinds for robotics adoption in 2024 and beyond | EY
²²The Robotics Revolution | BCG
²³UK Innovation Report Launch | Cambridge Industrial Innovation Policy, March 2024, London
²⁴Capturing the true value of Industry 4.0 | McKinsey
²⁵Predictive Maintenance | Deloitte
²⁶January 2023: “Our ability to run our equipment at, well, really flat maintenance costs in the face of very, very strong inflationary pressure…[is] a credit to that team and the work they’ve done from a predictive maintenance standpoint.” | Liberty Energy
²⁷May 2023: “[Discussing Ford Pro support package] Think about this future state, which we are executing to. 100% prediction of failure of components before they fail…The software gets better every day so you can predict failure more precisely and for more components. The customer gets a vehicle that never really goes out of service.” | Ford
²⁸What CEOs talked about in Q1/2023: Economic uncertainty, layoffs, and the rise of ChatGPT | IoT Analytics
²⁹AI for asset management and predictive maintenance webinar | Institute for Manufacturing, January 2024
³⁰Predictive maintenance market: 5 highlights for 2024 and beyond | IoT Analytics
³¹Reimagining industrial supply chains | McKinsey
³²A Unified Approach to End-to-End Supply Chain Transformation | BCG
³³How CEOs Can Balance the New Supply Chain Equation | Bain & Company
³⁴The EU Digital Product Passport shapes the future of value chains | BCG
³⁵The German Supply Chain Due Diligence Act and the Chemical Industry | Mayer Brown
³⁶Overview of the German Supply Chain Due Diligence Act | Taylor Wessing
³⁷Forrester analyst call
³⁸Rising energy prices and their impact on the manufacturing industry: which sectors are being hit the hardest? | Caixa Banks Research
³⁹Transforma Insights
⁴⁰The Fourth Industrial Revolution, 2020 | Deloitte
⁴¹Manufacturing Data Spaces | Forrester