How can AI Help Accelerate Europe’s Industrial Renaissance?
AI modelling and mapping can potentially create more visibility of the emerging new industrial ecosystem to support better, faster combinatorial innovation
Industrial AI is starting to transform the way we make things, and the cost model for doing so, which could provide a much-needed boost to the industrial renaissance that Europe needs in the face of growing geopolitical risk. But whilst AI will play a role in products and production processes, it could also find an important new role in visualising the changing industrial ecosystem to inform more agile industrial policy and defence procurement.
A good guide to where we are with overall AI adoption is Stanford’s comprehensive AI Index for 2025, recently published by its Human-centred AI group (and which thankfully includes the raw data that underpinned it). The report shows just how widespread and common AI adoption is becoming in areas as diverse as business operations, healthcare, science, and consumer applications; but it also highlights challenges and gaps in education, governance and safety. There is also a nod towards geopolitical debates, noting that China is close to catching up with the USA as an AI leader.
With so much capital and computing inputs required to evolve some of the leading AI models, it is unclear whether consumer GenAI will be profitable any time soon. But this appears to be a similar pump priming phase as we saw in the early dotcom era — and before that in telecoms — where early investors create the infrastructure that ‘fast followers’ and second wave firms build on to find new applications and opportunities.
But the most interesting, impactful and possibly also the most economically transformative applications of AI may not be in consumer applications but rather enterprise, industry, government and infrastructure.
For various reasons, including but not limited to its principal ally deciding to cosplay as a banana republic, Europe’s future now depends on a massive industrial renaissance in which AI, smart technology and more productive ways of working could be vitally important force multipliers that allow us to scale innovation in smarter ways than we have seen in the past.
This industrial renaissance could also help generate new high-value jobs in economies where low-level knowledge roles have probably peaked, and look set to be heavily impacted by automation, as the New York Times discussed recently:
The gap in wages between those with a college degree and those without one grew steadily beginning in 1980, but flattened during the past 15 years, though it remains high.
The flattening may partly reflect the fact that there are more college-educated workers for employers to choose from, as college attendance has increased. But some economists argue that it reflects employers’ reduced need for college graduates — for example, fewer jobs like bookkeeping as information technology has become more sophisticated. Such jobs don’t necessarily require a college degree but were often attractive to graduates because they once paid them a relatively high wage.
Adoption is happening, but readiness remains an issue
Despite concerns that old habits and bureaucratic barriers are still holding many European firms back, there are some positive signs of more rapid industrial AI adoption in recent months.
Capgemini recently wrote about how multi-agent systems are starting to transform factory operations and allow for greater modularisation and better coordination, comparing this shift to the rise of specialisation and standardisation during the first industrial revolution, which enabled mass production.
McKinsey suggest that geopolitical supply chain disruption will lead to a rise in remanufacturing to fill some gaps, and that AI can enhance this process at every stage including forecasting and price optimisation.
Mercedes-Benz, which established its Berlin-Marienfelde Digital Factory Campus (MBDFC) as the global centre of expertise for production digitalisation, are now trialling humanoid cobots alongside industrial robots and other automation systems as a way to scale production more effectively.
Diginomica recently shared some basic advice on scaling industrial AI and escaping the proof of concept (POC) trap, suggesting that data readiness, people readiness, and change management challenges were holding back some firms from adoption at scale.
We have written previously about the need for more focus on AI readiness, in particular focusing on data platforms and the shift from process-centric to service-centric ways of working. Too many firms want to go straight to POC and then question why they cannot scale them if they work well locally. They need to eat their greens if they want dessert, but that is difficult, often thankless work that does not win plaudits in C-suite presentations compared to AI enabled crowd-pleasing magic tricks.
There are also challenges around how firms are organised and integrated.
Large firms have traditionally operated in a very siloed structure, with supply chains that do a good job of sourcing components to be integrated by the end manufacturer in times of stability, but can be a barrier to agility and innovation during times of rapid change.
Smaller firms, especially the famous Mittelständler firms in Germany who supply larger manufacturers, often have strong technical capabilities, but lack an operating model that encourages them to explore adjacent areas or innovate rapidly.
This story from the Financial Times is a good example of the challenge Europe faces, painting a picture of advanced missile manufacture as a kind of cottage industry for one French company that suddenly faces huge demand but lacks an architecture that can scale.
There is such a variable pace of change in each of the component layers that make up a product that we need a smarter approach to modularity, service and component standardisation, and the unbundling and re-bundling of what a product ‘is’.
The world’s best precision engineering firms might buy in embedded chips and software necessary for their product to work, but they risk repeating the same mistake as the German automotive sector in not recognising that “everything is computer” now, as one orange-faced criminal famously put it. Each layer of a product stack needs to evolve, and that means each service layer of the firms that make them also needs to evolve towards a lateral platform architecture, rather than the old vertical alignment approach.
However, it is not just products that need a modular, pace layered approach to re-bundling — it is also the firms themselves and, in the case of industry, the wider supply chain ecosystem as well.
There has been much talk of ensuring supply chain resilience in key areas such as energy and defence, and this is important for known key components; but a rich, evolving ecosystem of industrial innovation can ultimately be more resilient and less brittle than vertical supply chains.
Ecosystems & Platforms are more adaptable than vertically-integrated supply chains
If we think of the Haier ecosystem model as an example, with lots of entrepreneurial micro-enterprises engaging in a kind of co-opetition that can lead to ad-hoc clusters of firms working together on bigger products or innovation projects, then how can we create same kind of dynamism as part of Europe’s industrial renaissance?
In addition to flexibility of structure among the firms that make up the industrial ecosystem, which is key to creating the agility to identify and seize opportunities, there are other factors that can influence whether such an ecosystem can develop, such as:
Investment and incentives — how can we create patient capital and the right incentives to encourage a new generation of industrial innovators to flourish, and not just seek to have them acquired by aging behemoths as a route to scale?
Bureaucratic and regulatory barriers that make sense individually but stifle innovation collectively — a particular challenge for the EU
Standardisation of components and inter-operability standards — we need these for some products (e.g. in defence), data, software and also for agentic AI, which is why Anthropic’s MCP and Google’s A2A initiatives are welcome
Visibility of what products, components and sub-systems are emerging that others might be able to build on or bundle — this is where the technique of ecosystem mapping that we wrote about last week can be transformative.
They say necessity is the mother of invention, and so it is no surprise that Ukraine has demonstrated the greatest progress of any European country towards creating this kind of industrial ecosystem to boost its defence production in the face of invasion.
So many small groups of drone developers and equipment hackers have created and tested breakthrough innovations that are now being studied by militaries around the world. After relying on the adaptation of commercial drones since the start of the war, the first 1000 locally-made First Person View (FPV) drones came off the production line last month, and at lower cost than even Chinese-imported models. And elsewhere, the country’s rich industrial heritage has enabled it to develop long-distance missiles, long-distance conventional drones, and famously some very clever naval drones that they hope will enable their survival even in the face of a newly-hostile United States government.
This kind of ecosystem demands speed, openness to experiment, and great networks that mean successful innovations can be adopted and scaled with government support, and the Ukrainian experience holds many lessons for Europe’s industrial renaissance.
A digital twin for industrial ecosystems?
Doing this within a single country is difficult enough, but coordinating across more than two dozen states with different relative strengths and capabilities is even harder.
This is where AI will have some incredible supply chain discovery applications — by enabling semi-automated capability mapping to uncover opportunities for combinatorial innovation across hardware, software, data and industrial production — and potentially across sectors and even countries. Imagine if we could create a digital twin of a complex industrial ecosystem to inform industrial policy, investment decisions and the planning of scaled production.
This could also help us identify and address missing capabilities at the ecosystem level, such as smart coordination and interoperability systems that enable products from multiple sources to work together either in a factory or perhaps even on the battlefield, if it comes to that.
Top-down supply chain design and planning was the Twentieth-Century approach to this, but we need something smarter in a period of such rapid innovation against a backdrop of increasing economic and geopolitical risks.
In a way, Chinese industrial policy has been a mix of top-down management approach and an evolutionary ecosystem approach, as this long essay from Conrad Bastable argues:
In the battery sector, China founded multiple manufacturers (BYD, CATL, EVE Energy, CALB and more) in the 10 year window between 1995 and 2005, supported by state subsidies and intense local competition. Instead of one big Electric Vehicle maker, China went upstream and built the largest companies in every industry that provides key Electric Vehicle components to manufacturers downstream.
Recognising the opportunity, they implemented industrial policy that has slowly assured Chinese manufacturers a dominant position in a key sector of the global economy, and along the way created a rich industrial ecosystem that uses over-supply to maintain very high levels of competition and therefore innovation. And by working with more advanced electric vehicle OEMs, they ... let’s say absorbed through osmosis … the key knowledge required to make their own vehicle companies, rather than just supply components.
Using state capacity to direct Capital and Labor into foundational industry and then clear regulatory and tax burdens for the new firms is the first move in the multi-decade play. Cultivating a fleet of national EV makers and letting them battle domestically for market share instead of choosing a winner by diktat is the next.
The redundancy in manufacturers at this stage is a feature, not a bug: a dynamic industrial base with multiple players keeps the market agile and forces firms to push for more innovation. You’re not picking winners and losers like a failed Soviet imitation of Capitalism.
In Conrad’s view, this approach was more similar to other periods of abundant growth, such as Victorian England, mid-century America and post-war West Germany than it was to Europe and America’s recent “post-industrial” phase.
Europe is well-placed to accelerate its industrial renaissance and transformation if it can learn some of these lessons and if it can agree and execute on long-term industrial strategy. AI can help at the product level and the production level, but it could also provide much-needed visibility and supply chain intelligence that helps us with the kind of breakthrough combinatorial innovation that we saw with Johannes Gutenberg, the invention of the combustion engine and more recently Ukraine’s wartime innovation and industrial development.
Want to lean more about this approach? Get in touch to see how our mapping approach can help you discover new service combinations, engage partners more strategically, and build capabilities that extend beyond your firm.