Adding AI to factories is not enough: Italy's industry needs a new vision.

Interview with Massimo Portincaso by Laura Magna.
Originally published in Italian in Corriere della Sera, 19 July 2026.

Translated with Claude. Read original article in Italian here

Massimo Portincaso, founder and CEO of Arsenale Bioyards: "The difference will not be between those who use artificial intelligence and those who do not, but between those who manage to rethink industry and set continuous evolution in motion."

We are not facing a new technological wave. We are at the beginning of a new industrial era. AI puts Italy's industrial system at a crossroads: this is no longer a question of software. Nor of automation. The real challenge of the next ten years, according to Massimo Portincaso, is far more radical: rethinking industry starting from artificial intelligence, rather than merely adding it to production structures designed for another century.

Founder and CEO of Arsenale Bioyards, the startup working to industrialise biomanufacturing through integrated hardware, software and artificial intelligence platforms, and previously partner and managing director at BCG, Portincaso is among the more original voices in the debate on industrial transformation. His company works precisely on the border between the physical and digital worlds: it uses precision fermentation, data and AI to program yeasts, bacteria and other microorganisms so that they produce bioidentical molecules for the food, pharmaceutical and materials industries.

How the West is rebuilding its industry (and what it gets wrong)

From this perspective comes the reflection he has developed in recent months on his Substack, Neo Industrial, through essays such as The Neo Industrial Age, The Generative Phenotype and The Inference Machine. Three texts which, read together, outline a precise thesis: the West is witnessing the rebuilding of its own industrial substrate, and artificial intelligence is not simply a new tool but the principle around which the companies of the future will be designed.

But there is a serious risk: Western industry today still operates on an infrastructure born between the end of the nineteenth century and the beginning of the twentieth. In the meantime, China has entered the new millennium and is racing ahead, leaving us irretrievably behind. If we do not change pace, the concrete and not too distant risk is industrial desertification, final and irreversible. An unforgivable loss, one that would essentially wipe out made in Italy.

From computerisation to intelligence

"What we have done over the past decades," Portincaso explains, "is add successive layers to that industrial substrate. First computerisation, then digital, finally artificial intelligence. But this logic of layers has a limit. Once you have added intelligence, there is no further level to superimpose. That is why today we are going back to the base. We are remaking the industrial substrate itself."

In his view, this passage marks a historic break comparable to the great industrial revolutions. For over a century, machines, factories and processes were designed starting from the physical component. Software, sensors and automation arrived only later. Today, Portincaso argues, the paradigm is inverting.

"We are no longer building machines and then adding intelligence. We are starting to design everything around intelligence. It is an enormous difference. Even machines built just five years ago are often already limited against the requirements for sensors, data collection and actuation needed to truly exploit AI in the physical world."

When intelligence steps out of the screen

The heart of the transformation is what Portincaso defines as the passage from cognitive AI to physical AI. Until now, artificial intelligence has operated mainly in the domain of information: analysing data, generating content, synthesising knowledge. Now it is beginning to interact directly with physical reality. "Until now, AI observed and interpreted. Now it is starting to act. This is where the change becomes historic. Factories are beginning to develop a form of operational intelligence. Not a soul, naturally, but a capacity for understanding and adaptation that did not exist before."

According to Portincaso, it is precisely this convergence between artificial intelligence and manufacturing that is redefining the very concept of the industrial company.

From digital twin to digital original

One of the most recurrent ideas on his Substack concerns moving beyond the concept of the digital twin, the digital copy of a physical object. For years, the digital twin was considered one of the more advanced tools of industry 4.0. For Portincaso, it belongs to a logic that has been overtaken.

"The digital twin presupposes that the physical object exists first and that the digital is a replica of it. In the new paradigm, the opposite happens. You start from the digital. That is why I prefer to speak of a digital original. The physical object becomes the real twin of the digital original." This is not, he argues, simply a question of terminology. "If the heart of innovation moves into the digital, then the entire industrial process changes. First you imagine, simulate, optimise. Only then do you build."

This is not an abstract theory. At Arsenale Bioyards, a large part of development work already happens through simulations and digital models that allow configurations and processes to be tested before physical realisation. Portincaso also cites the case of Dallara, the Italian company and world leader in the design and production of racing cars, which has built an important part of its competitive advantage precisely on the capacity to simulate and optimise in the digital before building in the physical world.

The new speed of innovation

Speed is exactly where the global competitive game is being played. According to Portincaso, artificial intelligence is enormously accelerating the phases of design and learning, compressing the time needed to experiment with new solutions. "Today the bottleneck is no longer design. We can simulate thousands of configurations extremely fast. The problem becomes physics. Ordering a component takes months. Building a plant takes months. In the meantime, some Chinese companies manage ten iterations of the same process."

Hence the centrality of in silico innovation. "Instead of doing ten physical cycles in two years, you can do hundreds of digital iterations before even building the first real object. The real twin serves precisely to connect simulation and reality continuously, creating a permanent learning loop."

The lesson of China, leader in industrial innovation

This is also why Portincaso invites us to look more closely at China's industrial transformation. "The Western error is to keep thinking that China competes only on labour cost. That is no longer the case. Today the Chinese competitive advantage derives above all from the speed of industrial innovation." In his view, the difference is also cultural.

"In the United States, artificial intelligence is often seen as a goal in itself, like the race to AGI. In China, it is understood above all as a tool to transform manufacturing, supply chains and production. The dark factories, the lights-out factories where only robots managed by AI work, are born from this vision: they are not simply automated factories, but systems designed from the start to integrate artificial intelligence into the physical world."

Industrial rescues (in vain): the risk of defending the past

From this reading also comes a critique of traditional industrial policies. "The biggest risk today is to keep investing to prolong the life of industrial models that no longer have prospects. Not everything can be saved. We need to distinguish between those defending an exhausted paradigm and those building a new industrial metabolism."

For Portincaso, the question goes well beyond adopting a few AI applications. "This is not about using ChatGPT to write emails faster. It is about creating an industrial nervous system capable of learning, simulating, reacting and adapting continuously. Companies designed to evolve, and to do more than survive."

Why Italy can still play its hand

And yet, paradoxically, Italy could be among the countries best positioned to face this transition. The reason lies in the structure of its productive fabric.

"Many large European organisations risk behaving like dinosaurs: they have legacy systems, complex decision-making layers and timelines incompatible with the speed required today. An important part of Italian SMEs is instead more flexible, more specialised and potentially more transformable." This, however, requires a change of perspective in public policy as well.

"Many companies are still completing the transition to Industry 4.0 while the paradigm has already moved beyond it. The question is no longer how to digitalise existing processes. The question is how to design processes built from birth around artificial intelligence."

It is a challenge that concerns the country's industrial competitiveness, but also the way we imagine the future of manufacturing. For Portincaso, the real dividing line of the coming years will not run between those who use AI and those who do not. "The difference will be between organisations capable of evolving continuously and those that limit themselves to automating the past."