Federico Martelli leads Forgis with a rare blend of strategy, hands-on engineering, and a geopolitical perspective, qualities born of studies in econometrics, international relations, and management, plus years in tech and strategy consulting. In this interview, he explains to Eqvista why Western manufacturing needs an “intelligence layer” above decades-old PLCs and heterogeneous hardware, how causal AI can eliminate microstops and predict failures in live production, and why re-industrializing Europe requires both technical depth and relentless execution.
Federico also walks through Forgis’s safety-first approach to agentic AI, the company’s early traction in automotive pilots, and how a recent $4.5M pre-seed raise will accelerate product development and enterprise adoption. Read on for a founder’s view of factories as living, adaptive systems and what it will take to keep production competitive in an East–West manufacturing race.
Federico,your path spans strategy, operations, and deep tech environments. How did that mix shape how you approach building a company?
Being a polymath allows me to build Forgis, while keeping all contributing forces in view, economic, social, political, and technological.
My background in tech consulting gave me a deep understanding of manufacturing systems and AI at an operational level. My time in strategy consulting taught me how to design a winning strategy and make hard trade-offs. Studying at the world’s best master’s in management program gave me the tools to structure and scale a growing organization.
My studies in international relations and sociology allow me to position Forgis within broader geopolitical and industrial power shifts while ensuring social impact is present and delivered. My first master in the UK in econometrics helps me understand the data foundations behind AI systems and manufacturing plants. That combination – strategy, operations, technology, and geopolitics – shapes how I sell, negotiate, build, and scale Forgis. I don’t see the company through a single lens; I build it by integrating all of them.
What drove you from observing manufacturing’s decline to founding Forgis and taking action?
I grew up watching Italy gradually lose the large manufacturers that once made it a global industrial leader. One by one, companies that defined our economic strength disappeared. Today, only a few remain.
Over the last five to ten years, I’ve seen the same pattern emerge in Germany. And that was the turning point for me. If Germany deindustrializes, Europe doesn’t just lose competitiveness – it permanently loses its ability to produce.
I believe in a strong, united, and integrated Europe, where member states support each other for the benefit of the whole. In that vision, a strong industrial Germany is not optional – it is foundational. Founding Forgis was my way of acting on that belief rather than just observing the decline.
Forgis positions itself as the intelligent layer for factories amid East-West competition in manufacturing. What core problem does your causal AI platform solve that legacy PLCs and robots cannot?
Forgis solves a problem that legacy PLCs and robots inherently cannot: the fragmentation of the factory hardware stack. Modern manufacturing plants are built from heterogeneous equipment – PLCs, robots, sensors, MES, SCADA – each operating in isolation, producing data that are not meaningfully connected. Without a shared semantic layer, data remains siloed. You cannot triangulate signals across systems, generate real understanding, or issue instructions that improve how the factory actually runs.
Our causal AI platform sits above this fragmented hardware layer and turns disconnected data streams into a coherent model of the factory. That semantic and causal understanding is what enables higher efficiency, lower downtime, better quality, and higher throughput. Without it, AI remains blind.
Today, legacy systems alone cause massive downtime, quality scrap, and throughput inefficiencies, amounting to trillions of dollars in lost value across global manufacturing (Siemens report: The True Cost of Downtime 2024).
This gap is also strategic. China benefits from both lower labor costs and more modern production hardware, having built much of its manufacturing base in the last twenty years rather than inheriting systems from the previous century. That gives them cleaner data, tighter integration, and an easier path to intelligence.
Forgis exists to close that gap, by upgrading the intelligence layer on top of Western factories, without replacing their physical assets.
In your automotive pilots with mixed vendors, what downtime and throughput improvements did you see? Can you walk through a real-world failure prediction example?
I can share some details of one example, although most of our deployments are under NDA. For a European machine builder, we eliminated recurring microstops caused by manual quality checks on dimensional accuracy. Traditionally, the machine had to stop every few hours to measure part dimensions and verify that tolerances were still being met.
With Forgis, the machine continuously inferred dimensional deviation in real time by correlating process signals such as motor current, vibration patterns, temperature drift, and cycle-time variance. The system could predict when parts were about to fall out of tolerance with very low error.
As a result, those inspection-driven microstops were completely removed. This enabled the machine builder to offer a clearly differentiated product and to retrofit the installed base, going back to existing customers with a measurable productivity and quality advantage.
Agentic AI raises concerns like goal misalignment or unpredictable inputs. What safeguards does Forgis use for safety in high-stakes manufacturing?
We don’t deploy agentic AI as an autonomous black box. In manufacturing, safety comes from constraint, structure, and control.
Forgis operates within clearly defined industrial boundaries: physical constraints, engineering rules, and validated process limits. AI agents don’t invent goals, they execute intent within a tightly specified operational envelope. Every action is bounded, observable, and reversible.
Second, we separate decision-making from execution. Agents propose changes, optimizations, or reconfigurations, but execution happens through deterministic, auditable layers that respect safety standards and plant-level approvals. Humans remain in the loop where stakes are high.
Third, the system is continuously validated against real production feedback. Unexpected inputs don’t propagate blindly, they are detected, flagged, and handled through fallback logic rather than cascading failure.
In short, Forgis doesn’t replace industrial discipline with AI. It embeds intelligence into it. Safety isn’t an add-on; it’s a first-class design constraint.
Who was Forgis’s first real buyer inside a manufacturing organization, and who typically blocks these deals?
Our first real buyers typically sit close to production: heads of digitalization, heads of R&D, plant managers, and in many cases automation engineers and operators themselves. In the last group, our users, we have a huge Slack community with 800 requests to join pending. So it is really a movement that goes both top-down and bottom-up. What usually blocks deals is not a specific persona. It’s the inertia of century-old manufacturing organizations.
Decisions require alignment across multiple divisions, IT, engineering, operations, procurement, and that coordination takes time. Forgis moves quickly; legacy organizations don’t. Bridging that gap is the real challenge, not convincing a single buyer.
With reshoring accelerating, how will Forgis expand beyond automotive to renewables or semiconductors, where supply chain volatility is extreme?
We start with the critical industries where Europe, and the West more broadly, is being challenged and losing ground. That is where we want to give the West an unfair advantage.
Regarding your question, Forgis is focused on solving a very specific problem: the lack of cutting-edge software on top of Western factories. That is our leverage point. Supply chain volatility, and the geopolitical or macroeconomic forces behind it, is largely outside our scope.
What we do control is how factories operate once constraints exist. By upgrading the software layer that runs production, we enable Western factories to be faster, more adaptive, and more competitive, regardless of external instability. That is where Forgis creates value, and where we choose to focus.
You recently raised $4.5M in a pre-seed funding round. Where does that capital create the most leverage for Forgis in the next 12–18 months?
The capital creates the most leverage in two areas: technology development and large enterprise customer acquisition.
On the technology side, it allows us to move faster than the market, building the industrial intelligence layer that makes factories adaptive, autonomous, and scalable. Speed here compounds: every capability we ship strengthens the platform and raises the barrier to entry.
On the customer side, it gives us direct access to large industrial players. Landing a small number of high-impact enterprise customers creates disproportionate leverage: real production environments, real constraints, and real scale. Those deployments accelerate learning, harden the product, and turn Forgis into infrastructure rather than software.
Together, these two levers, deep tech and enterprise adoption, reinforce each other and define our trajectory over the next 12–18 months.
You’ve advocated about work intensity, inspired by Jensen Huang. How do you sustain 105-hour weeks without burnout in Forgis’s high-stakes environment?
Europe is losing ground across multiple critical industries. Sectors where we led for over a century, exporting goods and importing value, are now under pressure from both the U.S. and China. When foundational industries are at risk, I believe leaders in those sectors have a responsibility to respond with exceptional intensity. This applies even more to startup founders, who face not only systemic decline but also the challenge of building a company from zero and establishing themselves as new incumbents.
How is that level of intensity sustainable? First, because I genuinely love what I do. The work is heavy, high-stakes, and comes with constant pressure but it is deeply meaningful to me. It doesn’t feel like something I have to do; it feels like something I choose to do. I’m not working, I’m the cause. Second, the time I invest aligns with a purpose I care deeply about: contributing to the re-industrialization of the West and bringing manufacturing capability back where it strategically belongs. When effort, identity, and mission are aligned, long hours don’t translate into burnout in the traditional sense, they translate into commitment.
Third, I don’t take a long time off, but I do take short, deliberate windows. A few hours each week fully dedicated to my hobbies and to the people I love are enough to recharge me completely.
In ten years, what factory capabilities will exist thanks to Forgis that would otherwise be impossible?
In ten years, a factory won’t be a fixed asset anymore. It will be a living system, plug-and-play, adaptive, and self-optimizing. Orders will move straight from digital demand into physical production, with AI agents translating intent into execution, reconfiguring machines, and tuning processes in real time. For the first time, corporate strategy will execute itself.
Change the strategy, and the factory changes with it. No retooling cycles, no months of delay, no friction between planning and reality. Humans won’t disappear from the factory floor, they’ll be elevated to higher quality tasks. Engineers and operators will act as conductors, not firefighters, focusing on quality, creativity, and innovation while the system runs in continuous orchestration beneath them.
As intelligence becomes embedded into factories, lowering prices purely thanks to cheap production costs – like China is doing now – will no longer be a viable strategy. Manufacturing will no longer be driven by low-cost labor. Cost efficiency will come from how effectively intelligence runs the factory. In this model, the manufacturers with the strongest industrial intelligence will achieve the lowest production costs. Competition between goods sellers will shift away from cost arbitrage and toward delivering the best product or executing the strongest marketing.
