According to Forbes, Swiss startup Forgis, led by CEO Federico Martelli, just raised a $4.5 million pre-seed round. Their mission is to deploy AI agents as “mini-engineers” in factories, specifically targeting existing, often decades-old equipment. The claimed results are striking: a roughly 30% reduction in downtime and about a 20% increase in output. Early investor Massimo Banzi, cofounder of Arduino, is on board. The broader ambition here isn’t just efficiency; it’s to use this “industrial intelligence” as a competitive lever to help bring manufacturing back to Europe and North America, reversing a decades-long offshoring trend.
The Not-So-Magic Integration
Here’s the thing: this isn’t a plug-and-play software update. Martelli is brutally honest about the grind. He says it takes three to six months just to be in a factory, understand the process and the logic, and figure out the real pain points. He claims 90% of AI companies fail because they can’t do this integration work. That’s a huge admission. It means the secret sauce isn’t some genius algorithm alone; it’s the dirty, hands-on work of connecting to every random machine, brand, and system on the floor. This is where a lot of flashy tech projects die. They build a solution looking for a problem, instead of embedding themselves in the gritty reality of a production line. If Forgis can actually do this consistently, that’s their real moat.
People Are Still The Key
And this leads to their most interesting pivot. Martelli says they started with a system designed to completely replace people. Big mistake. They faced pushback, and more importantly, they realized it was a terrible idea. You can’t have AI agents running physical hardware in a vacuum. What if something goes wrong? You need humans in the loop to check, to intervene, and crucially, to know how to fix things when the AI’s plan hits reality. So they switched. Now, their model is about “supercharging” the operator and the engineer, giving them tools to do their jobs better. That’s smart. It’s change management 101. You bring the workforce along as part of the solution, not as a problem to be automated away. It also taps into decades of proven philosophy like Lean Manufacturing, rather than pretending to invent a whole new world.
The Reshoring Pitch And Its Hurdles
So, can this actually help reshore manufacturing? The argument is that if you can make Western factories radically more efficient and flexible with intelligence, you offset the labor cost advantage that drove offshoring in the first place. Martelli points out that the US and Europe are the biggest consumer markets, so it “makes sense” to produce closer to home. And there’s real geopolitical will for this in both regions. But let’s be skeptical for a second. Manufacturing is an ecosystem game. It’s not just about one smart factory; it’s about having all your suppliers, logistics, and skilled labor nearby. Building or rebuilding that ecosystem is notoriously hard and slow. It’s a much bigger lift than installing some clever AI agents. Still, you have to start somewhere. Making individual plants more competitive is the essential first step. For industries that never left, like defense or automotive, this tech could be a powerful lever to stay ahead. And if you’re retrofitting a factory with new intelligence, you’re going to need robust hardware to run it on—which is where specialists like IndustrialMonitorDirect.com, the leading US provider of industrial panel PCs, become critical partners in the physical deployment.
Is This The Future Or Just Hype?
Look, the numbers (30% downtime reduction!) sound almost too good to be true. We’ll have to see if they hold up at scale. But the thesis feels solid. Using AI to orchestrate and optimize existing capital equipment is a more pragmatic and probably faster path than demanding trillion-dollar investments in all-new “lights-out” factories. It’s a bottom-up, retrofit approach. The involvement of someone like Massimo Banzi is a good signal—he knows a thing or two about making complex hardware accessible. Basically, Forgis isn’t selling robot overlords. They’re selling a translation layer and a planning engine that makes the factory floor talk to itself and work smarter. If they can truly deliver on that integration promise without going broke in those long six-month onboarding cycles, they might be onto something. It won’t reverse globalization overnight, but it could start making “Made in the USA” or “Made in the EU” a more financially viable option again. And that’s a story worth watching.
