Introduction
On Episode 4 of Bits and Atoms, Elias Can sits down with Adam Versteegh, Co-Founder and CEO of Inframesh, to explore how advanced hardware and AI are transforming the way we understand energy. Adam shares the personal journey that led him to founding Inframesh, explains why most current solutions fall short, and how Inframesh flips the model with plug-and-play hardware that uncovers previously invisible energy patterns. The conversation dives into the technical and practical challenges of building ambitious hardware, the discipline required to reach true product excellence, and Adam’s philosophy on fundraising.
Interview
Elias Can: Hello and welcome to Bits and Atoms, the Grishin Robotics founder series. I'm your host, Elias Can, and today I've got Adam Versteegh, Co-Founder and CEO of Inframesh, joining me. To get started, Adam, would you mind sharing a quick intro about you and Inframesh?
Adam Versteegh: I'm Adam Versteegh, the founder and CEO of Inframesh. Since 2022, our mission has been to see energy clearly using advanced hardware and AI.
Elias Can: Let's jump into some questions, maybe a little more personal one to get started. What originally inspired or motivated you to take the leap to become a founder?
Adam Versteegh: I saw there was a gap between what the research showed was possible in high-frequency, non-intrusive load monitoring and the existing products. The technology could clearly do more but it hadn't been built, perhaps due to perceived issues around the device size, cost, power consumption, and the processing demands required to make sense of the data. But with recent developments in AI, including transformer architecture introduced in 2017, and more recently new NPU chips, neural processing unit chips, and the ubiquity of ADC's (analog to digital converters) and other improvements in the supporting components, it became possible to realize a fully capable system. The evidence in front of me made it clear that this was the most impactful thing that I could be working on by a long way, and once I started working through those constraints, it became clear that there weren't really any downsides. At that point, I stopped being optional. I just had to build it.
I've always been drawn to building electronics projects and tinkering, seeing ideas turn into real systems. Growing up I was constantly experimenting, and I was often interested in the problems that felt underexplored. I spent my summers in Sweden where a neighbor was a carpenter and tinkerer who built his entire house from scratch himself and even built his own car out of bike parts. That made me realize you can just build whatever you can dream of, and that’s always excited me. Later on I was mentored by leaders in high-end electronics, digital conversion, and signal processing, which set a clear benchmark for world-class products and inspired me to ultimately start my own electronics company. When I looked at the existing products on the market I felt there was a whole layer missing; the ability to remember each device as a distinct instance with its own states, conditions, behaviors all linked together in the universal product solution.
Working in commercial smart buildings, I saw many bespoke systems with high cost and complexity. What I’m building is completely opposite; the system only uses analog inputs, there’s no integration effort required, and the AI figures everything out from the clip-on sensors only sensing the analog. The hardest part is that the value can sometimes seem abstract at first. People often ask, does it control things? But my mission has never been about control, it’s always been about fully understanding energy signals. From the property management perspective, the value is clear, and my focus has always been about fully solving that problem first.
Elias Can: Very interesting, and it sounds like a very technical problem. Clearly you had some meaningful exposure growing up and working with people that were great at solving technical problems, and I’m sure that was a nice inspiration and lead-up for you to take that path yourself in terms of becoming a founder.
Elias Can: I know you touched on this a little bit in the intro, but to double-click, what would you say are some of the key pitfalls in energy monitoring solutions on the market today? How is Inframesh different?
Adam Versteegh: The biggest issues are scalability and practicality. Most systems aren’t truly plug-and-play, making it hard to install, expand, and maintain. That's particularly true for intrusive systems, which would never be commercially viable at scale anyway. Compared with existing meters and monitors, low-frequency sampling, cloud-only processing and limited insight means these products can't understand the real usage or come close to their full potential. They lack the intelligence, the AI at the edge, and much of the data collected is too slow to be useful. A huge amount of value is lost simply because the signals aren't captured and processed properly.
Elias Can: That totally makes sense, and sounds like a wide range of challenges, pitfalls, and shortcomings for the products on the market today.
Elias Can: That leads into an interesting question, when you're meeting with potential clients, what is the best value-proposition for the Inframesh Supermeter? What’s the aha moment for potential clients?
Adam Versteegh: The moment is when potential clients see their blind spots for the first time, when they see what they’re missing and the potential of our technology. The Supermeter delivers deep visibility without the friction; it’s inexpensive, easy to install, no power-consumption penalty, and scales cleanly. It can identify devices in seconds rather than weeks and months, and hundreds of devices per device. It works in any building regardless of size or complexity. The AI is genuinely useful, high-frequency, edge-based, and focuses on real understanding. This seems like it should be impossible, but when potential clients see that the reaction is more like why wouldn’t you want this?
What’s been particularly interesting is how broad the call has been. We've seen strong interest from insurance, where the device-level risks and behavioral patterns suddenly become visible in a way that wasn't possible before in industrial settings. The ability to see everything at the device level delivers the outsized value with minimal upfront outlay, especially for facilities that wouldn't normally adopt advanced manufacturing tech. One example is a shoe factory we're piloting with in the UK. They run basic systems and spend around £120,000 a year on electricity, they wouldn't be the typical customer for a complex industrial platform. They're quite small, but the Supermeter ties everything together for them and makes their entire facility AI-driven with almost no integration effort. It's a game changer.
We now have pilot sales lined up across homes, offices, factories, and universities spanning regions from the UK, Saudi Arabia, Estonia, Serbia, Poland, Qatar and more. Our university partners are especially excited about the living laboratory aspect, it gives them the real infrastructure to build their own AI lab and we collaborate with them directly on that work. The challenge has been to make something very advanced feel simple and obvious so people can trust it immediately even if the underlying concepts are complicated.
Elias Can: Sounds like a no-brainer for potential clients to sign-up, and interesting to hear about the diverse set of use-cases especially across different customers in different industries.
Elias Can: Going in a new direction, but still on the product side, what would you say has been the biggest hurdle or unexpected challenge of building in hardware so far?
Adam Versteegh: Time is definitely the biggest constraint. Designing and manufacturing a device with the high-frequency digital and analog signals together in a small form-factor is inherently complex. With this design, there's no room for error. One control or layout misstep can stop the entire system from working and set you back for months. The ambitious, no-compromise hardware carries real risk and you either do it properly or you don’t do it at all. In my view, doing it properly takes time.
Elias Can: Absolutely, that’s a common thread I’ve heard from other folks building in hardware. It takes time, and if there are issues, it’s not like you can easily fix it. It’s more intensive to fix an issue in hardware.
Elias Can: Switching gears to chat about the fundraising side. With such a unique and more technical product, what's been some of your better strategies for finding potential investors?
Adam Versteegh: It took a long time to narrow the focus of the pitch, but in terms of fundraising, I try to avoid traditional selling. As a product develops, it starts to speak for itself. The focus has always been on development, and the last 1 to 2% of polish is disproportionately expensive but that's the difference between something that's good enough and something that people actually want to buy. Very few people want to buy a product at 98%, but when it reaches that 100% there’s no remaining friction and that’s the standard we aim for. Customers and investors notice that level of care and intent. One thing I’d add is don’t be afraid to work with the top-tier financial and legal advisors, it doesn’t need to cost you something upfront. Sometimes VCs say don’t do that, you should do it all yourself. For us, it really helped us get that last extra few percent of polish and helped us to stay fully focused on the product when we were already stretched thin. This is an exciting AI product, and customers and investors naturally get excited about it. It’s my belief that it’s part of the AI revolution and our technology will be widespread in the buildings of the future for a more connected and efficient society. I think the dedication to detail, excellence, and fully solving the problem is what makes a difference. The product’s quality becomes a story instead of having to put anything else in there.
Elias Can: Amazing, I think you hit a good thread there about storytelling. From the fundraising side, there’s certainly a story that the founding team can tell about the company and their evolution, but the product will also tell a story on its own. Here you can have both and it’s important to have both tell a great story for you.
Elias Can: Thank you, Adam, for joining me on Bits and Atoms, it was great to have you today!
Adam Versteegh: Thank you so much for having me!
Inframesh is creating the intelligence layer for the world’s buildings. The Supermeter reveals any building’s electrical fingerprints in real time using state-of-the-art hardware and AI, providing a complete understanding of energy for the first time. Without complex installation, it turns raw signals into unprecedented depth, unlocking far greater value across countless new and existing use cases.
Grishin Robotics is a Silicon Valley-based VC firm focused on investing in early-stage companies in broader consumer categories. Grishin Robotics has invested in category-defining companies such as Ring (acquired by Amazon), Spin (acquired by Ford), Zipline, Starship, and many others.
