top of page
RESOLV-Logo-Charcoal&Orange_Fill.png
Search

Hardware is the New Moat: A Post by Our CEO

  • Writer: Mike Ellenbogen
    Mike Ellenbogen
  • Jul 6
  • 4 min read

Software may be eating the world, but AI is eating software. 


According to a 2024 analysis of 211 million lines of code by GitClear, 41% of all code written globally is now AI-generated. The barrier to building software has never been lower. 


So if a competitor can ship a working product in days, where does durable competitive advantage actually live? I've been thinking about this a lot. My answer: proprietary hardware, sensors, and the data that only those sensors can generate.


Hardware is the new moat.


The Commoditization of Software

For most of the last two decades, software was the defensible asset. Building great software was hard - it required engineering talent, capital, and time. Those barriers created moats.


That era is ending fast. Today, you don't need to be an engineer to ship a software product. Bubble's 2025 State of Visual Development survey found that among people actively building apps, only 22.7% are professional developers. The rest are founders, non-technical business users, and complete beginners. The tooling has democratized to the point where anyone with a clear problem and an internet connection can build a functioning product.


The implication for competitive strategy is stark: if your only defensible asset is software, your moat is shrinking. A well-funded competitor - or an ambitious non-technical founder - can clone your app's core functionality with the same AI tools available to everyone. The software layer is becoming a commodity.


This isn't a reason to panic. It's a reason to think differently about where to build.


Why Hardware Creates Durable Moats

Yes, hardware is hard. The engineering cycles are long. Prototyping is expensive. Supply chains are brutal. I've lived it across multiple companies. But here's what I've come to believe: that difficulty is precisely the point.


The complexity of hardware development is the moat. It is the thing that keeps competitors out. And once hardware is deployed in the field, the dynamics shift dramatically in your favor.


Defensible, recurring revenue. A software subscription built on top of proprietary hardware is fundamentally stickier than software alone. The hardware creates a continuous reason to stay - and a natural platform for expanding value over time.


Unshakable customer retention. Once a physical tool is embedded into daily operations - purchased, trained on, integrated into workflow - switching costs become prohibitively high. 


Total quality control. When you control the physical device, you own the entire end-to-end user experience. You're not at the mercy of a third-party sensor, a browser update, or an API deprecation. You define what's possible.


A software-only competitor can clone an app overnight. They cannot clone hardware built on years of sensor engineering, field iteration, and real-world deployment.


The Resolv Thesis

This thinking is central to why we built Resolv the way we did.


The construction and trades industry is one of the last major sectors where critical decisions are still made by guesswork. Autodesk and FMI research shows that rework alone accounts for roughly 9% of total project costs, and that 52% of that rework stems directly from poor project data and miscommunication. Electricians, plumbers, GCs, and remediators have been working blind for generations, and paying for it.


We built the InSite Pro wall imager to change that. It puts best-in-class sensors directly in the hands of trade professionals. Real-time images of what's hidden behind every wall, ceiling, and drywall surface. No guesswork. No costly surprises. No invasive demolition.


But I want to be clear: the hardware is just the beginning.


The Data Nobody Else Can Replicate

Every scan of the InSite Pro generates something no AI startup or software competitor can replicate: real-world data from inside the built environment. Structural details. MEP layouts. Hidden defects. Conditions that have never existed in any training dataset because nobody has ever had a way to capture them at scale, until now.


The demand for this kind of structured, physical-world data is enormous. The global digital twin market is projected to reach $149.81 billion by 2030, growing at a CAGR of nearly 48%. The fundamental bottleneck to that growth isn't compute or AI capability. It's data. Specifically, accurate, structured data about the physical world that doesn't yet exist in clean digital form.


We are on a mission to build that dataset across thousands of real structures. And over time, that proprietary data becomes an asset that compounds. It trains smarter AI models. It powers better tools and more accurate predictions. It deepens the value we deliver to every customer. And it widens the gap between us and anyone trying to compete without access to it.


Software can be replicated overnight. A decade of proprietary scan data from inside the built environment cannot.


Where We Go From Here

We are still early. The InSite Pro is approaching its commercial launch, and we're building the team and infrastructure to scale. But the thesis has never felt clearer to me.


The companies that will win the next decade won't be the ones that wrote the best software. They'll be the ones that paired software with defensible physical assets - sensors, devices, data pipelines - that no one can replicate from a laptop.


Hardware is hard. That's the point. And at Resolv, we've chosen to do the hard thing because we believe it's the only thing that leads somewhere nobody else can follow.


Sources


bottom of page