At General Catalyst, we believe we’re at the forefront of a new industrial era. Geopolitical volatility, supply chain fragility, and aging infrastructure have exposed critical vulnerabilities in the systems that power our physical world. As we laid out in our Industrial Resilience thesis last year, the stakes are high. From energy to manufacturing, transportation to defense, we believe rebuilding our industrial base for global resilience must be a priority for us and our allies.

As great as this challenge is, it also presents a generational opportunity. This new era will be marked by a fundamental change, with the digital world catching up to the complexity of industrial environments. At the center of this shift is what we call software for hardware: an emerging category of purpose-built software tools and systems designed from the ground up for the challenges of the physical world.    

This is more than stacking new software onto old hardware. An emerging generation of hardware teams is adopting the principles that transformed software over the last two decades, creating a new end-product: software-defined hardware. These hardware platforms are designed by modern software, modular from the start, enable rapid iteration, and embed real-time intelligence as a foundational framework. 

Companies like Anduril, Applied Intuition, Helsing, and Saronic are early examples of what we believe is possible, but they are just the beginning. Modernizing critical infrastructure demands more entrepreneurs rebuilding the industrial stack from the very foundation, blending software-first design with real-world deployment. From the factory floor to the ocean’s surface, we are all in on a new class of software-defined platforms that have the potential to bridge the digital and physical worlds for mission-critical applications.

Software Founders for the Physical World

Hardware has always been a hard game. Long development cycles, high upfront costs, and rigid processes meant that even small mistakes could set teams back months or kill a project entirely. That reality bred conservatism. But a new generation of builders is changing the equation.

These founders are fluent in both physics and Python. They're bringing agile, iterative mindsets into industries that have long relied on fixed processes and hand-built systems. What’s emerging is a new kind of team: engineers who understand the messiness of physical systems paired with AI-native developers who, together, build tools that actually work in industrial settings.

We see this shift in motion at Senra Systems. Wire harness manufacturing (the dense electrical networks that power everything from cars to planes) has historically been a slow, manual process that hasn’t changed since the Cold War. Senra is flipping the model by re-engineering the entire workflow, combining automation, traceability, and intelligent routing into one integrated software-defined system. It’s not a digital layer on top of the old world. This is a redefinition of what the wire harness process should be.

Another critical challenge in building modern hardware is real-time, structured data from physical environments. Just as in the digital space, robust data in industrial environments allows for tracking, automation, and process improvements. Data-driven feedback is crucial: timing, precision, and safety are essential in these domains. But signal is often buried among messy inputs from a multitude of sensors, machines, and human activity, most of which doesn’t integrate well. The result? Heaps of unusable data.

That’s where Nominal comes in. They’re building tools to collect, structure, and activate test data collected in the field, giving operators and engineers visibility into how complex systems are performing in real time. Rather than porting software over from traditional enterprise data infrastructure, Nominal is building infrastructure specifically designed for the rigors and complexities of mission-critical engineering at factories, test sites, and labs. These tools enable operational velocity for a new generation of hardware companies that are iterative, data-driven, and automated.

We believe this combination of domain fluency, software-first thinking, and deep respect for real-world complexity will define the next generation of industrial software.

An Emerging Software-Hardware Stack

Senra and Nominal are early examples of a broader shift. We are now witnessing the emergence of a new category of applied software designed specifically for hardware-intensive industries.

Much of the existing tooling, including simulation platforms, test and measurement systems, product lifecycle management (PLM) software, and parts of electronic design automation (EDA), was built decades ago and has failed to keep pace with modern workflows. Tools like Ansys, National Instruments, and legacy PLM were not built for today’s realities: AI-native teams, vertically integrated manufacturing, and the speed at which hardware companies now operate.

PhysicsX is leveraging AI to redefine what’s possible in engineering and manufacturing. By delivering an AI-native engineering stack, they can significantly compress time to market, enhance performance, reduce waste and unlock new categories of machines and processes. PhysicsX is already re-defining key workflows at some of the most sophisticated engineering organizations in the world. We are also very excited about opportunities to transform the electronics design process with AI to unlock speed and efficiency.  

Beyond the engineering software stack, we need engineering and manufacturing solutions that can transform critical infrastructure. Chariot Defense is a prime example. By building software-defined hardware that fundamentally changes how power is delivered and distributed on the modern battlefield, they are providing a second life for traditional systems in the defense sector and enabling modern defense innovation. And companies like Re:Build, which is creating a new industrial platform to improve our advanced manufacturing capabilities, offer a window into a future redefined from first principles. With nearly forty percent of its team composed of engineers, Re:Build sits at the intersection of modern human-centric design, software-driven engineering, traditional manufacturing and digital manufacturing. 

Defining the next generation software for hardware stack can do far more than just modernize legacy systems; it can unlock entirely new markets. Autonomy, advanced robotics, and next-gen manufacturing workflows require new simulation layers, real-time data tooling, and developer-friendly infrastructure. As these systems mature, we believe they’ll open the door to major efficiency gains in physically intense sectors like services, agriculture, and food production.

The New Investor Toolkit for a New Industrial Era

Modernizing the industrial stack won’t be as easy as building traditional software. Investors will need to understand hardware development cycles and assess audacity of ambition against design pragmatism. Doing so requires knowledge of integration velocity, operational uptime, and deployment scalability across real-world environments. It also means understanding the capital requirements that come with building in the physical world, with creativity on both financing and business models.

At the same time, the operational business metrics that matter must expand. No longer is it just about tracking simple SaaS metrics. Founders and investors need to measure iteration pace, evaluate the quality and accuracy of testing metrics, understand R&D efficiency, and track total lifetime value of a system fully-burdened for cost. Most importantly, confidence in real-world performance, commitment to responsible innovation and the ethics of building for critical industries, and depth of customer trust will be paramount. 

Transforming industries like defense, industrials, and energy also requires a high level of engagement with public policy and industry incumbents. Companies, founders, and investors alike will need to be adept at navigating complex policy and regulatory environments, placing emphasis on uncommon partnerships between government and industry leaders.

We believe that the physical world is the next great software frontier, but this is not a place for overnight wins. Building in this space demands technical depth, perseverance, and sector fluency. True success will require patient capital, collaboration between the public and private sectors, and conviction in assembling teams across disciplines. At General Catalyst, we are committed to backing the founders rewriting the source code for a new industrial era.