Recently, we shared our thesis on next-generation tools for building hardware. We believe there is a significant opportunity to make the future of hardware engineering look more like software development, with modern collaboration tools, copilots, and continuous testing enabling rapid product development. Today, we’re excited to announce our investment in Cadstrom, a company bringing this vision to life in an AI expert system for electrical engineers.
Electronics—especially circuit boards—are often seen as a necessary, but not strategic, part of hardware design. However, from EVs to robotics, industrial equipment to consumer devices, products are increasingly defined by their electronics. Circuit boards that once contained dozens of components now house thousands. Simultaneously, development cycles that used to span years are now compressed into months. This transformation is enabling incredible innovation, but it's also creating unprecedented complexity in hardware development.
Yet the tools and processes for designing and validating printed circuit boards (PCBs) haven't kept pace. Schematics and layouts are largely created manually in design software, and limited innovation has been seen in the past decades. And simulations to test these designs are used sparingly due to the complexity and time required to implement tools today. Simulations tend to be complex or not comprehensive since no single simulation can cover a whole system nor a plurality of its states. As a result, engineers rely heavily on manual reviews and physical prototypes to catch errors. Each prototype iteration ("respin") costs 2-8 weeks and can be tens of thousands of dollars. Worse, some issues only surface during production, where fixes can cost millions. A lot of expertise in PCB design remains tacit knowledge with a generation of electrical engineers reaching retirement, magnifying the need for better tools to enable a new generation of engineers.
These challenges are particularly acute in testing, often called verification and validation. Despite consuming over 10% of development time and cost, testing remains largely manual. Engineers painstakingly review designs, configure complex simulations, and hope to catch issues before committing to expensive prototypes. The process is slow, error-prone, and increasingly untenable as complexity grows.
Verification and validation has increasingly become a bottleneck in hardware development cycles, exactly when we need to be moving faster. Enter Cadstrom—Cadstrom aims to reduce respins, accelerating the time to performant hardware. Conceptually, Cadstrom enables continuous testing by automating the setup and execution of simulation. This goal is accomplished in a few AI-enabled steps:
Cadstrom’s approach embodies our thesis that the future of hardware development will increasingly rely on AI-powered tools to augment engineers. Rather than replacing existing workflows, Cadstrom slots seamlessly into current processes, enabling engineers to spend more time developing rather than bug-fixing. In speaking to early design partners and potential customers, we heard strong demand for shortened product development timelines and reduced engineering time on complex simulations.
Moreover, we believe Cadstrom is the right team to execute this vision. Margot Blouin, Cadstrom co-founder and CEO, brings incredible technical chops to the problem—she single-handedly built the entire MVP. She was a principal program manager at Microsoft, most recently for Azure’s Quantum Computing service, and a lifelong tinkerer. Scott Bright, Cadstrom co-founder and CPO, has over 13 years of experience as an executive in their initial customer segment and product design consultancies, having co-founded and led Synapse. Scott understands the customer pain point intimately and brings both incredibly valuable knowledge and a network to Cadstrom.
Cadstrom’s vision is ambitious. Automating PCB verification and validation would meaningfully move the needle on electronics development, but Cadstrom aims to do more. In time, the Cadstrom team aims to automate regulatory certification, scale to cross-disciplinary system verification (i.e., beyond individual boards), and eventually tackle trustworthy generative design. We are eager to see Cadstrom bring this vision to fruition.
We think there are myriad more opportunities here: from upstream tooling to surface business tradeoffs early on in the design cycle, to downstream tools to support the transition to manufacturing, to data infrastructure layers that better support intelligence and decision-making throughout a product journey. If you’re transforming how we build in the real world, we’d love to hear from you.
