From hand-drawn sketches and scrapped prototypes to powerful AI-driven simulations at lightning speed, hardware product development has come a long way. As former hardware engineers ourselves, we know firsthand the grueling processes that have historically stymied innovation. Today, we’re betting on a future where AI and cloud computing are game-changers for hardware engineers.
Before the 1970s, the bulk of an engineering workflow involved isolated or loosely coordinated designs, with large sketch papers dominating massive desks. With each iteration, engineers often had to restart from scratch, leading to slow, ad hoc, and painstakingly repetitive cycles. To take designs from concept to production, every measurement had to be meticulously documented and passed on for collaboration with other engineers or manufacturing teams. And simulation? It was mostly done by building out prototypes, which were discarded after each test. It was a frustratingly slow process.
The 1970s and 1980s brought a wave of new software tools that fundamentally changed this process. Foundational companies like Autodesk, Ansys, PTC, and Dassault emerged, transforming engineering workflows and sparking a wave of rapid innovation. Despite these advancements, the past 50 years have seen only incremental software improvements, with today’s engineers still relying on tools their predecessors used decades ago. As software began to “eat the world” and tools to increase software productivity proliferated, hardware engineering was largely left behind.
Over the past two decades, SaaS investments have fueled an entire generation of category-dominating tools. Yet in hardware engineering, little capital has flowed into challenging incumbents. Nine of the largest 100 software companies (Synopsys, Cadence, Autodesk, Dassault, Ansys, PTC, Aspen Tech, Bentley Systems, and Altair) are tied to hardware engineering, together representing $350B in market cap and generating over $30B in annual revenues. Recent acquisitions (Ansys to Synopsys for $35B and Altair to Siemens for $10B) underscore this space’s value but there are significant challenges for new entrants.
Historically, incumbents had little incentive to innovate, buoyed by entrenched agreements, high switching costs, and a lack of competition. However, demand is now shifting, driven largely by the latest generation of engineers who are accustomed to modern software workflows and expect cloud-native solutions that deliver lower costs, flexible access from any device (especially critical in today’s hybrid work environment), and enhanced collaboration. While cloud-based licensing has gained traction, it only partially addresses the challenges of data integration across systems – an essential factor for achieving seamless workflows. Despite this shift, many existing platforms continue to operate in silos, limiting collaboration and efficiency. Meanwhile, AI is beginning to tackle hardware engineering’s most persistent obstacles, from design generation to large-scale optimization, enabling advances that were previously computationally unattainable. But to truly capture this moment, these paradigm shifts call for new platforms built with native functionality, not legacy add-ons, creating a unique opportunity for emerging players to reshape the field.
Setting the stage
Hardware engineering spans a vast range of disciplines—from the molecular precision of chemistry and biology to the structural projects of architects and civil engineers building bridges and skyscrapers. Despite this diversity, these fields share a similar process, highlighting the expansive potential and complexity within hardware engineering. For our purposes, we focus primarily on mechanical engineering and electrical engineering, with our discussion organized around four pivotal segments that are key to advancing these fields: design, collaboration, simulation/testing, and manufacturing.
Design: Redefining hardware design with AI-driven tools
Design is foundational to hardware engineering, often called “the drawing board” stage. Although this phase has evolved from physical sketches to sophisticated software models, its core remains a visual manifestation of ideas. Mechanical engineers, many familiar with CAD since university, have long relied on major platforms like Solidworks and AutoCAD, whose interfaces and functionalities have evolved only marginally since the 1990s. The result? Cumbersome, creativity-stifling tools that hinder the modern design process.
Today, AI is injecting new life into hardware design, balancing automation with human creativity in promising ways. Human-in-the-loop systems let engineers retain creative control, a crucial aspect for those who value the artistry of engineering. Text-to-CAD tools from startups like Cadify and Leo are building solutions that allow engineers to describe designs in natural language, with the promise of generating workable CAD models. Other platforms, such as Hestus, automate tedious tasks like applying constraints to lines and arcs, which could further boost engineering efficiency. Stealth companies like Backflip hold promise for delivering seamless, end-to-end design workflows.
But these new solutions face challenges. Design tools often serve as “anchor tenants” in an engineering workflow, deeply integrated and resistant to change. New platforms must either integrate seamlessly with legacy systems or provide a compelling enough value to overcome incumbents’ entrenched positions. Solutions acting as “copilots” to augment legacy systems may be the right near-term position to build customer usage but may need a long-term strategy to gradually displace established players over time.
Collaboration: Transforming engineering communication with seamless integration & data management
After design, collaboration is essential, but the limitation of current tools often hinder this process. Engineers frequently find themselves scrambling to correct outdated file versions, triggering a confusing chain of emails and file names like “vFFF_AbsoluteFinal.” Traditional collaboration — driven by email exchanges and limited version control — underscores a critical need for innovation. Just as tools like Google Docs and Figma have transformed collaborative work, hardware engineers are beginning to expect the same cloud-based experience.
We see two key categories being created in the collaboration layer: data integration, which ensures seamless data flow across systems, and team collaboration, which enables efficient project teamwork.
Hardware engineering collaboration tools are tackling the need for streamlined communication and accelerated iterations by focusing on the data integration layer. Traditional tools like email and Slack are often inadequate, as feedback can easily get buried among unrelated messages. Companies such as Flow Engineering, Integrate, Violet Labs and Prewitt Ridge are building solutions that consolidate inputs across platforms (e.g., Office, CAD, Solidworks, ERPs), creating an easily accessible, unified source of truth that prevents data from getting lost in siloed environments. Others like Quarter20 are automating the documentation process, reducing the time engineers spend on this tedious yet essential task. These advancements not only safeguard the creative and technical integrity of projects but also foster a more dynamic and efficient collaboration environment.
Collaboration tooling is one of the most promising areas for boosting efficiency in hardware engineering without requiring teams to abandon their existing platforms or overhaul workflows. In the SaaS world, collaboration tools transformed productivity a decade ago, with document repositories (Box, Dropbox), software versioning (Github), communication tools (Slack) and project management systems (Asana) becoming indispensable. But collaboration in CAD brings unique challenges – MCAD and ECAD files aren’t simple “black boxes” that can be passed around like basic documents. They contain intricate layers of data, from individual parts and subassemblies to interconnected components, each requiring specific access, versioning, and collaborative input.
Today’s emerging solutions are addressing CAD complexities directly, enabling engineers to collaborate on modular components within larger assemblies. Colab, for example, offers a platform much like Google Drive but specifically for CAD files, allowing teams to manage project details, assign review responsibilities, and set timelines in one cohesive platform. Meanwhile, Duro’s PDM One integrates with existing CAD tools to maintain a single, synchronized master BOM, enabling file checkouts and locks during edits – similar to document management systems like Box (shoutout to a former Scale port-co).
This modular approach not only enhances real collaboration within CAD but also integrates smoothly with broader systems like requirements management platforms, made possible by these new data integration capabilities. By linking MCAD and ECAD data with requirements tracking and other project tools, these platforms ensure that data remains accessible and current, streamlining the entire development process. This results in faster, more flexible, and fully synchronized hardware product development across every stage.
Simulation / testing: Accelerating innovation with AI-driven tooling
In today’s innovation-driven market, the demand for faster, more accurate simulation and testing capabilities has never been higher. Industries ranging from aerospace to consumer electronics need to test complex designs quickly without sacrificing quality. Traditional simulation methods, dependent on physics solvers and finite element methods, are becoming increasingly difficult to use as Moore’s Law reaches its limitations and computation costs continue to climb. Companies are redefining these processes by leveraging AI-powered models that dramatically reduce the time and cost of CFD simulations. By encoding physical constraints directly into their models, Navier claims to enable engineers to iterate a thousand times faster than traditional approaches – signaling a massive productivity leap.
Real-world complexity further underscores the importance of AI-driven simulation. For instance, Archetype AI uses foundational models to interpret sensor data and uncover underlying physical laws without explicit human guidance. This could allow for impact in sectors where accurately predicting system behavior can mean the difference between seamless operations and costly downtime. Additionally, Arena AI’s “SimCity for the real world” concept leverages transformer models to simulate and anticipate complex behavioral patterns, which could help train specialized AI agents for roles like inventory management or revenue optimization.
Others are addressing the cumbersome and costly simulations for high-stakes industries such as aerospace. DeepSim’s advanced 3D physics simulation platform automates the setup and execution of simulations which could be a scalable solution for intricate designs in industries such as semiconductor manufacturing. Meanwhile, Elodin focuses on facilitating massive Monte Carlo simulations for the drone and UAV industry, testing scenarios ranging from turbulence to collisions. This would not only accelerate the testing process but also bring improvements to reliability and resilience of autonomous systems, paving the way for safer, more capable hardware in a rapidly evolving technological landscape.
These innovations highlight why AI-driven simulation and testing are critical today: they enable organizations to iterate, innovate, and bring products to market with higher levels of speed and confidence.
Manufacturing: streamlining processes with AI-enhanced production tools
Once designs pass prototyping and testing, manufacturing is the final step, which involves sharing requirements with production teams, optimizing part procurement, and managing paperwork and supplier communications. Although it may seem straightforward, transforming designs into manufacturing-ready formats is complex, often requiring manual tasks like creating 2D slices with geometric specifications. Companies like DraftAid and Drafter leverage AI to automate this process, which could improve accuracy and further save engineers time. Others like Dirac take in CAD files and generate interactive work instructions allowing manufacturing and assembly teams to better understand the manufacturing requirements, which would represent a significant improvement over the standard static, difficult-to-read PDFs used today.
In electrical engineering, AI is revolutionizing PCB layout, speeding up what was once a slow and labor-intensive process. For example, Quilter is pioneering a “compiler for circuit boards” that translates schematics into manufacturable designs, and using AI to autonomously handle placement, routing, and stackups while accounting for complex physics like crosstalk and electromagnetic interference.
Closing thoughts
The landscape of hardware engineering tools is evolving rapidly, driven by cloud and AI advancements. While Silicon Valley often overlooks hardware in favor of software, the computers we work on, the cars we drive, and the spaces we inhabit all rely on hardware built through complex engineering. Today’s startups have an opportunity to upend incumbents and introduce new perspectives on the hardware engineering development lifecycle. This is a space where we expect several large outcomes, mirroring the first generation of incumbents. Though these new players have a long journey ahead, the potential in this domain is immense and we are excited to support the next generation of software tools for hardware engineering. If you are building in the space or have any thoughts to share, reach out at omar@scalevp.com.
