Why C2 Ventures invested in Shelfmark: giving production lines a memory. Shelfmark is helping continuous-flow manufacturers catch defects, uncover their causes, and prevent them from happening again. C2 Ventures is pleased to announce its investment as part of the second close of Shelfmark's $3.5 million seed financing, led by Armory Square Ventures. Shelfmark is an AI-native production intelligence platform built for continuous-flow manufacturing, where industrial films, labels, textiles, coatings, and other essential materials can move through production lines at hundreds of feet per minute. This is exactly the type of industrial AI application C2 Ventures find compelling: a large, durable market underserved by modern software; a repetitive task that is difficult for people to perform consistently; and a technology capable of producing tangible operational returns. A difficult job with expensive consequences. Quality inspection on continuous-flow production lines remains an incredibly manual process. Workers are often tasked with monitoring material as it moves continuously in front of them, looking for defects that may be smaller than a human hair. Maintaining that level of focus and precision across an entire shift is nearly impossible. When a defect is missed, the impact can extend well beyond the material itself. Manufacturers face wasted inventory, rework, line stoppages, downgraded products, rejected shipments, and unhappy customers. Traditional machine-vision systems have attempted to address this problem, but they can be expensive, rigid, and difficult to adapt across different materials, products, and operating environments. Shelfmark has developed a more flexible approach. From detecting defects to understanding their causes. Shelfmark combines deep-learning computer vision, specialized line-scan cameras, and production data to inspect 100% of material as it moves through the line. The platform identifies defects in real time, alerts operators, and creates a digital record of what happened during each production run. Shelfmark reports inspection accuracy of up to 99.5%, including on materials previously considered too fast, reflective or variable to inspect reliably with automation. Catching the defect is only the first step. Shelfmark connects defects with the production conditions surrounding them, including temperature, humidity, pressure, and line speed. This helps manufacturers understand what went wrong, identify patterns across multiple runs, and determine how to prevent the same problem from returning. Instead of treating every defect as an isolated incident, manufacturers begin building an operational memory of the line. Tangible results on the factory floor. Industrial technology has to produce measurable value under real operating conditions. Technical novelty alone will not carry a product beyond the pilot stage. Shelfmark reports that it has worked with more than 40 manufacturing facilities and converted nine out of ten pilots into customers. Across its deployments, customers have reported: * Waste reductions of up to 90% * Inspection labor reductions of 50% * Inspection accuracy of up to 99.5% * Returns of up to seven times the investment The value is direct: less scrap, less manual inspection, fewer production interruptions, and more usable output from the same equipment and workforce. Shelfmark also manages the complete system, including cameras, edge hardware, software, and machine-learning models. Manufacturers gain the benefits of production intelligence without having to build and maintain an internal machine-vision operation. The larger opportunity. The immediate value proposition is clear, but the longer-term opportunity is even more compelling. Every production run creates a digital record of what happened on the line. Over time, that gives manufacturers a growing dataset they can use to identify patterns, improve processes and make better operating decisions. This creates a path from defect detection to root-cause analysis, predictive recommendations, and, ultimately, production lines capable of adjusting before problems occur. Shelfmark has the potential to become the persistent intelligence layer across continuous-flow manufacturing: a system that helps production lines see every inch, understand what happened, and improve with every run. Why C2 Ventures invested in Shelfmark. C2 Ventures has spent years investing in AI-supported software, robotics, and automation for legacy industries. C2 Ventures look for companies that apply technology to expensive operational problems in markets where the status quo has been accepted for too long. Shelfmark fits that thesis on several levels. Continuous-flow manufacturing is a massive and fragmented market that has historically received less attention from modern software providers. The inspection workflow is repetitive, difficult for people to perform consistently, and enormously costly when mistakes occur. Shelfmark's ROI is also highly tangible. Manufacturers do not have to rely on an abstract promise of future efficiency. They can measure the value through reduced scrap, lower inspection costs, fewer interruptions, and increased production capacity. The team's approach matters just as much. Pat O'Donnell, Craig Markovitz, and the Shelfmark team have spent time on factory floors, working with operators to understand how these lines actually run. That practical experience is reflected in a product designed to deliver results without adding another complicated system for plant teams to manage. Chris Cunningham, Founder of C2 Ventures, states, "Shelfmark is exactly the type of company we built C2 Ventures to support. The team is bringing practical AI into an enormous industrial market where the problems are real, the value is measurable, and existing technology has fallen short. Shelfmark is helping manufacturers reduce waste and improve quality today while building toward production lines that get smarter with every run." C2 Ventures' sector experience was also an important part of the fit. Shelfmark wanted an investor that understood industrial technology, the realities of selling into legacy markets, and the work required to turn a successful deployment into a durable company. Founder and CEO, Pat O'Donnell, notes, "Validation of Shelfmark's mission from a sector-aligned fund like C2 Ventures means the world to Shelfmark and our partners. As a company building with our customers and for our customers, directly from the factory line, an investor that shares this worldview is incredibly important." What comes next. The funding will support continued product development, commercial expansion, and the growth of Shelfmark's Pittsburgh-based team. C2 Ventures is excited to back Pat, Craig, and the entire Shelfmark team as they build the intelligence layer for some of manufacturing's most overlooked production lines. The company has already shown that it can help manufacturers catch problems earlier, reduce waste, and improve output. The next step is helping those production lines retain what they learn - and prevent the same problems from happening again.

Why Grand Ventures invested in Shelfmark. August 6, 2026 By: Tim Streit Walk through almost any factory making rolled goods, fabric, labels, industrial film, webbing, and you'll find the same quality control system: a person standing next to a production line moving hundreds of feet a minute, looking at it. That has been the state of the art in continuous flow manufacturing quality control for decades. Not because manufacturers haven't tried to do better. Because until recently, nothing better existed. Grand Ventures is proud to partner with Pat O'Donnell and Craig Markovitz, co-founders of Shelfmark, who are building the AI inspection platform that continuous-flow manufacturing has been waiting for. Legacy machine vision players like Keyence, Cognex, and ISRA were built decades ago for discrete, static products. Their systems rely on rules-based algorithms that require controlled lighting, fixed geometry, and tightly scripted defect libraries. This type of technology hasn't transferred well to continuous flow manufacturing, where lines run at extremely high speeds. The result: continuous-flow manufacturers, the companies making the fabric in your clothes, the labels on your prescriptions, the material in military parachutes, were left out. Grand Ventures spoke with a 30-year industry veteran who put it plainly: the incumbents won't bid on a product if they don't already have the algorithm for it. And they don't have algorithms for continuous flow. That's the gap Shelfmark fills. Why now? Advances in computer vision and sensor technology have unlocked data capture at a resolution and scale that was previously infeasible. Paired with modern AI, this enables dynamic modeling, pattern recognition, and near-real-time detection and resolution. Shelfmark's platform pairs deep-learning computer vision with line-scan camera hardware and IoT sensors to inspect every inch of material at full production speed. No sampling, no spot checks, no hoping the defect happened to be in the section someone glanced at. What makes this compound over time is what happens after detection. Shelfmark generates a "roll map report card" for every production run, a structured quality record tied to that specific roll of material. That record is beginning to travel up and down supply chains. Customers are pushing Shelfmark on their own suppliers. Grand Ventures has seen a lot of go-to-market strategies; 'your customers do it for you' is one of the best ones. One Fortune 500 customer built their own internal ROI model, presented to their CEO to justify rolling Shelfmark out globally, showing $40 million in five-year savings net of cost. Usually the optimistic spreadsheet is the VC's job. This time the customer did it for Grand Ventures. Twenty minutes into my first call with Pat, the CEO, it was obvious he hadn't built a tool and gone looking for a problem. He'd spent months watching the problem happen and built what the people would buy. Before Pat wrote a single line of code, he spent months on factory floors understanding this market firsthand. He rode alongside workers on early shifts at factories and production floors, absorbing what the job actually looked like before designing software to change it. His background spans industrial engineering, Deloitte consulting, and software product studios. He speaks the language of plant managers and the language of engineers, which matters in a market where the buyer is a floor operations lead, not a CTO. Co-founder Craig Markovitz previously founded Blue Belt Technologies, a CMU Robotics Institute spinout in surgical robotics, acquired by Smith & Nephew. He's taught entrepreneurship at CMU for a decade, which means he's seen more startup pitches than almost anyone alive. Shelfmark is the only one he joined. Physical hardware on production lines creates switching costs that no software-only competitor can bypass without replacing line equipment. Shelfmark has patents filed covering vertical-specific inspection methods. Each deployment generates labeled defect data that makes models more accurate and shortens time-to-value for the next customer in the same vertical. And as the roll map report card becomes a shared quality standard, pushed upstream and downstream by customers, the company earns a structural position in industrial infrastructure that doesn't come from a product roadmap. It compounds from deployment. Shelfmark is operating in a very green space. The incumbents will eventually notice the market they've been ignoring - they usually do, once someone else proves it's worth the trouble. Shelfmark's head start is leading in deployments - what compounds and accelerates is the data they will generate and analyze, and the network effects of manufacturers, vendors, and customers. At Grand Ventures, Grand Ventures look for companies quietly becoming mission-critical in large, overlooked markets. A significant segment of global manufacturing is still doing quality control by eyeball in 2026, not for lack of trying, but for lack of a product that worked. Shelfmark built the one that does. Grand Ventures is proud to welcome them to the Grand Ventures portfolio. Extra icing on the cake... * My first summer internship nearly 30 years ago was on the plant floor at Ford Motor Company (I studied ME at U of M) and observed quality control first hand...manually and painstakingly. * Pittsburgh has been on my radar for over 10 years, waiting for the right company. The engineering and computer vision expertise at CMU is unparalleled. * Kudos to Jacqueline for sourcing this opportunity, and Anthony at Armory Square for sharing! * Grand Ventures has been spending more time on Physical AI the past 6-12 months. This is the product of the early work. * I love it when a plan comes full circle! If you go back to my Ford days/engineering undergrad... it's been 30 years in the making!