Moving Beyond the Lens: Voxel's Bid for an Industrial 'System of Reality'

SAN FRANCISCO – October 13, 2026 — For all the breathless hype surrounding artificial intelligence over the past few years, the corporate bottom line is rarely transformed by chatbots. Real operational change happens on the warehouse floor, in the loading bay, and across the sprawling footprints of hyperscale data centers. It happens in the physical world, where the friction of moving goods and materials dictates profitability.

Voxel, a San Francisco-based industrial intelligence company, has been quietly digitizing this physical friction. Today, the company announced the integration of Ventral AI, an advanced computer vision and industrial automation engineering firm based in Croatia. The move—which industry observers characterize as a strategic acqui-hire—signals a critical shift in how enterprises are deploying artificial intelligence. We are moving past the era of static observation and entering a phase of dynamic, real-time operational intelligence.

The transaction brings Ventral AI’s co-founders, Matko Kulaš and Luka Žagar, along with their specialized engineering team, into Voxel's fold following a year-long strategic partnership. But more importantly, it brings a specific set of technical capabilities that solve one of the most stubborn bottlenecks in physical AI: the moving camera.

The Edge of Operations: Untethering the Camera

Until recently, industrial computer vision has been largely tethered to the wall. Companies have successfully layered AI algorithms over existing, fixed security cameras to monitor safety compliance or track inventory. But a modern industrial environment is not static. It is a chaotic ballet of moving parts, transport fleets, and heavy machinery.

Deploying computer vision in these dynamic environments is notoriously difficult. When a camera is mounted on a forklift moving at fifteen miles per hour through a dimly lit warehouse, the resulting video feed is plagued by motion blur. Objects are constantly occluded as workers step behind pallets or machinery. The background is a complex, ever-changing tapestry that easily confuses standard vision models. Furthermore, processing these high-volume video streams requires robust edge computing—analyzing the data on the machinery itself rather than sending it to a distant server, which introduces unacceptable latency.

Building on a foundation of solving these exact problems, Ventral AI established its reputation by deploying real-time computer vision and vision language models on moving equipment. By absorbing this intellectual property and engineering talent, Voxel is effectively untethering its platform from the warehouse wall.

This capability extends beyond logistics and warehousing into the high-stakes world of infrastructure and data center construction. Construction sites are uniquely challenging for AI because they are inherently evolving environments. A model trained to monitor a site on Monday might be obsolete by Friday as new structures are erected and material piles shift. Ventral AI’s expertise allows Voxel to deploy models tailored specifically to these high-precision, heavy industrial environments, ensuring that safety protocols and progress tracking keep pace with the reality on the ground.

The financial impact of getting this right is substantial. Voxel’s platform, trained on over five billion hours of real-world industrial scenarios, already boasts a 96 percent detection accuracy. Enterprise case studies from their portfolio—which includes heavyweights like Americold, Dick's Sporting Goods, and Berry Global—report staggering results: a 50 percent reduction in ergonomic incidents and a 73 percent reduction in workers' compensation costs. By expanding these capabilities to moving assets, the potential return on investment multiplies for their global enterprise clients.

Mining Europe's Technical Veins

Beyond the technology, the integration of Ventral AI highlights a broader strategic trend among Silicon Valley enterprise startups: the race to secure elite engineering talent in Europe.

While the financial terms of the deal remain undisclosed, the integration functions as a classic acqui-hire, absorbing Ventral AI's headcount and establishing a vital European bridgehead for Voxel. Based in Zadar, Croatia, the acquired team provides a direct pipeline into the region's top-ranking technical universities, allowing the parent company to scale its engineering workforce outside the highly competitive Bay Area market.

"From our earliest collaboration, we saw the incredible potential of uniting our expertise with Voxel's leading operational intelligence platform," said Žagar, Co-Founder and former CEO of Ventral AI. "Together, we are unlocking a new level of real-time intelligence for physical operations, enabling enterprise teams to solve complex frontline challenges faster while also bringing applied AI innovation into the physical world."

This expansion arrives during a period of aggressive growth for Voxel. The company recently ranked No. 71 on the Inc. 5000 list of America's fastest-growing private companies and has raised $83 million in total funding, including a $44 million Series B and strategic backing from Ericsson Ventures. To sustain this momentum and serve a growing international customer base, establishing a dedicated engineering and talent hub in Europe is not just strategic; it is a logistical necessity.

Can Computer Vision Become Core Enterprise Software?

At the heart of this expansion is an ambition to pioneer a new category of enterprise software, which executives have dubbed the "System of Reality."

For decades, corporations have relied on Systems of Record—like Enterprise Resource Planning (ERP) software—to act as the single source of truth for their financial and operational data. But these systems are inherently retrospective; they rely on human input and historical data. Voxel is betting that computer vision can serve as a proactive, continuous layer of operational intelligence.

"As industrial leaders navigate supply chain volatility, labor constraints, and rising throughput demands, they need a single source of truth, a System of Reality, for their entire physical footprint," said Vernon O'Donnell, CEO of Voxel. "The Ventral AI team gives us even greater engineering horsepower to meet that demand head on. Their expertise in complex computer vision further accelerates the evolution of our operational intelligence platform."

The market for this technology is fiercely competitive. Voxel is vying for dominance against well-funded players like Protex AI and Intenseye, who also focus on workplace safety, as well as industrial IoT giants like Samsara, who provide AI-powered fleet management.

However, the push to define a new software category suggests a broader vision. Voxel does not want to be viewed merely as a safety tool or a specialized analytics dashboard. By transforming passive video feeds into a system that diagnoses root causes in real-time and directs immediate action, the firm aims to make its software as sticky and indispensable as a traditional ERP.

If an enterprise can definitively track not just what was shipped, but exactly how it was handled, the specific movement of the machinery involved, and the ergonomic safety of the worker executing the task—all in real-time—the camera ceases to be a simple security device. It becomes the foundational sensor of the modern industrial enterprise. With the engineering horsepower of Ventral AI now fully integrated, Voxel is betting it has the technical leverage to make that vision a reality on a global scale.