Zurich-based construction robotics firm Gravis Robotics has secured $200 million in Series A funding from SoftBank, marking what the company describes as the largest Series A financing in construction robotics to date. Emerging as an ETH Zurich spinout in late 2022, the company develops hardware and AI software designed to retrofit existing heavy machinery for autonomous infrastructure projects worldwide.
Untangling the Mechanics of Jobsite Autonomy
Autonomous deployment in construction presents an engineering problem fundamentally distinct from warehouse automation or standard autonomous vehicles. A self-driving vehicle generally attempts to understand and safely navigate an environment without changing it. By contrast, an excavator is expected to do the opposite.
Every bucket movement alters the terrain the machine needs to understand next. Machines must contend with soil, rocks, changing slopes, underground resistance and other conditions that cannot be fully predicted in advance. To process these unscripted variables, Gravis integrates data from machine hydraulics, 3D LiDAR, cameras, and Global Navigation Satellite System (GNSS) positioning.
According to CEO and co-founder Ryan Luke Johns, physical infrastructure work has long served as an industry bottleneck. “To build the future, we need to change the world, literally. Whether we are building housing, modernizing energy grids, or scaling data centers, every project starts with moving earth,” Johns stated in details released by Gravis Robotics. “Our machines are built for the messy, unscripted reality of live jobsites that breaks traditional automation.”
Retrofitting Fleets via the Gravis RACK Architecture
It is common for construction fleets to comprise machinery from various brands, typically gathered over years through a mix of rental agreements and direct purchases.
Rather than requiring operators to buy purpose-built autonomous machinery from scratch, Gravis engineered a modular rooftop unit called the Gravis RACK. This system combines automotive-grade edge computing, GNSS RTK positioning, cameras, and 3D LiDAR coverage onto a single deployable hardware package.
- Caterpillar
- John Deere
- JCB
- Hitachi
- Volvo
- Yanmar
- Case
- Develon
- Sumitomo
Since data processing happens directly on board the equipment, autonomous operations can keep running seamlessly even when a stable network connection is absent—a vital capability on unfinished or remote jobsites.
Simulation Training and Ecosystem Expansion
To train models against the chaos of real-world earthmoving, Gravis subjects its learning-based control systems to extensive simulation. This allows the AI to encounter large numbers of virtual excavation scenarios before being deployed to physical machinery.
Beyond basic digging routines, the perception system manages downstream workflows. By combining multiple sensor inputs, the system spots dump trucks and determines precise drop-off points for the excavated material, enabling autonomous tasks to expand past simple digging into loading and material management processes.
Furthermore, continuous LiDAR scanning serves a dual purpose. While supporting autonomous excavation, it simultaneously produces 3D site data, cut-and-fill visualizations, and records of completed work.
With funding from SoftBank, the company plans to rapidly scale its engineering team, accelerate its international rollout, and deploy its autonomous systems into more construction and infrastructure projects globally.
