Gravis Robotics closed a $200M Series A led by SoftBank on August 17, 2026.

Before the Series A, Gravis raised $23M in November 2025 from IQ Capital, Zacua Ventures, Pear VC, Sunna Ventures, Armada Investment, and Holcim. Holcim participated as a strategic investor alongside its quarry deployment partnership. An earlier seed round of approximately $4M was raised in March 2023.

Total publicly announced funding stands at approximately $223M across all rounds.

Gravis Robotics makes a retrofit autonomy system for excavators and wheel loaders that turns a contractor's existing machines into terrain-aware, semi-autonomous earthmoving equipment, without requiring the purchase of new hardware from a single OEM.

The system has two physical components. The Gravis Rack is a modular rooftop kit that bolts onto the machine and houses LiDAR, cameras, RTK/GNSS positioning, onboard edge compute, and Wi-Fi, and it runs offline when site connectivity is poor. The Slate is a rugged touchscreen tablet that serves as the operator interface, where the operator defines the job, visualizes the terrain, and switches between operating modes.

A typical workflow starts with the operator importing a grading plan from CAD or BIM, or defining the target geometry directly on the tablet. The Rack scans the site and renders a live cut/fill visualization on Slate, color-coding where material needs to come out and where the ground is already at grade. The operator then picks one of three modes: in-cab Copilot, which provides AR-style guidance and people-detection alerts while the operator stays in the seat; Field Command, where the operator supervises from outside the cab; or Remote Orchestration, where one operator monitors and controls one or more machines via low-latency video from a distance.

Once a repetitive task is defined, such as trenching a utility corridor, loading trucks from a stockpile, or feeding a quarry screener hopper, the machine executes autonomous dig-swing-dump cycles while the system continues scanning and updating the as-built model in the background. The machine adapts to changing ground conditions in real time using data from hydraulics, LiDAR, cameras, and GNSS rather than replaying a fixed motion path.

The platform is compatible across a range of brands including Caterpillar, Case, Develon, John Deere, JCB, Hitachi, Sumitomo, Yanmar, and Volvo, which matters because most contractor fleets are mixed and sticky. Every equipped machine also contributes field data back to Gravis's models, which improves performance across soil types and task types as the installed base grows.

Gravis Robotics sells B2B through a mix of direct enterprise contracts, OEM and dealer channel partnerships, and rental-market distribution. Its monetization combines a one-time hardware and installation fee for the Gravis Rack with an annual software license covering autonomy, analytics, and workflow software, with leasing available as an alternative to outright purchase.

The rental channel, anchored by a partnership with Flannery, the UK's largest plant-hire provider, is important because it lets contractors access autonomous capability as an operating expense rather than a capital commitment. That lowers adoption friction for mid-sized firms that would not buy a bespoke robotic excavator outright, and gives Gravis access to an established equipment distribution network rather than relying entirely on direct sales.

Mixed-fleet compatibility is the main differentiator in the business model. Because the same software stack runs across many machine brands and sizes without custom reprogramming, Gravis can sell a neutral autonomy layer across a contractor's heterogeneous fleet. That expands the revenue opportunity within each customer relationship and increases switching costs once the system is deployed across multiple machine types on a site.

Gross margin is mixed: hardware and installation carry lower margins than software, and early deployments require meaningful field engineering and support. As the installed base grows and channel partners take on more of the distribution and service burden, the recurring software layer should become a larger share of revenue and improve overall margins. Each additional deployed machine also adds field hours across diverse soil types, which can improve the autonomy models and reduce friction in new customer deployments.

The retrofit autonomy market for heavy earthmoving equipment is contested on three fronts: machine-guidance incumbents selling lower-risk operator-assistance tools, teleoperation-first startups targeting the same equipment budgets, and OEMs embedding autonomy into their own platforms. Gravis competes most directly in retrofit autonomy, but in practice it also competes with guidance systems and OEM distribution.

Built Robotics is the closest historical analog in the U.S. market, offering an aftermarket autonomy upgrade for excavators called Exosystem with a cloud monitoring layer called Everest. Built has accumulated over 20,000 hours of autonomous field operation and has concentrated commercially on solar piling and utility trenching, giving it a documented ROI case in a specific high-volume segment.

Bedrock Robotics is the clearest emerging rival. It markets a same-day retrofit system for existing fleets and draws on Waymo-derived safety and ML expertise to argue that its machines can generalize across variable real-world conditions. Bedrock has named deployments with major U.S. civil contractors including Sundt, Zachry Construction, and Austin Bridge & Road, giving it a foothold in the large-scale infrastructure segment where Gravis also wants to compete.

Teleo competes for the same budget line by leading with remote operation and supervised autonomy rather than terrain-aware autonomous excavation. For contractors whose primary constraint is operator shortage or hazardous-site staffing rather than fully autonomous dig cycles, Teleo's simpler deployment model can win procurements that Gravis's more complex platform may lose on sales-cycle length.

Develon is both a channel partner and a competitive threat. The jointly developed Real-X autonomous excavator entered practical use with KIBAG in Switzerland in 2026, validating the Gravis stack commercially while also giving Develon direct exposure to customer requirements and distribution through its own dealer network. Each successful OEM-integrated deployment increases the risk that the OEM internalizes the autonomy layer instead of continuing to source it externally.

Caterpillar is the highest-consequence long-term threat even if it is not yet a like-for-like retrofit rival. Cat is expanding its Command autonomy portfolio from mining into construction, quarry, and aggregates, and its distribution power, dealer service infrastructure, financing relationships, and buyer trust can make a slightly less capable native solution more attractive to large contractors than a third-party retrofit that requires separate procurement, insurance, and support arrangements.

Trimble and Komatsu Smart Construction target a meaningful share of Gravis's value proposition from a lower-risk angle. Trimble Earthworks and Komatsu's retrofit guidance kits give excavator operators 3D grade control, cut/fill visualization, and payload monitoring without crossing into autonomy, and both are sold through established global dealer and reseller networks.

For buyers who mainly want better operator precision and less rework rather than robotic execution, these products can absorb budget that might otherwise go to Gravis's Copilot layer. The practical competitive question is whether Gravis can outperform other autonomy vendors and whether the market treats advanced guidance as a long-term substitute for full autonomy.

Gravis Robotics's expansion logic runs along three tracks: adding autonomous task types and equipment coverage, moving into adjacent industrial verticals, and entering geographies where infrastructure spending is rising.

The nearest-term TAM lever is expanding the number of autonomous tasks the platform supports on machines already in the field. Gravis currently covers trenching, bulk excavation, truck loading, grading, stockpile management, and quarry hopper feeding, tasks with the repetitive, bounded operating conditions that make autonomy more practical. Each added task can raise utilization of the same Rack hardware and Slate interface, increasing revenue per machine without incremental hardware cost.

The Copilot and operator-assist layer extends the addressable market beyond jobs that are fully automatable today, because contractors can capture value from guidance and surveying before they are ready to run machines unsupervised. Over time, the terrain scanning Rack generates on every shift could support a second software revenue layer in jobsite intelligence, including production analytics, terrain change tracking, safety event logs, and as-built verification, monetizing the installed base even when autonomy adoption remains partial.

Quarries and aggregates are the clearest near-term vertical expansion. Workflows are typically more repetitive and controlled than open construction sites, routes are more predictable, and the value of 24/7 utilization and removing operators from hazardous zones is higher. The Holcim quarry trial at Lee Moor indicated commercial potential, and Holcim's strategic investment in Gravis included the opportunity to expand from construction into quarry applications.

The broader aggregates sector is also moving toward automation independently. Epiroc extended its autonomous haulage system into aggregates in 2026, which may reduce buyer education costs for Gravis in quarry sales. The Flannery rental partnership also broadens access by giving mid-sized contractors a way to use autonomous excavation without a capital commitment.

Gravis was live in seven countries across the UK, EU, U.S., LATAM, and Asia by late 2025, and by August 2026 reported deployments across four continents, with offices in Zürich, Austin, and Oxford, indicating concurrent European and U.S. commercialization. The U.S. market is timely: grid expansion tied to AI data center load growth, domestic manufacturing reshoring, and electrification is increasing transmission and utility infrastructure activity, which is excavation- and grading-intensive, with annual data center construction investment projected to exceed $250B by 2027.

APAC remains a longer-term target. PwC's 2026 global infrastructure outlook projects total global infrastructure investment rising from $4.4T in 2024 to $6.9T by 2050, with Asia-Pacific accounting for the largest share of new build volume, a market where earthmoving intensity and labor dynamics may favor autonomous equipment adoption.

OEM enclosure: As Caterpillar, Komatsu, Develon, and other OEMs accelerate native autonomy programs, they could close machine hydraulic and control interfaces or favor their own stacks commercially, compressing Gravis from an independent platform into a component supplier or excluding it from new machine generations entirely.

Deployment complexity ceiling: Because every new site requires physical installation, safety validation, operator training, and site-specific calibration, the business scales more like a field-services operation than a pure software company and could face a ceiling on deployment velocity before channel partners and OEM dealers absorb that operational burden.

Guidance substitution: A large share of Gravis's near-term addressable market may settle for advanced machine-guidance products from Trimble, Komatsu Smart Construction, or OEM-native grade-control systems rather than progressing to full autonomy, because guidance delivers a meaningful share of the productivity gain at lower procurement risk, lower insurance complexity, and through already-trusted dealer relationships.