Gravis Robotics raises $200M from SoftBank to scale autonomous excavators

Why it matters

SoftBank's check gives Gravis the capital to compete in a market where mixed-fleet support, field data and equipment distribution will decide the winners.

Ryan Luke Johns and Dominic Jud's Gravis Robotics raised a $200 million Series A from SoftBank (@SoftBank) to expand its autonomous construction machinery business worldwide. Tech.eu reported that SoftBank was the round's sole investor. Gravis announced the financing on August 17, about nine months after raising $23 million from a group led by IQ Capital and Zacua Ventures. (gravisrobotics.com)

The size of the check reflects how far Gravis has traveled from its academic roots. Johns is an architect and roboticist whose doctoral work focused on autonomous dry-stone construction. One project produced a 65.5-meter-long, 6-meter-high retaining wall containing 938 unique elements, according to Johns' ETH dissertation. Jud and Burak Cizmeci developed teleoperated and semi-autonomous heavy machinery for dangerous work, including disaster-response road clearing and unexploded-ordnance removal. They tested those systems with Swiss Federal Railways and Swiss defense procurement agency armasuisse. (ethz-foundation.ch)

Gravis was founded in October 2022. ETH materials identify Jud, Cizmeci, Simon Kerscher and Johns among the founding team. Gravis has since narrowed the work into a commercial proposition: add sensors, computers and controls to equipment contractors already own, then let that machinery move among manual, remote and autonomous operation.

Gravis is betting on retrofits

Gravis' core product is the Gravis RACK, a retrofit system combining sensors, cameras, LiDAR, positioning technology and machine telemetry. The accompanying Slate interface lets an operator switch a machine among manual, teleoperated and autonomous modes.

The retrofit model addresses a practical barrier to autonomy in construction. Contractors have already invested in fleets assembled from several manufacturers, often around regional dealer and maintenance relationships. Selling an intelligence layer for existing excavators and heavy equipment gives Gravis a route to market that does not require customers to replace their fleets with purpose-built robotic machines. Gravis has not disclosed how many systems have been installed for paying customers.

The technical problem is also distinct from autonomous driving on roads. An excavator changes its environment with every bucket movement, encountering soil, rocks and hydraulic resistance that cannot be fully mapped in advance. Jud said Gravis feeds those physical inputs into machine telemetry so its control system can adjust to changing ground conditions. "We didn't try to simplify the world for our software; we gave it the physical intuition to handle real job sites," he said in the funding announcement. (Tech.eu)

Gravis claims the system can improve jobsite productivity by as much as 30% compared with peak manual operation. That figure has not been independently audited, and Gravis has not published the underlying customer count, measurement period or testing methodology. Public demonstrations provide firmer evidence that the system works outside an ETH lab. In a 2026 demonstration, Hitachi Construction Machinery showed an operator train a retrofit excavator before it repeated a trenching task autonomously.

SoftBank is funding the distribution race

Gravis said the Series A will fund hiring and a global rollout. As of August 17, the company listed Zurich as its headquarters, described Austin as its next growth location and reported deployments across seven countries. Gravis has also announced a commercial partnership with Flannery Plant Hire to offer turnkey rental excavators equipped with the Gravis RACK. That channel matters because contractors commonly obtain equipment through manufacturers, dealers and rental providers rather than buying robotics software directly.

SoftBank's investment lands during a surge of financing for companies applying autonomy to heavy construction. Bedrock Robotics, a construction autonomy company founded by former Waymo engineers, raised a $270 million Series B in February 2026 to develop systems coordinating autonomous construction fleets. TerraFirma, a heavy-equipment autonomy startup, raised $115 million in July, including a $100 million Series A led by Kleiner Perkins, for software, remote operations and retrofitted equipment. Built Robotics, another construction robotics company, already markets an aftermarket excavator kit focused on autonomous trenching.

Gravis claims its financing is the largest Series A in construction robotics history. Recent comparable rounds support the scale of that claim, though no independent dataset establishes the record across the entire category. The valuation and SoftBank's resulting ownership stake were not disclosed. Gravis has now announced approximately $223 million in funding since November 2025, excluding earlier financing for which a complete figure is unavailable. (gravisrobotics.com)

The capital gives Johns and Jud room to move beyond demonstrations and limited deployments. Their next test will come from contractors measuring output, downtime, safety and rework across entire projects. The technology has already moved real earth. The round gives Gravis more time and capital to prove it can do so repeatedly across equipment brands, countries and construction workflows.