Boston Dynamics, a US-based developer of mobile robots used in industrial environments, and FieldAI, a US-based developer of embodied AI systems, have announced a partnership that aims to extend autonomous robotics into construction and other dynamic, unstructured environments.
The collaboration combines Boston Dynamics’ robotic platforms, including its quadruped Spot, with FieldAI’s autonomy software built on what it describes as Field Foundation Models, designed for decision-making in unpredictable physical settings.
Construction sites, with shifting terrain, active workflows, and constantly changing layouts, are being used as a proving ground for the combined approach.
Spot robots moving into adaptive site environments
Spot (robot), previously used mainly for structured inspection tasks, is now being applied in construction environments that change from day to day.
In its existing deployments, Spot has primarily been used for industrial inspection, mapping, and monitoring in controlled environments such as manufacturing facilities and infrastructure sites. The new approach extends its role into active construction settings, where conventional programming methods often struggle to maintain consistent performance.
FieldAI’s system is intended to allow robots to interpret and respond to environmental uncertainty, including uneven terrain and evolving site layouts. The system is being applied to functions, including 3D progress tracking, hazard detection, and continuous site monitoring without relying on fixed routes or pre-mapped layouts.
From standalone robots to coordinated fleets
The partnership also moves towards multi-robot coordination, with robots designed to function as part of coordinated fleets that share environmental understanding and operational data.
FieldAI’s system enables this coordination by allowing multiple robots to operate across sites while building a shared view of conditions on the ground. The companies say this approach could help standardise data collection across projects and improve consistency in site monitoring at scale.
Continuous data capture also feeds back into model refinement, linking deployment activity with ongoing system training.
Construction sites as a testbed for autonomy
The collaboration positions construction environments as a stress test for autonomous robotics. Sites typically combine shifting terrain, overlapping human activity, and changing spatial constraints, creating conditions that are difficult to replicate in controlled settings.
FieldAI says its models are designed to operate without relying on pre-mapped environments, GPS, or cloud connectivity. Instead, the system is intended to support on-device decision-making, allowing robots to function in remote or infrastructure-limited locations.
If such systems can operate reliably in construction settings, the companies suggest the same approach could extend to other complex environments with similar levels of variability.
Claims around efficiency, safety, and scale
The announcement points to early deployment results that include reductions of more than 90% in inspection and documentation time compared with manual processes, along with earlier detection of deviations that can reduce rework. It also highlights the use of robots in hazardous areas to support safety monitoring.
Spot robots are described as operating across multiple regions, including Asia, Europe, and North America, with deployments continuing to expand. In construction use cases, they can operate overnight, generating updated site insights through 3D scans and visual documentation used for progress tracking.
FieldAI and Boston Dynamics also indicate that further expansion of Spot deployments is planned in the coming months as part of broader fleet-based rollouts.
Shifting how construction sites are observed
The partnership reflects a gradual shift in how construction sites are being documented and monitored. Instead of periodic inspections and manual surveys, the model described here moves towards continuous, machine-led observation.
The partnership also reflects a broader shift in how robotics companies are approaching construction, not simply as a deployment market, but as a real-world testing ground for autonomy systems operating under constantly changing conditions.
*All Images: Courtesy of FieldAI*
