The Site Is Becoming a Machine
Sites are being instrumented, measured, and fed into software systems that can act on them in real time.
This week, DroneDeploy crossed 20 trillion square feet of visual site data, MicroVision reported revenue-generating LiDAR shipments into mines and heavy trucks, Crewline AI raised $7.1M to scale self-driving rollers across 100 construction sites, and Boston Dynamics deployed Spot at an operating copper mine in Utah, integrated into the mine’s software stack. All four are products, revenue lines, or live deployments.
DroneDeploy crosses 20 trillion square feet of visual site data
On April 22, DroneDeploy reported more than 20 trillion square feet of captured visual site data across 3 million sites in 180 countries. The data comes from aerial capture, ground robots, and 360 walks across construction, energy, agriculture, and infrastructure. The company describes it as the largest visual dataset in construction.
Progress AI tracks construction progress across more than 50 active commercial projects in minutes, replacing multi-day manual review. Safety AI has flagged more than 90,000 risks since launch. The models improve with additional data, which turns the milestone into a distribution and moat signal.
DroneDeploy reached break-even in September 2025 and raised $15M from existing investors to fund its AI and robotics roadmap. Docked drones are active on more than 100 projects. Autonomous ground robots enter beta in 2026. The company is also moving toward FedRAMP alignment, which opens access to US government deployments.
MicroVision ships LiDAR into mines and trucks
On April 22, MicroVision reported paid shipments of its Iris LiDAR sensor into mining, hauling, and autonomous and semi-autonomous trucking systems. At least one mining programme has moved from field validation into production deployment.
Iris detects obstacles beyond 250 metres and is designed for dust, vibration, and temperature variance in industrial environments. The company has shifted away from passenger automotive toward industrial verticals, where deployment timelines are shorter and procurement is tied more directly to operating needs.
Revenue is still early. MicroVision reported $1.2M for full-year 2025 and guides $10–15M for 2026. The signal is not scale. It is that demand for sensing is emerging from sectors already running or preparing to run autonomous systems in the field.
Crewline raises $7.1M to scale autonomous construction rollers
Crewline AI raised a $7.1M seed round led by Initialized Capital and Nebular, with Ford Street Ventures, Cocoa, Begin, Entrepreneurs First, and Transpose participating.
The company builds retrofit autonomy kits for vibratory drum rollers. Installation takes around 30 minutes and requires no permanent modification. Machines operate autonomously within geo-fenced zones, with remote operators stepping in when conditions change.
A 30-day deployment on a 30-acre airport extension in Austin produced the initial performance data. Daily downtime fell from six hours to under one hour. Productive hours nearly doubled. The machine recorded zero accidents. Crewline reports $220K in ARR, a $26M pilot pipeline, and more than $50M in waitlisted demand. The 2026 target is 100 units on live earthwork jobs.
Compaction is the starting point because it is repetitive, GPS-compatible, and easier to automate than articulated machines. The roadmap moves into dozers, rock trucks, and excavators. Contractors pay a monthly fee with no upfront cost, which removes procurement friction and shifts uptime risk to Crewline.
Boston Dynamics deploys Spot into a copper mine
On April 21, Boston Dynamics and Mariana Minerals announced deployment of Spot at Copper One, an operating copper mine and refinery in southeastern Utah.
Spot connects directly into MarianaOS, the mine’s software stack. Inspection data is ingested, structured, and fed into production, maintenance, and environmental workflows in real time. Anomaly detection triggers automated escalation. Inspection shifts from periodic visits to continuous monitoring tied to planning and scheduling.
Boston Dynamics integrated Gemini Robotics-ER 1.6 into Spot and Orbit in mid-April. The system can read analogue gauges, detect chemical spills, and perform multi-view visual analysis with limited human input. Several thousand Spots are deployed globally. The Copper One integration shows how those robots are being used, as part of the operating system rather than a separate inspection tool.
Source: Mariana Minerals
Context
Construction productivity has declined roughly 30% since 1970 while the broader US economy has doubled. The US faces a shortfall of around 800,000 construction workers over the next two years. Project backlogs exceeded eight months as of December 2025. Investment in manufacturing facilities is at record levels, driven by reshoring, data centres, and energy infrastructure.
These four signals sit on different layers of the same constraint. DroneDeploy builds the data layer. MicroVision supplies sensing. Crewline adds machine autonomy through retrofit. Boston Dynamics integrates robots into operational systems. None solves the full problem alone. Together they show how sites move from manual environments toward systems that can be measured, monitored, and acted on in real time.