Research · October 28, 2026

Bedrock Robotics: Paid Autonomous Work, on Machines Contractors Already Own

On October 26, 2026, Bedrock Robotics announced a $600 million Series C financing at a $4.1 billion post-money valuation, co-led by CapitalG and Emergence Capital. New investors US Innovative Technology Fund (USIT), Bessemer Venture Partners, and Temasek joined existing investors 8VC, Georgian, Valor Equity Partners, Xora, Eclipse, Perry Creek Capital, NVIDIA, and others, bringing Bedrock's total funding to nearly $1 billion. The company says the money will accelerate the volume of Bedrock-equipped machines across commercial construction sites in the United States, helping contractors put existing heavy equipment to work more safely, predictably, and at greater capacity.

The proof point is paid work, not a demo. The announcement follows what Bedrock calls a defining milestone for physical AI: in mid-August 2026, fully autonomous excavators began performing paid work on live commercial job sites. That distinction matters. A demonstration proves a machine can move dirt; paid work on a live site proves a contractor will trust it with real work, alongside the crews and workflows already in place. Co-founder and CEO Boris Sofman put it directly: "We have proven that fully autonomous excavators can perform real work of value alongside the crews and workflows already in place on job sites. The next challenge is delivering that capability to more contractors as a dependable part of how projects get built." Bedrock says its total customer volume has more than doubled since 2025.

The blitz mechanism is the retrofit. Bedrock's systems retrofit existing machines, starting with excavators, and the company is building toward mixed fleets that can self-orchestrate across an entire job site. The retrofit choice changes the adoption math: a contractor does not replace its fleet or learn a new machine; it adds autonomy to equipment and workflows it already understands, and judges the system against work it already knows how to price. That is distribution through the installed base rather than through a new product cycle. It also fits the team's origin: Bedrock was founded in 2024 and is led by former Waymo engineers who helped pioneer the systems that made fully autonomous vehicles a reality on public roads. The company is applying that road-autonomy expertise to construction equipment first.

The demand side is a labor gap, stated in the company's own numbers. The announcement frames the gap plainly: through 2035, employers will need to fill approximately 1.7 million openings each year, both to support growth and to replace workers who retire, move into other occupations, or leave the field. In construction specifically, 87% of firms report openings for hourly craft positions, and 88% say those roles are as hard or harder to fill than a year ago. Emergence Capital general partner Jake Saper framed it as a physical problem, not a software one: "More than 20% of global GDP runs on industries that can't hire the skilled labor they need, and construction is the sharpest example... Bedrock's AI lets heavy equipment reason through and adapt to the unpredictability of a live job site."

The newest customer shows the expansion pattern. Haydon, a 100% employee-owned contractor and one of the largest general contractors in the Southwest, based in Tempe, Arizona, is the latest partner to deploy Bedrock's technology, on a remote site in Texas where crews will use the system for fully autonomous earthwork. Haydon plans to expand its use of Bedrock-equipped machines across building and heavy civil projects, including construction for advanced manufacturing, transportation, water and wastewater, data centers, and healthcare facilities, and will also test more advanced applications, including trench work and the automation of larger machine models. Haydon CEO and Chair Katie Haydon described the appeal as operational, not futuristic: "For us, the biggest win is a safer, more predictable job site... Bedrock's technology helps us work more efficiently and take on more of the work our communities need, without placing additional strain on our crews." Other new customers named in the announcement — Wagner Construction in Minnesota, Smith Trucking Company in Florida, B&S Site Development in Virginia, Native Excavating in Colorado, Clarkson Construction in Missouri, C.W. Matthews in Georgia, and selected ArchKey Solutions projects in Arkansas — extend Bedrock-equipped machines beyond the West and Central South into the Midwest and South Atlantic.

The $600 million funds the operating foundation, not just more machines. Bedrock says it is building the field, commercial, and operational infrastructure required to make autonomy a dependable part of everyday construction work, and it expanded its leadership team accordingly: Amy Jones Satrom joins as Vice President of Operations, bringing experience deploying complex physical systems at Amazon, Nuro, and Uber, and Lando Juarez joins as General Counsel after serving as Head of Legal for Uber's Autonomous Mobility & Delivery business, where he built commercial partnerships and regulatory pathways for driverless vehicles across multiple markets. CapitalG general partner Derek Zanutto noted the speed: Bedrock "moved with extraordinary speed from technical development to autonomous work on live commercial sites, just one year after emerging from stealth," and has "the rare combination of technical depth, operational discipline, and contractor partnership required to define a new category of industrial autonomy."

What this is not. This is not a measured result beyond what the announcement states. No machine count, no revenue figure, no pricing, and no safety statistics were disclosed, and none are claimed here. Trench work, larger machine models, and mixed-fleet self-orchestration are tests and a direction, not deployed capability. The honest test of this round is whether machine volume grows with the new customers, whether Haydon's trench-work and larger-model tests convert into deployments, and whether the field operations behind every machine keep pace — because "dependable" is the part of the pitch that has to be earned site by site.

Sources