Eight months ago we invested in Bedrock Robotics’ $270M Series B on the belief that the construction industry is an ideal market for the application of autonomous technology and that the Bedrock team was best positioned to build, deploy and operate it. Today, Bedrock’s technology is live on customer job sites. Bedrock is the first company in the world to have announced going from supervised testing to operating fully autonomous excavators on active commercial job sites, with no one in the cab and no one operating the machinery remotely. Bedrock’s autonomous excavation is powered by a proprietary, end-to-end machine learning model trained on real job-site data.
We're excited to announce our participation in Bedrock’s $600M Series C round at a USD $4.1 Billion valuation alongside co-lead investors CapitalG and Emergence Capital, plus new investors US innovative technology fund, Bessemer Venture Partners and Temasek Holdings.
We first invested in Bedrock via Georgian’s Pathfinder program, which lets us back emerging markets early, before the outcome is obvious. It gave us a front-row seat as Bedrock set out to move autonomous equipment from the test yard onto real projects. Now that it has, we believe the opportunity is even larger than we first imagined.
Eight Months, Six Job Sites, One Major Product Milestone
- Fully autonomous excavators are now working on customer job sites. In August, Bedrock announced its first 100% operator-out deployments across major U.S. infrastructure projects, including a water treatment facility in Nevada with Sundt Construction, a large-scale earthmoving project in Texas with Champion Site Prep and a 1.2-million-cubic-yard civil sitework project with Zachry Construction. These deployments are not test sites or demos, they are live commercial projects that represent an important step beyond controlled testing, demonstrating that Bedrock's technology is effectively operating in real construction environments where conditions change constantly.
- Bedrock hit its goal ahead of schedule. At its Series B, the company set a goal of fully autonomous customer deployments in 2026. It got there in under six months, roughly two years after its founding. In our experience, that kind of predictable execution is rare in a field where hard tech can take years to reach the real world and is often subject to unforeseen delays.
- A growing partner network and virtuous data flywheel. Bedrock now works with over 15 partners across 23 states and has more than doubled its commercial job site volume from when we first invested. Many of these partners have moved beyond the design stage. Bedrock has signed its first commercial agreements for fully autonomous deployments, a commercial milestone to match its technological one and continues to expand into new regions. The Company’s increasing in-field experience will help improve Bedrock’s autonomy stack by adding more real-world data to its model.
Why We Are Investing Again
Four things stood out when we invested in Bedrock at the start of 2026: a market with significant tailwinds, a team purpose-built for autonomy, an approach designed around how contractors actually work and an autonomy stack that improves with every deployment. Georgian’s experience with Bedrock over the past eight months has reinforced our belief in each of these items and contributed to our conviction that Bedrock has a leading position in the space over competitors.
1. Market Opportunity: Construction Has a Labor Capacity Problem
Bedrock CEO Boris Sofman has described the industry’s challenge as “a capacity problem, not a priorities problem”. Global construction spending is projected to grow from ~$13 trillion in 2023 to $22 trillion in 2040. Demand to build is rising across data centers, energy, manufacturing and aging infrastructure and the AI boom is adding to it: U.S. data center construction reached a record annual pace of about $85 billion in August 2026, up roughly 73% from a year earlier. Meanwhile, the construction industry must attract roughly 349,000 net new workers in 2026 and contractor backlogs stretch past eight months (nearly 10-months for data center work). The labor shortage is only projected to grow with ~40% of the current U.S. construction workforce expected to retire by 2031.
Bedrock works alongside contractors’ existing crews to fill growing labour gaps and perform repetitive, high-volume earthwork so skilled operators can focus where their experience matters most. Excavation sits at the front of that bottleneck. Every data center, factory and power plant starts with moving earth. Concrete, steel and electrical work can’t begin until the ground is dug and graded. Excavators are also hard to staff: they can take five years or more to master and often make up about 25% of a contractor’s fleet. In our view, that makes earthwork a high-value starting point for autonomy.
Excavation is also a starting point, not the ceiling of Bedrock’s opportunity. Construction sites rely on several types of heavy equipment working together: excavators dig and load, dump trucks haul and bulldozers and loaders move and level the ground. Bedrock specifically built its technology to extend far beyond a single machine type and plans to expand into each category of heavy equipment over time, ultimately aiming to automate coordinated fleets that work together across a site. We believe every new machine type addresses similar gaps for the same customers, expanding the market Bedrock can ultimately serve.
2. Team: Experienced Operators Who Have Deployed Autonomy At Scale
Building autonomous construction equipment is a complex technical challenge. Machines must understand their surroundings, respond to changing physical conditions and make decisions without constant human intervention. That is why the founding team’s experience was central to our original thesis.
Before founding Bedrock, members of the team held leadership roles at Waymo. CEO Boris Sofman led Waymo’s trucking and freeway driverless programs and CTO Kevin Peterson was Waymo’s head of perception. Co-founders Ajay Gummalla and Tom Eliaz bring hardware and software engineering leadership from Waymo and Twilio/Segment and COO Laurent Hautefeuille previously led business development at Uber Freight.
That experience matters for Bedrock’s roadmap. Perceiving the environment, planning movement and operating safely are challenges that carry across machine types and we believe a team that has already deployed autonomy on freeways is well positioned to extend it to dump trucks, bulldozers and loaders. Since our initial investment, Bedrock has continued to attract senior talent from companies such as Waymo, Uber, Meta and General Motors. With the recent addition of Operations Exec Amy Jones Satrom and General Counsel Lando Juarez, Bedrock Robotics now employs 137 people and will continue to hire top talent as a key use of investment funds.
3. Approach: A Practical, Intelligent Retrofit Working With Existing Machines and Labor
Contractors don’t need to buy new equipment to benefit from Bedrock’s autonomy models. The Bedrock Operator is a set of sensors, computers and software that can be retrofitted onto existing excavators in a same-day installation, with no permanent modifications. We believe this suits an industry where contractors have significant capital tied up in their fleets, since a retrofit adds autonomy without replacing heavy equipment at a large upfront cost.
Safety is not just one of the highest priorities at Bedrock, it is one of the main reasons the company was founded. Construction is one of the most dangerous industries in the world and we have conviction that Bedrock’s autonomy technology can change that. Bedrock’s system has been rigorously tested in real world environments and is built upon an extensive construction data set. The R&D team at Bedrock also brings hard won experience from Waymo - a system that has now driven more than 200 million miles at 10x human safety records. Bedrock’s system gives each machine a 360-degree field of view, stops automatically if a person or object gets too close and only operates autonomously in approved conditions, pausing if it detects a fault. Contractors described Bedrock to us as the only autonomy provider they considered credible for the constantly changing conditions of a job site.
4. A Proprietary Vertical-Specific Autonomy Stack that Improves With Every Deployment
Bedrock’s autonomy platform runs on a proprietary end-to-end machine learning model that perceives the site, plans its motion and carries out the work. That model sits within a broader proprietary autonomy stack. Rather than relying on hand-coded rules for each task, Bedrock’s model learns from real job-site operation. Job sites change daily, with shifting soil, weather, terrain and crews. In our view, a trained AI reasoning model can adapt to the inherent variability of construction sites in ways pre-programmed rules based systems struggle to match.
That is one of the core differences we see between Bedrock and many others in the market: Bedrock does not rely on a remote operator running the machine, or on automation that follows a fixed path across a controlled site. In our view, this is the difference between automation and true physical AI: traditional automation systems repeat a pre-set path, while Bedrock’s machines understand their surroundings, make their own decisions and adapt as conditions change, with no one at the controls in the cab or remotely. Bedrock’s model will continue to compound as it expands to more commercial job sites alongside its partners. Every new deployment collects tens of thousands of hours of diverse field data, making the whole fleet smarter in a continuous learning and improvement flywheel. We believe this expanding set of real-world operating data will help Bedrock build upon its leading position, relative to the company’s peers and become one of its most durable advantages as deployments continue to scale.
What’s Next
In eight months, we have seen Bedrock go from promise to proof: operator-less machines on live job sites, a growing commercial partner network and the team and technology to support its ambitions.
The next stage is scale: expanding deployments, demonstrating consistent and safe performance across a wider range of conditions and proving that autonomous equipment delivers lasting economic value to contractors. We believe Bedrock has the technology, team and market opportunity to execute on it. We’re excited to keep partnering with Boris, Kevin, Tom, Ajay, Laurent and the entire Bedrock team as they bring autonomy to the construction industry.