Bedrock Robotics

America needs to build like never before. Data centers to power the AI revolution, energy infrastructure, new housing units, and manufacturing facilities. But we can’t keep up. Construction projects that should take months stretch into years, while others sit stalled. The constraint? 500,000 open construction jobs that can’t be filled fast enough. We’re here to change that. Today, Bedrock Robotics is coming out of stealth with $80M from Eclipse and 8VC to bring advanced autonomy to the built world. Our Bedrock Operator transforms existing fleets of worksite equipment into fully autonomous machines, enabling builders to meet growing construction demands in the U.S. with greater quality, predictability, and safety. Founded by leaders that helped put the first driverless vehicles on public roads at Waymo, and joined by executives who helped scale companies like Uber Freight from zero to $5B in revenue, our early team has a powerful mix of technical know-how and operational expertise. And in only just over a year, it’s incredible we’ve accomplished so far. Bedrock is operating on test sites in AR, AZ, TX, and CA, with five forward-thinking partners, including  Sundt Construction, Zachry Construction Corporation,  Champion Site Prep, Inc., and Capitol Aggregates, Inc. And we’ve got so much more to build. Read more from our CEO Boris Sofman, on the company vision and where we’re headed: https://bit.ly/4kCAT2x Additional thanks to all our other investors, including:  Two Sigma Ventures, Valor Equity Partners, NVentures (NVIDIA's venture capital arm), Crossbeam Venture Partners,  The Raine Group, Tishman Speyer, Atreides Management, LP, AlRajhi Partners, and Samsara Ventures.

Construction gets labeled as slow to adopt new technology. But that hasn't been our experience. When the economics are compelling to contractors, and companies can unlock autonomous capabilities without replacing million-dollar equipment, adoption happens fast. Bedrock's partners are proving that every day. Our co-founder and CEO Boris Sofman spoke with Robotics and Automation News about how this moment in construction automation is different, and why the technology is finally ready for the complexity of real job sites. Read more of the Q&A here: https://lnkd.in/gty_VM6n

Our CEO Boris Sofman spent 5 years at Waymo watching self-driving cars go from "stressing over the first hundred miles" to millions of miles every week. The breakthrough? A fundamental shift in how autonomy systems learn. On the latest episode of "What's Your Problem" from Pushkin Industries, Boris sits down with Jacob Goldstein to share what happens when that same technology meets construction sites, and why heavy equipment might be where autonomy makes its biggest impact yet. Listen now to learn: - Why excavators can take 4-5 years to master (and what that means for the labor shortage) - The unexpected physics problems engineers face when machines start changing their environment - What expert operators can actually do with excavators, and why the diversity of use cases is both the hardest problem and the biggest opportunity Episode link in comments below ⤵️

Attending Foxglove Actuate this week? On Wednesday at 11:15am, don't miss our Co-Founder and CEO Boris Sofman share a behind the scenes look at how we took our autonomous construction systems from zero to one.

Physical AI represents one of the most significant technological shifts of our time. The convergence of AI with robotics is already unlocking transformative possibilities for construction and manufacturing to healthcare and beyond. We're proud to be selected for the inaugural Physical AI Fellowship with MassRobotics, Amazon Web Services (AWS), and NVIDIA. Along with the seven other companies in this first cohort, we're excited to bring intelligent automation to the physical work that builds and sustains our world.

Our digital Operator is learning to think ahead... Meet Moonrock, a Caterpillar 336 equipped with our autonomous systems, doing its thing on our test site in Texas. When Moonrock recently ran out of material in its target digging area, it adapted. Rather than dig at nothing, or stop work, the machine began strategically gathering nearby loose dirt into a pile to get a better scoop. We never had to hard code this behavior. The Operator developed a new strategy entirely on its own, with only a couple weeks worth of experience under its belt. This is the fundamental difference between traditional automation and Bedrock’s advanced autonomy— instead of relying on a system that requires engineers to anticipate every edge case and program specific responses, our machine learning approach lets the system discover solutions and react to changing conditions with the kind of intuition real-world job sites demand.

Same equipment, heavier impact. Did you know that Bedrock’s upfit process can add autonomous capabilities to an excavator in less than three hours? We’re not redesigning things from scratch, we’re making use of the well-built machines already in our partners’ fleets. No permanent modifications, no downtime, no hassle.

Three states. Three completely different challenges. All in one week. Recently, our excavators worked through blazing 96-degree heat and humidity in Texas, waist-deep flash flooding in Arkansas, and intense dust storms in Arizona. Anyone who's spent time on construction sites knows that "controlled conditions" don't exist in this industry. The weather doesn't care about your schedule, and the job has to get done regardless. The crews on site deal with everything Mother Nature throws at them—and they do it with skill and determination that never stops impressing us. Building autonomous systems for construction means building for the real world in all its messy, unpredictable reality. Every scorching day, every muddy site, every rapidly approaching haboob teaches our technology something new about what it takes to work alongside the best operators in the business. Huge respect to everyone who keeps projects moving no matter what the weather brings.